Research Collection

Doctoral Thesis

Synthesis of mixed α/β-peptides and NMR-solution structure ofF/OH-substituted and other β-peptides

Author(s): Mathad, Raveendra I.

Publication Date: 2007

Permanent Link: https://doi.org/10.3929/ethz-a-005423911

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

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

ETH Library

Diss.ETHNo. 17209

Synthesis of Mixed a/ß-Peptides and NMR-Solution Structure

of F/OH-Substituted and Other ß-Peptides

A dissertation submitted to the

EIDGENÖSSISCHE TECHNISCHE HOCHSCHULE

ZÜRICH

for the degree of

Doctor of Sciences

Presented by

Raveendra I. Mathad

M. Sc. Chemistry

Bangalore University, India

born June 16,1976

Citizen of India

Accepted on the recommendation of

Prof. Dr. Bernhard Jaun, Examiner

Prof. Dr. Donald Hilvert, Co-Examiner

Prof. Dr. Dieter Seebach, Co-Examiner

Zürich, 2007

I am deeply indebted to my

'Doktorvater'

Prof. Dr. Bernhard Jaun

and

Prof. Dr. Dieter Seebach

For their encouraging guidance, intellectual support

throughout my dissertation

Acknowledgements

I would like to take the opportunity to thank people who have been cooperative and

helpful throughout my studies.

First of all, I wish to thank my thesis supervisor Prof. Dr. Bernhard Jaun for giving me

the opportunity to work in his group and for his encouraging guidance, his intellectual

support and the freedom he gave me throughout my dissertation.

I am grateful to Prof. Dr. Dieter Seebach for the fruitful collaboration, and I have

benefited from his teachings which often led to interesting discussions and results. I am

thankful to him for providing me financial support during all these years.

I would like to thank Prof. Dr. Donald Hilvert for agreeing to be the co-examiner.

I would like to thank Albert K. Beck for his helpfulness, suggestions, and for the proof¬

reading of the experimental part of this work, and his kind support in many

administrative tasks.

I am thankful to NMR service team, especially, Brigitte Brandenberg for her patience and

recording excellent NMR spectra, Philipp Zumbrunnen and Rainer Frankenstein for their

kind cooperation.

In addition, MS-analysis of peptides by Mass-service team at Analytical Chemistry

should be greatly acknowledged.

I wish to thank my colleagues Zeena, Siegi, Stefan, Rafal, and Silvan who created a

warm and cooperative atmosphere and for the good times, which turned the duration of

my studies into a very pleasant period. I wish to mention that I had a nice time especially

during afternoon "coffee breaks".

Many thanks go to all the present and past members of the Seebach group for the

excellent cooperation and pleasant atmosphere to work in all these years. Thierry

Kimmerlin, Gerald Lelais, Dr. Francuise Gessier, Dr. David Hook, Dr. Marino Campo,

Dr. Estelle Dubost, Dr. Markus Löweneck and Dr. Stefania Capone for their willingness

to help and useful suggestions in the lab. I appreciate fruitful collaborations with Dr.

Oliver Flögel, Dr. Michael Limbach and their advice during the peptide synthesis, and

Dr. James Gardiner and Dr. Michael Badine for their critical reading ofthe thesis.

I thank all my friends Anil, Deva, Shiva, Srinivas, Anu, Vishwa, Vikas, Trishul, Sai and

Shailesh who made my stay pleasant. I could not forget Dr. Gopi and Dr. HanBo for

maintaining a lively atmosphere in the part of the lab in the beginning ofmy studies.

My sincere thanks to Prof. K. V. Ramanathan, Prof. N Suryaprakash, Dr. S. Raghothama

and Dr. G. A. Naganagowda at SIF, IISc, Bangalore, and Prof. G. Nagendrappa, Prof. N.

M. Nanjegowda, Prof. A. Pasha, Dr. Hariprasad and Dr. V. V. SureshBabu, Dr. H. N.

Gopi at the Chemistry department, Bangalore University.

I am very grateful my friends Chandru, Vasanth, and Karthik for sharing lots of

unforgettable moments and their continuous support.

I thank all my family members for their understanding and support all these years,

without them I would not have accomplished the degree. Last but not least, my warmest

thank goes to my beloved wife Savita for her support.

Publications

Raveendra I. Mathad, François Gessier, Dieter Seebach and Bernhard Jaun.

The Effect of Backbone-Heteroatom Substitution on the Folding of Peptides - A

Single Fluorine Prevents a ß-Heptapeptide from Folding into a 3u -Helix (NMR

Analysis)

Helv. Chim. Acta 2005,88, 266.

Dieter Seebach, Raveendra I. Mathad, Thierry Kimmerlin, Yogesh R. Mahajan, Pascal

Bindschädler, Magnus Rueping, and Bernhard Jaun.

NMR-Solution Structures in Methanol of an a-Heptapeptide, of a ß3/ß2-

Nonapeptide, and of an ö//-ß3-Icosapeptide Carrying the 20 Proteinogenic Side

Chains

Helv. Chim. Acta 2005,88, 1969.

Alice Glättli, Xavier Daura, Pascal Bindschädler, Bernhard Jaun, Yogesh R. Mahajan,

Raveendra I. Mathad, Magnus Rueping, Dieter Seebach, and Wilfred F. van Gunsteren,

On the Influence of Charged Side Chains on the Folding-Unfolding Equilibrium of

ß-Peptides: A Molecular Dynamics Simulation Study

Chem. Eur. J. 2005,11, 7276.

Gerald Lelais, Dieter Seebach, Bernhard Jaun, Raveendra I. Mathad, Oliver Flogel,

Francesco Rossi, Marino Campo, and Arno Wortmann.

ß-Peptidic Secondary Structures Fortified and Enforced by Zn2+ Complexation -

On the Way to ß-Peptidic Zinc Fingers?

Helv. Chim. Acta 2006,89, 361.

Dieter Seebach, Bernhard Jaun, Radovan Sebesta, Raveendra I. Mathad, Oliver Flögel,

and Michael Limbach.

Synthesis, and Helix or Hairpin-Turn Secondary Structures of "Mixed" a/ß-

Peptides Consisting of Residues with Proteinogenic Side Chains and of 2-Amino-2-

methylpropanoic Acid (Aib)

Helv. Chim. Acta 2006, 89, 1801.

Daniel Trzesniak, Bernhard Jaun, Raveendra I. Mathad and Wilfred F. van Gunsteren.

Simulation of an a//-Beta(3)-Icosapeptide Containing the 20 Proteinogenic Side

Chains: Effect of Temperature, pH, Counterions, Solvent, and Force Field on Helix

Stability

Biopolymers 2006, 83, 636.

James Gardiner, Anita V. Thomae, Raveendra I. Mathad, Dieter Seebach, Stefanie D.

Krämer.

Comparison of Permeation through Phosphatidylcholine Bilayers of N-Dipicolinyl-

Alpha- and Beta-Oligopeptides

Chem. & Bidivers. 2006, J, 1181.

Parts of thesis have been presented as posters:

Raveendra I. Mathad, Christian Noti, François Gessier, Dieter Seebach, Bernhard Jaun.

Structural Investigation of Fluoro Substituted ß3-Peptides

Swiss Chemical Society Fall Meeting 7th October 2004, Zürich University, Zürich.

Raveendra I. Mathad, Marino Campo, Francesco Rossi, Gerald Lelais, Dieter Seebach,

Bernhard Jaun.

Design and NMR Structural Analysis of Hairpin Turn In ß-Peptides: Stabilized by

Zn+

Ion Chelation

Swiss Chemical Society Fall Meeting 13th October 2005, EPFL, Lausanne.

Thesis Summary 1

Zusammenfassung 5

1 Peptide Containing Non-Natural Amino Acids 9

1.1 Introduction 9

1.2 Peptidomimetics 10

1.3 Peptides Containing Non-Coded Amino Acids 12

14 p-Amino Acids 13

1.5 Naturally Occurring ß-Amino Acids 15

1.6 Synthetic ß-Peptides and their Derivatives 18

1.7 The Discovery of Helical Structure in ß-Peptides 19

1.8 Conformational Analysis of Peptides 20

1.9 Interactions Determining the Secondary Structure ofa Peptide 21

1.10 Qualitative Analysis of Local Conformation 24

1.11 ß-Peptides 26

1.12 Determination ofthe Solution Structure of Peptides by NMR 29

Spectroscopy1.13 Conformational Properties of Peptides in Solution 30

1.14 Strategies for Resonance Assignment of Small Peptides 31

1.15 Adaption of the Canonical Procedure to Small Peptides 34

Consisting ofNon-Natural Amino Acids

1.16 Identification of Secondary Structures 35

1.17 Structure Calculation Using NMR Data 37

1.18 Evaluation of Structures 40

1.19 Conclusion 41

1.20 References 42

2 Structural Aspects of ß-Peptides Containing a-F and a-OH 45

Substituted ß-Amino Acids

2.1 Introduction 45

2.2 Peptides Containing Heteroatoms Directly Bound to the 47

Backbone

2.3 Results and Discussion 48

2.4 NMR Analysis ofthe a-Fluoro and Hydroxy-ß-Heptapeptides 50

2.5 3i4-(iW)-Helix Forming ß3-Tridecapeptide with a Central like-a- 57

F-ßhAla Residue: NMR Structure Determination

2.6 The Structural Analysis of a//-«-a-Fluoro-, a//-a,a-Difluoro-, 61

and all-u- Hydroxy ß-Hexapeptides2.7 Conclusion 66

2.8 References 67

3 Mixed-Helices: Synthesis and Structural Studies on Mixed 69

a-, ß- and y-peptides

3.1 Introduction 69

3.2 Synthesis of cc/ß-, ß/y- and a/y-Mixed Peptides 74

3.3 CD Spectroscopy 75

3.4 The NMR-Solution Structure of Mixed a/ß-Peptides 77

3.5 Conformational Analysis of Mixed a/y and ß/y-Peptides 82

3.6 Conclusion 84

3.7 References 85

4 Solution Structure of a ß3-Icosapeptide Containing 87

Homologues of the Twenty Common ProteinogenicAmino Acids

4.1 Introduction: The 3i4-Helix 87

4.2 Structural Analysis of a ß3-Icosapeptide by NMR 91

4.3 NMR-Solution Structures of an a-Heptapeptide 95

4.4 Conclusion 97

4.5 References 98

5 ß-Peptidic Secondary Structures Enforced by Zn2+ 99

Complexation: Spectroscopic Evidence

5.1 Introduction 99

5.2 Design of ß-Peptidic Hairpins 102

5.3 The Structural Analysis by NMR-Spectroscopy: Zn2+ 108

Complexation of Hairpins5.4 Conclusion 118

5.5 References 119

6 NMR-Solution Structure Investigation of Cyclic ß- 121

Tetrapeptides as RGD and Somatostatin Anologues

6.1 Introduction 121

6.2 Cyclic RGD Peptides 124

6.3 Design of ß-Peptidic RGD Peptides 125

6.4 Results and Discussion 127

6.5 Cyclic ß-Tetrapeptides as Somatostatin Analogues 131

6.6 Design and Structural Aspects of Somatostatin Analogues 134

6.7 Results and Discussion 136

6.8 Conclusion 140

6.9 References 142

7 Experimental Part 145

7.1 Abbreviations 145

7.2 General Methods and Materials 146

7.3 Chemical Shifts of Peptides 151

7.4 Synthesis of a/ß-, a/y-, and ß/y-Mixed Peptides 153

7.5 NMR Measurement and Structure Calculation 164

7.6 References 189

1

Thesis Summary

The perhaps most important group of ß-peptides consists of ß-amino acids that are

derived from the corresponding a-amino acids. They have been shown to fold into well

defined secondary structures such as 3u-helices, turns, and sheets at short chain lengths

in methanol solution. ß-Peptides have exceptional proteolytic stability towards a variety

of proteases, which makes them a promising class of cc-peptide analogues for the design

of low molecular weight drug-like molecules.

In the present work, a number of ß-peptides, mixed a/ß-peptides and their analogues

have been designed and synthesized through the Solid Phase Peptide Synthesis procedure

in search of new secondary structures. Furthermore, the solution-structures of a number

of ß-peptides and their analogues have been investigated by high resolution NMR-

spectroscopy. Resonances were assigned by a combination of DQF-COSY, TOCSY,

HSQC and HMBC techniques and the three-dimensional structures were derived from

nuclear Overhauser effects (NOE) and coupling constants, subsequently used in

SimulatedAnnealing molecular-dynamics calculations to arrive at the solution structure.

a) A systematic study of the structural preferences of ß-heptapeptides with central 2-

fluoro-, 2,2-difluoro- or 2-hydroxy-3-amino-butanoic acid residues of like and unlike

configuration has been undertaken by NMR-spectroscopy in MeOH solution. 3U-

Helices were found for the gew-difluoro and for the F- and OH-substituted derivatives of

»-configuration. The two compounds containing the central /z'&e-configured ß-amino-acid

moieties do not form helical structures over the full lengths of the chains. In all

structures, it was observed that the there was a preference for a conformation with an

antiperiplanar arrangement of the C-F and the C=0 bonds as depicted in I.

F H

Hw//£e-a-F-ßhAla

/NHF

antiperiplanarconformation

Me-a-F-ßhAla

Nil

gaucheconformation

2

However, the longer ß-tridecapeptide with a central like fluoro-ßhAla residue forms an

uninterrupted 3i4-helix by positioning the fluorine atom in a lateral position on the helix.

The C-F bond is not anti but nearly gauche to the C=0 group. This local conformation is

considered to be unfavorable due to stereoelectronic effects. The fact that the molecule

nevertheless forms an uninterrupted helix along the whole sequence points to a

cooperative effect in ß-peptides.

b) To extend our knowledge about the secondary structures of 'mixed' peptides, a series

of cc/ß-, ß/y- and a/y-peptides were designed and synthesized on solid support with

standard Fmoc-strategies. The NMR structural investigations of a/ß-peptides which

contained an Aib residue in alternating positions of the sequence showed that these types

of peptides form a new 14/15-helix stabilized by the alternating intramolecular H-

bonded-ring structures. On the other hand, certain mixed a/ß-peptides with proteinogenic

side-chains were found to form hairpin structures in solution.

c) A ß -icosapeptide II containing all of the 20 ß-amino acids with proteinogenic side

chains in a sequence that should stabilize a helix by hydrophobic interactions along one

strand on the helix surface, and by hydrophilic interactions on the two other strands

including two salt-bridges (cf. the 3i4-helical-wheel presentation) was analyzed by a

comprehensive NMR study.

Salt Bridge

The NMR-spectra of ß3-icosapeptide II in MeOH showed a large chemical shift

dispersion of the backbone protons that allowed the complete assignment of all protons.

A detailed NMR-structural investigation revealed that the ß3-icosapeptide formed a left-

3

handed (M)-3u-he\ix over its entire length in MeOH solution. The structure is

characterized by 14-membered hydrogen bonded-rings between the NH of residue i and

the OO group of residue i+2. The length ofthe 14-helix was ca. 30-35 À with six helical

turns, which is the longest helix ever observed.

d) To investigate whether 'short' a-peptides adopt stable helical conformations in

solvents such as MeOH, which are less polar than H2O, an a-heptapeptide III with a

central Aib unit was designed. In this sequence all a-amino acids were chosen to be

different to facilitate interpretation ofNMR spectra, and the inclusion of the central Aib

residue was chosen to promote helix formation.

jj Ë H J A H g S H

III $

The NMR studies showed that the a-heptapeptide did not assume any helical structure in

MeOH solution, although the sequence contained Aib, a strong helix inducing residue.

To our knowledge this is the first study of an a-peptide under the conditions which cause

ß-peptides of the same length to adopt secondary structures.

e) A series of ß-octapeptides was designed to test their coordinating ability towards Zn2+

and the influence of Zn2+ on stabilizing or destabilizing the secondary structure in MeOH

and in water. The ß-octapeptides contained a central ß2/ß3 tum-inducing segment and

ßhHis and ßhCys residues at the TV- and C-termini as potential ligands for Zn2+ ions. The

effect of the absolute configuration of amino acid residues in the sequence on the

formation of ß-peptidic hairpins has been studied. Further, a ß-oetapeptide containing u-

ß2'3-amino acid residues on both strands was studied for its propensity to form a hairpin

in water, in which unfolding is more likely to occur. From NMR analysis, it became

evident that the ß-oetapeptide IV that was designed to form a 314-helix changed its

conformation to a hairpin structure upon addition of Zn2+ in water, demonstrating that

Zn+

ions are capable of switching the secondary structures of ß-peptides.

-Y"

4

Zn

or"11 n - \J^\

,p 'S/ N-H 0=<

-^S OK NH

N-H 0"<NH

t „S—'

\_7 IV h^ X -,

—NH (^N'

HN-*

314-helix in the absence of Zn2+ hairpin-turn upon addition of Zn2+

f) Based on previous work on somatostatin mimics in the Seebach group, a series of

cyclic ß-tetrapeptides such as V were designed and synthesized to mimic RGD and

somatostatin, which are known to inhibit integrin receptors and SRIF receptors,

respectively, on the cell surface. The conformational preferences of these ß-tetrapeptides

in solution were investigated by NMR to correlate structure and activity. The ß2/ß3-

segment flanked by appropriate side-chains was incorporated into cyclic peptides to

induce a ß-turn.

H2NWNH2

T

P'hArg f yH

ohß'bAsp

}' \ 1=0

ß3hTrp i ym H2Nß3hLys

O

—m yHN

(° Hr/°

ß3hPhe HN

ß3hVal \_NH^

ß2hPhe

^=0 >=0

NH \ // v_.-

ß-tetrapeptidc as RGD analogue

HN

( PH ß3hThr

NHH^

V ß-tetrapeptide as somatostatin analogue

By NMR structural analysis, it was found that these cyclic ß-tetrapeptides form well

defined, rigid structures in solution, characterized by intramolecular hydrogen bonded-

rings and with all side-chains occupying approximately equatorial positions on the

macrocyclic ring. These low-molecular-weight ß-peptides were tested for biological

activity, the binding affinities were in the sub-micromolar range.

5

Zusammenfassung

Die vielleicht wichtigste Klasse von ß-Peptiden besteht aus ß-Aminosäuren, welche sich

von den entsprechenden a-Aminosäuren ableiten. Es wurde gezeigt, dass diese

wohldefinierte Sekundärstrukturen wie 3i4-Helizes, Schleifen und Faltblätter Kettenlänge

in methanolischer Lösung schon bei geringer ausbilden. ß-Peptide zeigen

ausserordentliche Stabilität gegenüber einer Vielzahl von Proteasen. Somit stellen sie

eine vielversprechende Klasse von a-Peptid-Analoga dar, um pharmazeutische

Wirkstoffe mit niedrigem Molekulargewicht zu entwerfen.

In der vorliegenden Arbeit wurde bei der Suche nach neuen Sekundärstrukturen eine

Reihe von ß-Peptiden, gemischten a/ß-Peptiden sowie deren Analoga mittels

Festphasensynthese hergestellt. Ferner erfolgte die Strukturaufklärung in Lösung für eine

Vielzahl von ß-Peptiden nebst Analoga durch hochauflösende NMR-Spektroskopie,

wobei sämtliche Signale über eine Kombination aus DQF-COSY, TOCSY, HSQC und

HMBC zugeordnet wurden. Aus Kem-Overhauser-Effekten (NOE) und

Kopplungskonstanten konnten geometrische Beschcränkungen abgeleitet und in Folge

mittels Simulated-Annealing-Molecular-Dynamics-Berechnungen die Struktur in Lösung

hergeleitet werden.

a) Zur systematischen Untersuchung der Strukturpräferenzen von ß-Heptapeptiden mit

zentralen 2-Fluor-, 2,2-Difluor- oder 2-Hydroxy-3-aminobutansäure-Resten in like- und

w«Me-Konfiguration wurden NMR-Messungen in Methanol durchgeführt. Dabei lagen

die ge/w-Difluor-sowie die F- und OH-substituierten Derivate in w-Konfiguration als 3h-

Helizes vor. Die beiden Komponenten mit den zentralen ß-Aminosäure-einheiten in /-

Konfiguration bilden keine helikalen Strukturen über die gesamte Kettenlänge aus. Es

Hess sich beobachten, dass eine Konformation mit antiperiplanarer Anordnung der C-F-

und der C=0-Bindung in allen Strukturen bevorzugt vorlag (vgl. Abb. I).

oß<\

IT W^JTF

w-ot-F-ßhAla antiperiplanare /-a-F-ßhAla SynklinaleKonformation ' Konformation

6

Das längere ß-Tridecapeptid mit zentralem /-Fluor-ßhAla-Rest bildet jedoch eine

ununterbrochene 3j4-Helix aus, in der das Fluor-Atom eine laterale Stellung einnimmt.

Die C-F-Bindung liegt nicht in antiperiplanarer, sondern vielmehr synklinaler

Orientierung zur Carbonyl-Gruppe vor. Diese lokale Konformation gilt infolge stereo¬

elektronischer Effekte als ungünstig. Die Tatsache, dass das Molekül nichtsdestotrotz

entlang seiner gesamten Sequenz als ununterbrochene Helix vorliegt, weist klar auf einen

kooperativen Effekt in ß-Peptiden hin.

b) Zur Erweiterung unserer Kenntnisse über Sekundärstrukturen „gemischter" Peptide

entwarfen wir eine Reihe von a/ß-, ß/y- und a/y-Peptiden und synthetisierten sie auf

Festphase mit standardisierten Fmoc-Strategien. Die NMR-Strukturaufltlärung von a/ß-

Peptiden mit einem Aib-Rest in sequentiell wechselnden Positionen ergab eine neuartige

14/15-Helix, stabilisiert durch alternierende intramolekulare H-Brücken-Ringstrukturen.

Andererseits bildeten gemischte oc/ß-Peptide mit proteinogenen Seitenketten in Lösung

Haarnadel-Strukturen aus.

c) Das ß3-Icosapeptid II wurde in einer umfassenden NMR-Studien untersucht. Dieses

enthält alle 20 ß-Aminosäuren mit proteinogenen Seitenketten in einer Sequenz mit

Helix-stabilisierender Wirkung durch hydrophobe Wechselwirkungen entlang einer Ecke,

sowie hydrophile Interaktionen und zwei Salzbrückcn in zwei Ecken auf dem 3i4-Helix-

Rad.

Satzbrücke

hydrophil

Salzbrücke

hydrophil

hydrophob

II

Die NMR-Spektren des ß3-Icosapeptids in Methanol zeigten breit dispergierte Signale für

die Rückgratprotonen und erlaubten eine vollständge Zuordung, Die detaillierte NMR-

Strukturuntcrsuchung von II in Methanol ergab eine linkshändige (M)-3i4-Helix über die

7

gesamte Länge. Charakteristisch sind die 14-gliedrigen, über H-Brücken geschlossenen

Ringe zwischen NH von Rest i und der Carbonylgruppe von Rest i+2. Die Länge der 3U-

Helix beträgt ca. 30-35 Â mit sechs helikalen Umgängen. Dabei handelt es sich um die

längste jemals beobachtete Helix.

o : noNo v. o

III i^

d) Zur Klärung der Frage, ob „kurze" a-Peptide stabile Helix-Konformationen in

weniger polaren Lösungsmitteln als Methanol einnehmen, entwarfen wir das a-

Heptapeptid III mit zentralem Aib. Innerhalb dieser Sequenz wurden nur

verschiedenartige Aminosäuren eingesetzt, um die Interpretation der NMR-Spektren zu

erleichtern. Ausserdem sollte der Einbau des zentralen Aib-Restes die Helix-Bildung

fördern. Gemäss den NMR-Studien nahm das a-Heptapeptid III keinerlei helikale

Struktur in methanolischer Lösung an, trotz des integrierten starken Helix-Promotors

Aib. Soweit bekannt, ist dies die erste Untersuchung an einem a-Peptid unter

Bedingungen, bei denen ß-Peptide derselben Länge Sekundärstrukturen ausbilden.

e) Eine Reihe von ß-Octapeptiden wurde entworfen, um deren Koordinationsneigung

gegenüber Zn2+ sowie den destabilisierenden oder stabilisierenden Einfluss von Zn2+ auf

die Sekundärstruktur in Methanol und Wasser zu prüfen. Diese Peptide enthielten ein

zentrales ß2/ß3-Segment, welches bevorzugt Schleifen ausbildet, sowie ßhHis- und

ßhCys-Reste am N- und C-Terminus als potentielle Liganden für Zn2+. Zusätzlich wurde

der Einfluss der absoluten Konfiguration der Aminosäurereste auf die Ausbildung von ß-

Peptid-Haamadeln untersucht. Darüber hinaus stand ein ß-Octapeptid mit w-ß2,3-

Aminosäureresten auf dem einen oder anderen Strang der Haarnadel im Mittelpunkt des

Interesses, sowie dessen Tendenz zur Haarnadel-Bildung in wässriger Lösung, worin die

Sekundärstruktur sonst eher aufbricht. Die NMR-Analyse zeigte klar, dass das ß-

Octapeptid IV, welches ursprünglich als 3i4-Helix vorlag, bei Zugabe von Zn2+ in Wasser

eine Haarnadel-Konformation einnimmt. Somit sind Zn2+-Ionen in der Lage, die

Sekundärstruktur von ß-Peptiden zu schalten.

8

314-Helix in Abwesenheit von Zn2+ Haarnadel-Schleife nach Zugabe von Zn2+

f) Auf der Basis früherer Arbeiten innerhalb der Gruppe Seebach zu Somatostatinen

entwarfen und synthetisierten wir eine Reihe von zyklischen ß-Tetrapeptiden (vgl. V),

um RGD und Somatostatine nachzuahmen, welche wiederum beide für ihre inhibierende

Wirkung gegenüber Integrin- bzw. SRIF-Rezeptoren auf der Zelloberfläche bekannt sind.

Zur Erschliessung von Struktur-Aktivitäts-Beziehungen untersuchten wir die

konformationellen Präferenzen dieser ß-Tetrapeptide in Lösung mittel NMR. Das ß2/ß3-

Segment, flankiert durch geeignete Seitenketten, wurde in zyklische Peptide integriert,

um eine ß-Schleife zu induzieren.

HZNL ^NH2

ß3hArg

HN

ß3hVal \NH

\

H ß2hAspN OH

) J=0

ß3hTrP çym HZNß3hLys

)—NH )-

^Oß2hPhe

ß-Tetrapeptid als RGD-Analoges

( HN

ß3hPhe HN^ (pH ß^Thr

^-•^NH^O

ß-Tetrapeptid als Somatostatin-Analoges

Die NMR-Strukturanalyse ergab, dass diese zyklischen ß-Tetrapeptide wohldefmierte

starre Strukturen in Lösung ausbilden, gekennzeichnet durch intramolekulare, über

Wasserstoffbrücken geschlossene Ringe, wobei sämtliche Seitenketten im Makrozyklus

annähernd äquatorial positioniert sind. Diese niedermolekularen ß-Peptide wurden auf

biologische Aktivität getestet. Dabei bewegten sich die Bindungsaffinitäten im sub-

mikromolaren Bereich.

9

Peptides Containing Non-Natural

Amino-Acids

1.1 Introduction

Proteins are essential for any living organism as catalysts (enzymes), receptors

and structural elements. Peptides, their shorter counterparts, act mainly as

neurotransmitters, neuromodulators or hormones. As such, peptides and their analogues

have long been used and are actively investigated as potential drugs in medicinal

chemistry. To design and make new proteins that rival or even exceed the properties of

their natural counterparts is a long sought-after but still very ambitious goal. The

exchange of specific amino acids or even whole stretches of residues of a protein by

other proteinogenic amino acids is now a standard procedure of molecular biology and

has become an invaluable tool for studying structure/function relationships in proteins.

However, the scope of possible modifications is widened enormously if non-coded,

synthetic amino-acids or their analogues can be introduced into a natural protein. Such

non-natural elements might provide stability against peptidases, stabilize secondary or

tertiary structure and/or alter folding pathways. The incorporation of a range of non-

coded amino acids is commonly achieved either by bioconjugation followed by NRPS

(non ribosomal protein synthesis), by chemical ligation1 or by the tRNA engineering

method [1]. Nature herself uses post-translational modification of proteins extensively in

order to change the properties of the initial ribosomal product, e.g. by making disulfide

bonds, by attaching sugar residues as address tags for transport, or by ./V-methylation to

change the hydrogen-bonding ability of a given amino acid. Even backbone C-alkylation

or the replacement of the carbonyl oxygen by sulfur in a particular peptide bond has been

found in proteins [2]. Post-translational phosphorylation is one of the most common

protein modifications that occur in animal cells. The vast majorities of phosphorylations

The most widely applied ligation method is the reaction of Af-terminal cysteine of a peptide with

C-terminal thioester of a peptide which is termed as native chemical ligation.

10

occur as a mechanism to regulate the biological activity of a protein and are transient.

The enzymes that phosphorylate proteins are termed kinases and those that remove

phosphates are termed phosphatases. In animal cells serine, threonine, tyrosine and

histidine are the amino acids undergoing phosphorylation. The largest groups of kinases

are those that phosphorylate either serines or threonines and as such are termed serine or

threonine kinases. So far, however, no larger units consisting of non-coded amino acids

or having a non-peptide based backbone have been found in naturally occurring proteins.

If protein chemists want to eventually replace elements such as an a-helix or a ß-sheet in

a protein by entirely non-natural building blocks it is certainly a good strategy to first

learn to build peptidomimetic units that assume a desired secondary structure.

1.2 Peptidomimetics

Native peptides often make poor drugs. Their bioavailability is low because they are too

polar to diffuse across the cell membranes and they are proteolytically unstable when

taken orally. The inherent flexibility of peptides enables interaction with multiple

receptors besides the target, and could result in undesired side-effects. Using an

endogcneous peptide as a drug can be risky because this might disturb a complicated

network of regulatory equilibria in which the natural substance is involved. One way to

overcome these limitations is to use a peptidomimetic rather than a peptide made of

coded amino acids. A peptidomimetic is a molecule that does not (solely) consist of

coded amino acids but mimics a peptide in terms of its three dimensional structure,

charge distribution, hydrophobicity and ability to engage in hydrogen bonding. The

design ofa good peptidomimetic rests on the knowledge of the conformational, chemical,

and electronic properties ofthe native peptide. It usually starts from the structure of an a-

peptide made from coded a-amino acids and introduces modifications either of

individual amino acid residues, the nature of the backbone linkage, or by replacing two or

more amino acids with a more rigid three dimensional template mimicking a desired

peptide conformation. Some partial structures which have been employed successfully

for the latter strategy are depicted in Figure 1.

11

Figure 1. Synthetic analogues used for stabilization of a particular local conformation in a peptide.

A more systematic analysis leads to the following approaches towards a modified peptide

(see Figure 2):

a) A change of the absolute configuration of an oc-amino acid (l to d).

b) Modification ofthe side chain to a non-proteinogenic one.

c) Introduction of cyclic elements within an amino acid.

d) Introduction of additional side-chains (e.g. a-alkylation).

e) Insertion of additional carbon atoms into the backbone (ß-, y-, ö-amino acids).

f) Introduction of cyclic elements connecting sequential residues: a bridge between

two side chains or backbone units, (e.g. disulfide bonds, metal binding,

dipeptidomimetic analogues [3], see also Figure 1).

g) Replacing backbone atoms with hetero atoms (0, N, S) or carbon, leading to

depsipeptides, azapeptides, sulfonamido-pseudopeptides, or carbapeptides. This

type of modification creates a peptide bond surrogate that encompasses

significant changes in polarity, H-bonding capability, and local conformational

preference.

sside chain modification

ß-alkylation . j^^R

cyclization f~* f *\cyclizatioili a!ß.del)ydration

^H2N/|NCOOH

N-alkylation a.alkylation

change of absolute

configuration

Figure 2. Possible modifications within an amino acid.

12

1.3 Peptides Containing Non-Coded Amino-Acids

Aib (a-aminoisobutyric acid), a non-proteinogenic amino acid with a quaternary Ccc-

center, is known as strong promoter of helices, attributed to steric interactions involving

the gem methyl groups [4]. The preferred conformations of Aib-containing peptides have

been studied extensively over the years [5]. Aib-homo and co-peptides forming 3i0-, a-

helical and ß-tum structures have been characterized. Extensions of these studies to

peptides containing non-proteinogenic amino acids with linear alkyl and cycloalkyl side

chains inducing both helical and turn conformations have been documented [6-9]. Most

acyclic, Aib-containing polypeptides are of fungal origin and known to form voltage-

gated channels in phospholipid bilayer membranes. The peptide antibiotics alamethicin

and zervamicin have a high proportion of Aib residues, which aggregate into amphiphilic

helices to function as voltage-dependent multilevel ion channels in bilayers [10]. The

naturally occurring peptaibols, which contain several a-aminoisobutyric acid (Aib)

residues, posses antimicrobial properties [11]. Figure 3 illustrates the structures of

symmetrical cc,a-dialkylglycines, which are higher homologs of Aib. Two classes of

amino acid residues can be considered, residues with linear alkyl side chains (Deg =

diethylglycine; Dpg = di-n-propylglycine; Dbg = di-n-butylglycine) and residues with

cycloalkyl side chains (1-amino-cycloalkane-l-carboxylic acids, AcnC, where n is the

number of carbon atoms in the cycloalkane ring) [12].

ÇH3 CH3

H3C CH3

1 1

n(H2C) (CH2)nf^\ /f\ s—S

UP L)-c"' (Ui2)n

H2N Y°H0

h2n\oh0

H2N>VH0

H2N Y H2N Y0 0

Aib dialkyl amino

acidcyclic dialkylamino acid

Diphenylglycine Adt

n=l,Degn=2, Dpgn=3, Dbg

n=l, Ac3c

n=2, Ac4c

n=3, Ac5c

n=4, Ac6c

Figure 3. Some a,a-disubstituted a-amino acids.

13

./V-Substituted poly-glycines termed peptoids [13], are a family of non-natural oligomers

with polyglycine backbone in which the side chains are attached to the peptide nitrogen

atom instead of the a-carbon. This structural modification has important consequences:

the substituent on the nitrogen atoms makes the formation of hydrogen bonds, usually

considered to be essential stabilizers of secondary structure, impossible. Because the

individual residues are achiral, both right- and left-handed helices can be expected .N-

Alkylated amino-acids have a reduced preference for trans amide bonds. Hence, the

occurrence of eis amide bonds, which - with the exception of proline - are scarce in

native proteins, should increase. This could have a profound effect on the secondary

structure, and therefore biological activity. TV-Methyl amino acids are found as natural

products and in naturally occurring peptides, cyclic peptides like cyclosporine [14], and

depsipeptides, exhibiting wide ranges of biological roles, including antibiotic, anticancer,

antiviral, and immunosuppressive activity [15]. They are also useful tools for stabilizing

various peptide backbone conformations (e.g. turns), and for obtaining structure-activity

information about peptides.

1.4 ß-Amino Acids

ß-Amino acids are another class of modified amino acids that have received great

attention recently because ß-peptides (oligomers of ß-amino acids) were found to form a

variety of secondary structures such as 14-, 12-, 10/12-, 8-helices, ß-shects, turns and

hairpins. ß-Amino acids are homologs of a-amino acids with an additional methylene

group in the backbone. Peptides built of ß-amino acids exhibit a variety of activities

ranging from antibiotic, antifungal to cytotoxic and are resistant to enzymatic degradation

both in vitro and vivo. So far, the use of ß-amino acid building blocks ranges from a

single substitution of the analogous a-amino acid, incorporation of cyclic ß-amino acids

and the synthesis of complete ß-peptides which mimic the function of an a-peptide.

Their applications are rapidly expanding in the design of bioactive peptide analogues

acting as receptor agonists and antagonists, MHC-binding peptides, antimicrobial

Theoretical conformational analysis of both a- and ß-peptoids have been carried out employing ab initio

calculations, showing that peptoids form helical structures with peptide bonds either eis or trans despitemissing hydrogen bonds.

14

peptides or peptidase inhibitors [15-18]. The group ofSeebach [19] and other groups [20]

have shown that "amphiphilic" helical ß-peptides (Figure 4) can inhibit apoUpoproteins

involved in lipid uptake and transport.

Hydrophobic cationio

Figure 4, General presentation of an amphiphilic 14-helix formed by the arrangement of positively chargedside chain and a hydrophobic surface of a helix.

Furthermore, highly bioactive cyclic ß-peptides (Figure 5) have been designed to mimic

the peptide hormone somatostatin with good affinities [21, 22]. Positively charged ß-

peptides such as oligo-ß-arginines and oligo-ß-lysines have been shown to be cell

penetrating agents [23]. Designed peptides with ß-amino-acids have been found to bind

to class MHC-I protein and they can also complex with both DNA duplexes and DNA-

single strands [24, 25]. These properties make ß-peptides potential candidates for the

development of peptide-based therapeutic agents.

Figure 5. Cyclic ß-tetrapeptide as somatostatin analogoue.

15

Our group has mainly focused on studying ß-peptides with proteinogene side chains

which are analogous to the peptides of natural origin. Unlike a-peptides, ß-peptides or

higher analogues are able to form secondary structures at short lengths, while retaining

the same or additional types of interactions which are essential for stabilizing any

secondary or tertiary structure in a protein. Studies on ß-peptides or higher analogues

might serve to deepen our understanding of complex features such as folding pathways,

protein-protein interactions, ultimately leading to applications of biological importance.

1.5 Naturally Occurring ß-Amino Acids3

Although ß-amino acids are generally regarded as unnatural amino acids, there are a

number of examples of naturally occurring peptides containing substituted ß-amino acids

isolated from marine organisms and various prokaryotes [26, 27]. The ß- and y-amino

acids are found in nature as free constituents or as building blocks in a number of

biologically active natural products. They have emerged as an increasingly important

class of compounds with diverse structures. Substituted ß-amino acids can occur in up to

four diastereoiosomeric forms which, compared to their a-counterparts, significantly

expands the structural diversity of ß-amino acids. A selection of natural products

containing ß-homoamino acids is displayed in Figure 6.

The most commonly encountered ß-amino acids in nature arc Asp (which is an a- and a

ß-amino acid at the same time) and ßhGly or ß-Ala. The formation of isoaspartate

moieties is the result of a rearrangement of the backbone in Asp- or Asn-containing

peptides during non enzymatic peptide and protein degradation [28]. Such ß-Asp residues

have been found in aggregated ß-amyloid-peptides in brain tissues affected by Alzheimer

disease. ßhGly is an essential amino acid, a component of pantothenic acid in coenzyme

A and of carnosine in muscle tissue. The ßhGly moiety can be incorporated both in

helices and turns [29, 30], and adds conformational flexibility due to its additional

torsional angle. A simple y-amino acid, y-aminobutyric acid (GABA) is the major

inhibitory neurotransmitter in the central nervous system (CNS) of mammals. The main

3The occurrence, properties and structural diversity of ß- and y-amino acids have been surveyed in a recent

review [26].

16

source or pathway for GABA biosynthesis is the enzymatic conversion of glutamate to

GABA by glutamic acid decarboxylase in neurons [31].

Other ß-amino acids are found in natural products possessing of variety of biological

functions. Examples include the antibiotic streptomycin F, which contains ß3-hOm. The

peptidic antibiotic Edine contains ß2-tyrosine and isoserine residues [32]. a-Hydroxy-ß-

amino acids are constituents of the natural product taxol (in the side chain), which is an

active antitumor agent. Cryptophycin, a potent tumor selective cyclic depsipeptide,

contains ß2-hAla. ß-Aryl-ß-amino acids are found in metabolites (cyclic peptides)

isolated from a sponge and showing antimicrobial and insecticidal activities [33].

ß3hGly

OH

H2N ^f OH

OH

phenylisoserine

O

GABA

OH

C02H

H,N',CONH,

ßAsn

oH2N" 'COOH

cispentacin

0

H2N' ^y "0H

NH2

2,3-diaminobutanioc acid

H2N'^^/^OH

p3hPgl

(CH3)3NH2

H2N^C°^H

ß3hOrn

H,N'"vAOH

ß2hAla

NH:

Dap

p3hAib

H2N^Ny'^0HÔH

isoserine

XxH2N^*^T>H

ß3hVal

H2N' "Y" ^0H H3N^Y^C02NH2 NH2

emenaminc

HoN^^OH ^N^^OH

ß3hPro

H2NOH H2N^Y^° I0" H2N^Y^°^f°Vll"N^YC

OHOOH

OA

HO

yC02H

,OH

OH OOHOOH

0

/\"

O

statin bestatin amastatin

Figure 6. Some of ß- and y-amino acids occurring as components of natural products.

17

The cyclodepsipeptide jasplakinolide contains ß-tyrosine [34]. Emericedins, recently

isolated from a microbial source, are betaines inhibiting long chain fatty acid oxidation.

Cispentacin possesses antifungal properties, bestatin and amastatin contain hydroxyl-

substituted ßhPhe and ßhLeu esters in depsipeptides, known inhibitors for

aminopeptidases [35].

Taurine (2-aminoethanesulfonic acid) is a small sulfur-containing ß-amino acid present

in large quantities in the intracellular space of the brain, retina, liver, kidney, heart, and

muscle in vertebrates. In mammals, taurine is known to serve many important biological

functions, including osmo-regulation, bile acid conjugation, membrane stabilization,

antioxidation and Ca2+ modulation [36].

2,3-Diamino-acids (ß-amino acids) are another important class of non-proteinogenic

amino acids present in plants and microorganisms. They show activity as antibiotics and

as inhibitors of proteases [37, 38]. The non-coded amino acid ßLeu (ß3hVal) is formed

by the migration of the a-amino group to the ß-carbon, during the metabolism of leucine

in mammals and plants. Derivatives of ß-amino acids are part of well known ß-lactam

antibiotics, penicillins and cephalosporins. Pepstatin, isolated from microorganisms,

contains statin (y-amino-ß-hydroxyearboxylic acid), which acts as an inhibitor of peptidic

asparatyl protease.

a,ß-Dialkyl substituted ß-amino acids (ß2'3-amino acids)4 are also found in natural

products, but their role remains open. One example are the dolastatins, which are cyclic

peptides containing a ß2'3-amino acid (2-methyl-3-aminopentanoic acid) [39].

ß -Amino acids occur rarely in nature. They have been found in some plants and in

bovine brain peptides which contain exclusively ß2hAla derivatives. The simplest

structure (Figure 7) is the dipeptide a-L-glutamyl-ß-isobutyric acid ethyl ester. The

ß2hAla found in plants has the (K)-configuration [40].

a) b)

>2C...v\^N^^C02Et H02C. ,.»v^AN^s^CNH. = \TH.

H I

Figure 7. y-Glutamyl dipeptides extracted from a) plants and b) bovine brain represented by the dipeptide

a-L-glutamyl-ß-isobutyric acid or ethyl ester. In plants the ß2hAla building block has (^-configuration.

The structural preference of ß2J-amino acids have been studied. (2S,3S)- ß2'3 -peptides form a 14-helix.

On the other hand (25,3/?)-ß2,3 -peptides were found to form a ß-sheet.

18

1.6 Synthetic ß-Peptides and Their Derivatives

Based on the type of side-chain modification, ß-peptides may be classed into three

different groups. 1) ß-amino acids bearing proteinogenic side chains [26] 2) 'rigorously

unnatural' ß-peptides (peptide side chains conjugated with carbohydrates or nucleic acid

bases [41-43]) and 3) ß-peptides derived from cyclic ß-amino acids. Lastly peptides

incorporating backbone N-, O-, S-atoms, hydroxy amine, sulfonamide or sulfonimide

groups in the backbone [44, 45], The classes of ß-amino acids which are derived through

side-chain or backbone modification are shown in Figure 8.

• Ç 9 /"Nucleic AciiT"^ .si R

(CH2)n H^

./ iL A-^A h v

W*C02Me

p-amino acids with ß-amino acids with ß-amino acids ß-amino acids with

proteinogenic side chains "conjugated" side chains with ringsmodified backbone

Figure 8. Classification of ß-amino acids based on the type of side chain substitutions.

19

1.7 The Discovery of Helical Structure in ß-Peptides

Background

Only speculative interpretations concerning the structure of monodisperse ß-peptides had

been reported in the literature before 1996 [46-51]. For oligomers like poly (ß-aspartic

acid), poly (ß-homoglycine) and poly [(5)-ß-amino butanoic acid], a ß-sheet type of

structure was proposed based on fiber X-ray, NMR and CD studies. A helical structure

was suggested for poly-(ß-aspartic acid) derivatives. Further structural investigations by

Drey and coworkers on linear and cyclic ß-peptides were unsuccessful. Several groups

were working on oligomers of ß-hydroxy acids, "oxa-ß-peptides", in order to obtain

information on their intra and extra-cellular degradation. The poly-ester, poly-((Ä)-3-

hydroxybutanoate) (PHB) is a naturally occurring unbranched biopolymer, produced by

microorganisms as an energy storage material in the cytoplasm of the cell. Apart from

occurring as a high molecular weight polymer as storage material, it also occurs in a low

molecular weight form detected in eukaryotic and prokaryotic cells. Oligomers of 3-

hydroxy butanoate are found to act as selective ion channels [52] for Ca2+ in membranes.

The exact structure of PHB that allows the ion transport in the membrane is not known.

However, different models have been proposed for transmembrane ion channels

involving PHB oligomers based on X-ray investigations of fiber structure, lamellar

crystals and crystal structures of cyclic oligomers. For poly-[(.ß)-3-hydroxy butanoates], a

2]-helical structure was determined. Seebach and coworkers prepared cyclic-oligomers

(oligolides) of (i?)-3-hydroxy butanoate and determined their structure by X-ray

scattering [53]. Two different types of structures were observed in the crystal, called d-

type and S-type. The backbone torsion angles of S-type were identical to those in the 2\-

helical structure suggested by Marchessault. A linear model of the oligomer was

generated using torsional angles from the d-type structure observed in the crystal,

resulting in a 31-helical structure with a pitch of 6 Â, which has not been found

experimentally, neither in the oligomer nor in poly-3-HB, Upon examination of the 3i-

helix model, it was observed that the C=0 bonds point in the direction of O-atoms bound

in the chain, which led to the idea that the replacement of the O-atom by NH might

stabilize helix formation in solution through H-bonding. In 1996, Seebach and coworkers

showed by NMR-spectroscopy that small ß-peptides with as few as six ß-amino acid

residues fold into a 3i4-(M)-helix in pyridine solution. This was the start of a structural

20

exploration of "ß-Peptides" and the wide variety of secondary structures they form in

solvents such as pyridine, and methanol.

The 3i4-helix is characterized by a 14-membered hydrogen bonded ring between the NH

of residue i and the C=0 group of residue i+2. There are three amino acids per turn with

a pitch of the helix of car. 5 À and the helicity is (M) or left handed when homologues of

L-amino acids are used [54].

Around the same time, Gellman et al. also reported the 3i4-helical structure for ß-

peptides containing cyclic ß-amino acids in the solid state and solution [55]. For y-

peptides consisting of only four residues with the appropriate substitution pattern, a 2.6i4

helical structure has been characterized both in the solid and solution states [56].

In the 3i4-helix, the carbonyl groups are parallel to the helix axis and point in the

direction of the C-terminus, resulting in a macrodipole opposite to that of the a-helix in

natural proteins which points from the C- to the TV-terminus. Depending on the type of ß-

amino acid used (ß3 and/or ß2, mono and/or disubstituted, cyclic and/or open), oligomers

fold into a variety of secondary structures. Besides the 3i4-helical structure, a 2.5i2-helix,

a 2.7io/i2-helix and a 2g-helix have been characterized to date. In analogy with a-

peptides, a pleated-sheet and a hairpin have been characterized [57, 58],

1.8 Conformational Analysis of Peptides

a-Peptides

Our knowledge of backbone conformations adopted by linear a-peptides rests on a large

database of protein structures as determined by X-ray diffraction and NMR-spectroscopy.

They revealed the canonical secondary structural elements such as the a-helix, the

parallel and antiparallel ß-sheet, or the 3io-helix and allowed to compilation of the

dihedral angle combinations that are characteristic for each.

Statistical analysis of the sequential combinations of several amino acids also provided

insights into the sequence dependence of secondary structure and tendencies of

individual amino acids to induce helices or turns. However, X-ray structures do not

reveal energetic differences between possible alternative conformers, and moreover, they

reflect the context of folded proteins where tertiary as well as secondary interactions are

21

important. In order to separate the contribution from tertiary structure from those inherent

in the peptide sequence of the secondary structure element, it would be helpful if one

could study a-peptides with typical lengths of an a-helix or p-sheet in solution.

Unfortunately, such a-peptides do not generally assume well defined conformations in

solution, cither in water, or, as will be shown herein in the less polar solvent methanol.

1.9 Interactions Determining the Secondary Structure of a Peptide

1) Local conformational preferences

The classical non-bonded steric interactions (1,2 1,3 and 1,5) have a profound influence

on the conformational stability of peptides. Analogous to alkanes, staggered

conformations (0 = ± 60, 180) are stabilized more than eclipsed arrangements due to

repulsion between vicinal occupied bonding orbitals (Pitzer strain). The backbone favors

a conformation in which OO and C-H bonds arc oriented in parallel, avoiding 1,3 and

1,5-repulsions between non-hydrogen atoms.

Stereoelectronic effects are another important contributing factor to the conformational

stability of proteins.

Electrostatic forces due to charge-charge (Coulomb), charge-dipole and dipole-dipole

interaction: the typical charge-charge interactions that favor protein folding are those

between oppositely charged groups. Dipole-dipole interaction refers to the interaction of

ionized side chains ofamino acids or with the dipole ofthe solvent molecule.

2) Sequential (i to i+1, i+n) interactions

a-helix

: !OR2 0 R4 H 0 R6

in extended3 hdix

structure

22

Hydrogen bonds: polypeptides contain numerous hydrogen bond donors and acceptors

in their backbones that can form hydrogen bonds between residues which are distant in

the sequence. Stacking side-chain interactions involving salt bridges, hydrophobic and

van der Waals interactions can be crucial for the stabilization of a particular secondary

structure.

3) Energy terms involving the whole peptide

Macrodipole: in an a-helix, the hydrogen bonds are almost parallel to the helix axis and

all carbonyl dipoles point in the same direction (TV- to C-terminus as shown in Figure 9).

This results in a macrodipole with the negative end pointing towards the C-terminal. The

partial negative charge on the C-terminal carboxylate as well as the positively charged

ammonium group at the TV-terminus opposes the macrodipole and this leads to a

destabilization ofthe helix.

N-terminus

v

0

C-terminus

Figure 9, Illustration of macrodipole in the a-helix.

4) End effects

In accordance with the helix dipole model, the charges imparted to the terminal ends of

the helix by the dipoles necessitate oppositely charged side chains at the termini to

counterbalance the charge.

The presence of the protonated amino group at the TV-terminus and of the deprotonated

carboxyl group at the C-terminus is helix destabilizing. The unfavorable interactions with

the dipole can be eliminated by 'capping' the charges with protecting groups [59-61],

23

which may also offer additional interactions with intra peptide amino acids in the form of

hydrogen bonds.

5) Solvation

The conformational preference and stability of a given protein or peptide is influenced

significantly by the surrounding solvent. Polypeptides contain numerous hydrogen bond

donors and acceptors in the backbone and side chains. The hydroxylic solvents in which

peptides are soluble are both good H-bond donors and acceptors themselves. Therefore,

surrounding solvent molecules compete with the peptide backbone groups for H-bonding.

The amino acids are in contact with the solvent to a different degree, causing a large

variability of strengths of other potential interactions. The hydrophobicity (solute-sovent

attraction/repulsion) of certain amino acids can result in the burial of the hydrophobic

residues in the core of the protein, thus restricting the available conformations into which

a protein may fold. A substantial component of the energy involved in protein folding is

attributed to charge-dipole interactions. This refers to the interaction of ionized side

chains of amino acids with the dipole of the solvent molecule. It is, therefore,

understandable that the majority of the amino acids found on the exterior surfaces of

globular (non-membrane) proteins contain charged or polar side chains.

Charge attenuation: polarization of the solvent near the charged solute atom results in the

formation of an induced solvent charge of the opposite sign, attenuating the electrostatic

interactions of the solute. This effect is considered as an effective reduction of the solute

partial charges.

The polar solvents methanol and trifluoroethanol (TFE) can induce stable conformations

in peptides that are unstructured in aqueous solution. Several properties have been

suggested to be responsible for this stabilization. Peptide-fluoroalcohol association is

possibly driven by hydrophobic effects. The dielectric constant of TFE more closely

approximates that of the interior of proteins. It is about one third that of water, which

should strengthen interactions between charged groups. TFE is less polar than water; it

favors intra-strand hydrogen bonding within the polypeptide over inter-strand hydrogen

bonding. Also, the poor hydrogen bond acceptor property of fluoroalcohols, limits its

ability to insert into a C=0-H-N bond, as compared to water [62, 63].

A regular secondary structure becomes possible, if a low energy local conformation in

regular repetition (3i4Jielical unit) or pattern repetition (10/12-helical unit) allows

24

favorable long range interactions (i+n) and is sterically possible in space without serious

van der Waals repulsions.

1.10 Qualitative Analysis of Local Conformation

The amide bond, the characteristic element of a polypeptide, gives the backbone a certain

rigidity. The atoms Cai-i, C(0)j-i, On, Ni, and Co,; lie essentially in the same plane, as

shown in Figure 10.

Figure 10. Presentation of structural elements present in a-peptide with Ca^, C(0)n, On, Nu and Ceci

atoms in plane.

The Ca centre plays the role of a pivot. All bonds connected to Ca are single bonds and

function as torsional axes. The conformation of each residue with respect to the rotations

about Ca-N and Ca-C(O) can be described by two sequential torsion angles: cp for C(O)-

N-Ca-C(O) and \|/ for N-Ca-C(0)-N. A plot of cp vs \p (Ramachandran diagram) is often

used to illustrate local conformations with respect to these two torsional angles that

correspond to stable secondary structures. Relative to glycine, the side chain of the other

amino acids leads to a rather restricted conformational space.

Local conformational analysis of polypeptides rests on the analysis of idealized Newman

projections along Nj-Ca; (cp) and Cai-C(O);. The backbone conformations with cp= 0°

and cp= 120° can be ignored since they result in an eclipsed arrangement of Ni-Cn and

Ca;.C(0)i or Ca;-Cßi bonds and 1,5-repulsion between 0m and C(0);or Cß; as depicted

m Figure 11a.

It is important to note that C(O) is actually bound to two non-hydrogen atoms. The

idealized conformations possessing minimum Pitzer strain are \p= 120° and \|/

= -60° in

which C(0)i-Ni+i or C(0);-Oj bond is sp-oriented relative to the Caj-Haj (Figure lib).

25

a) <(> -angles

H CO,.!

'eV^.Cß, ,aHs/tY°Pl^0 (fjT

,oc ]îo,<|> = 0° (H 180°

HMOC

*r^CßiTmCsL

VSo,, Vp^H1CO,

CO,

<t> = 60° <t> = -120°

*"r?f:' ,aHx

H

CO,

0,.iCCO,

<j) = 120° <t> = -60°

b) y-angles

Ha

1

(/si^^^Cp,

Nl+1

y = -60°

0

ißC^^Sr^(yJi+inI

Ha

V|/ = 120°

Figure 11. a) The idealized six NrCc^ conformations shovving repulsion for <p = 0, 60', 120' and

conformations with minimum strain for 9= 180°, -120°, -60° b) conformations about Ca,-Ci(0) with

minimum strain for y= 120° and -60'.

A stable secondary structure can be anticipated as soon as translational propagation along

the polypeptide chain allows formation of H-bonds without major alterations in local

conformation. An insignificant alteration leads to the natural a-helix with torsions cp = -

57° and v|/= -47°, in the vicinity of ideal values (cp = -60°, y = -60°). The deviation of

experimentally found values from the idealized pair of torsion angles has the effect of

reducing the number ofresidues per helix turn.

ß-Sheets, formed by peptide strands with intermolecular H-bonds, become possible if the

backbone atoms are ö/>-oriented5. The conformation with idealized torsion angles cp = -

120° and \|/= 120° is free from 1,3-interactions and suitable for such sheets. Typical

ap antipenplanar and sp. synpepriplanar arrangement.

26

experimental torsion angles are (9 = -119°, \j/= 113°) for parallel sheets and (cp = -139°, \|/

= 135°) for antiparallel sheets.

1.11 ß-Peptides

a)

v^-ev-11* \x-

a-amino acid

<p 6 v\

X

ß-amino acid

b)

V^V-y'" "Ytv" HVtVc^*»*!>*# i„^^R> xj<t>**

H H I H

H i^c^ H

(+)-SC -)-sc ap

Figure 12. a) Presentation of torsional degrees of freedom in a- and ß-amino acids b) conformations

present in ß3-amino acid.

Formal homologation of a-amino acids to ß-amino acids introduces one additional

degree of torsional freedom. The additional dihedral angle Ni-Cß-Ca-C(O); is called 0

and gives a total set of three torsional angles cp, 0, y to describe the local conformation of

each residue.

In contrast to the torsion along cp and v|/, which have a six-fold periodicity because the

corresponding bond joins sp2 and sp3 centers, the potential energy profile of 0 shows a

classical ethane-like three fold periodicity.

Therefore, conformations of ß-peptides can be studied by analyzing the main three

torsional degrees of freedom (cp, 0, \|/) of a ß-amino acid monomer. The additional

conformations on ß-peptides can be represented as staggered ethane like conformers as

shown in Figure 12.

27

ß3-amino acids

favor a gauche conformation

ß2-amino acids

favor a gauche conformation

like ß2,3-amino acids

favor a gauche conformation

unlike ß2,3-amino acids

favor the trans conformation

ßhGly: conformational^ flexible

favors gauche or trans

Figure 13. The preferred conformation of differently substituted ß-amino acids.

The newly introduced bond has a higher barrier of rotation (C(sp3)-C(sp3)) than the two

sp2/sp3 bonds in a-peptides and the three resulting conformers differ significantly in their

stabilities. By qualitatively analyzing the local conformations of differently substituted ß-

amino acids, the preference for secondary structures in the corresponding ß-peptide

oligomers can be predicted. Helical or turn-like conformations of ß-peptides require a

gauche (-spl+sp) conformation around the Ca-Cß torsional angle 0, while the trans (ap)

conformation around 0 leads to an extended conformation which allows formation of

sheet-like structures. At first, it was expected that the additional degree of freedom 0 in

ß-amino acids would lead to higher conformational flexibility and large conformational

space. Hence, the effects of substituents on the conformation of ß-amino acids have been

the subject of extensive experimental and molecular mechanics/dynamics calculation

studies. It turned out, however, that the conformational space and the flexibility are

R O

jtNMt

O H

R1 O

*N-\-V

R1 O

*nAA62

O H

*YjuH J

H On

*N*

*N^At

28

severely restricted by substitution at Ca and/or Cß positions. The unsubstituted ß-amino

acid ßhGly, shows flexible conformations very similar to Gly found in a-amino acids [9].

In the crystal structures of small peptides containing ßhGly, 0 values corresponding to

both trans (0 = ± 180°) or gauche conformations (0 = ± 60°) were found [29, 64] When

carrying a substituent in either the Ca or the Cß position, however, ß-amino acids

strongly favor a gauche conformation.

One means of enforcing a particular 0 value is to incorporate the Ca-Cß bond into a ring.

This has been demonstrated for cyclopropane, cyclobutane, cyclopentane or cyclohexane

and it has been observed that the ring size6 actually determines the type of helix in such

relatively rigid backbones [65, 66]. In the non-cyclic residues, substitutions at a, ß

positions with like configuration, strongly favor a helical conformation as both side

chains can occupy equatorial {lateral) positions in the helix, whereas with the unlike

configuration a trans conformation is favored, because gauche would lead to an axial

orientation of the substituents in the helix which is impossible due to clashes between

successive helix-turns. For the same reasons, ß-amino acids geminally disubstituted7

either at the a or at the ß carbon are not anticipated to form helical structures, in contrast

to dialkyl a-amino acids such as Aib, which are known to induce helical and turn-like

conformations in a-peptides. The effects of substituents on the local conformation of a ß-

amino acid are summarized in Figure 13-14.

"X ;

Tf

ioft>\û4&n*

axial

positions

-N, "alioweci'

Intern!

positions

Figure 14. Schematic representation of a (A/)-3i4-Helix model.

The cyclopentane, pyrrolidine, piperidine and cyclohexane rings prevent full rotation around Cot-Cßbonds of amino acids in their corresponding ß-peptides and determine the type of structure.

7 ß2,2-Disubstituted amino acids induce a turn stabilized by 10-membered H-bond as found in crystalstructure of a ß-tripeptide with geminally disubstituted residues.

29

1.12 Determination of the Solution Structure of Peptides by NMR

Spectroscopy

An overview

It is known that, in many cases, peptide fragments of proteins exhibit significant

conformational preferences in solution. To identify their preferred conformation is of

great interest and significance, in particular to those who seek insights into the earliest

events in protein folding [67].

Among the various spectroscopic methods, circular dichroism (CD), fluoroscence, IR,

Raman, and nuclear magnetic resonance are widely employed for the elucidation of

peptide or protein conformation. In terms of spatial resolution, NMR is, in principle, the

most powerful of these methods. In favorable cases, it can derive nearly the same three-

dimensional structural information as X-ray crystallography. However, each

spectroscopic technique has its own time scale: the various forms of optical spectroscopy

can distinguish species that interconvert on a subpicosecond timescale, whereas NMR-

spectroscopy can distinguish exchanging species only if they have life times in the range

of milliseconds. This "slowness" ofNMR is simultaneously a weakness, namely when it

comes to unraveling a complex mixture of rapidly exchanging conformers, and a strength

because NMR allows the nature and even the rates of such dynamic processes to be

investigated.

A clear advantage of the NMR technique is that molecules can be studied in solution at

the specific pH, temperature, ionic strength, etc. that is characteristic for the

physiological environment. The activity of peptides depends on their three-dimensional

solution structure: molecules having a predominant conformation in free solution that is

close to the conformation they adopt when bound to a receptor will exhibit higher

binding constants, because the entropy of binding is less negative (they are

"preorganized").

High resolution liquid-phase NMR-spectroscopy is ideally suited to study intra- and

inter- molecular interactions, and conformational changes induced by them. It is regularly

used to investigate processes such as protein folding or protein/receptor interaction and to

map structure activity correlations.

30

Secondary structure motifs (e.g. helices, turns and sheets) play an important role as

nucleating centers in protein folding. The presence of such secondary structures can be

recognized by NMR based on characteristic vicinal coupling constants (local

conformation) and nuclear Overhauser effects between protons that are close in space but

belong to neighbouring or even sequentially remote residues.

1.13 Conformational Properties of Peptides in Solution

NMR-spcctroscopy is the only technique to date for experimentally determining the

structures of flexible or semi-flexible linear peptides experimentally with atomic

resolution. The NMR experiments provide local (site-specific) conformational

information. However, due to the slow timescale of the experiments, all information is a

population-weighted average over all structures in the conformational ensemble, Thus,

the interpretation of the NMR parameters for deriving peptide conformation requires

some caution. When applied properly, however, NMR is an exceptionally useful

technique to determine the structures of the dominant conformers formed by peptides in

solution.

Protein folding involves the transition of largely disordered conformations into a compact

and highly ordered native structure and shows acceleration by cooperativity effects due to

local and non-local interactions within and between the subunits. Peptides do not show

such cooperativity effects but nevertheless, retain all of the steric, hydrogen bonding and

electrostatic interactions present in proteins and therefore, peptides are anticipated to fold

into defined structures in spite oftheir modest sizes.

31

1.14 Strategies for Resonance Assignment of Small Peptides

Structure determination of unlabeled biopolymers, peptides and oligonucleotides by

NMR is usually done according to the following protocol:

1. Residue specific assignment; 2D DQF-COSY to assign spin systems through 2-

and 3-bond connectivities, and, in the case of strong overlap, TOCSY to identify

the protons belonging to the same amino acid.

2. Sequential assignment by 2D-NOESY to identify flanking residues through short

range NOEs.

3. Correlation ofnon-neighbouring residues through long range NOEs.

4. NOEs are classified as intra residue, sequential and long range and converted into

distance restraints for structure calculation,

5. Torsional angles (dihedral) can be derived {Karplus relation) from three-bond

coupling constants that are extracted from ID-spectra or qualitatively estimated

from DQF-COSY cross peak intensities.

6. Once the assignment of the coupling network and unambiguous interpretation of

NOE spectra has been accomplished, three-dimensional structures of a molecule

that are consistent with the NMR-derived constraints can be generated with the

help of restrained simulated annealing molecular dynamics calculations.

Although well established by now, this procedure is still not trivial: even after the

assignment and extraction of geometric constraints (distances, dihedrals) the relation

between the quantity of experimental data and degrees of conformational freedom may

still be insufficient to solve the structure. Spectral overlap and redundant amino acid

sequences can also prevent extraction of sufficient NOEs to arrive at an unambiguous

well defined structure.

Whereas 2D-NOESY is certainly the method of choice for the acquisition of precise

NOE data in the case of large proteins, medium sized molecules such as peptides often

have correlation times that are near the zero-crossing of the NOE vs ic curve (cûo.Tc ~ 1).

In these cases, one has to resort to the measurement of ROESY spectra (NOE in the

transverse plane). While NOESY spectra may show COSY-type artifacts, and NOESY

32

cross peaks cannot be distinguished from cross peaks due to chemical exchange, NOESY

is still a much less artifact-ridden technique than ROESY,

In addition to COSY-type artifacts, ROESY spectra can show TOCSY type peaks

between coupled spins. Since these have a phase opposite the NOE cross peaks, strong

TOCSY artifacts may partially or totally cancel the NOE between the same protons, a

situation that can not be diagnosed readily by inspection. Perhaps the most invidious

errors in ROESY are TOCSY-NOE relay transfers ofthe type shown below:

large J -*- TOCSY

Modified ROESY techniques (clean ROESY) have helped to reduce these errors as well

as the notorious offset dependence of ROESY cross peak volumes but in terms of

reliability of measured NOEs ROESY remains clearly inferior to NOESY. On the other

hand, both exchange cross peaks and cross peaks due to spin diffusion have the opposite

phase as the true NOEs peaks in ROESY spectra, and can easily be identified by

inspection. This is a distinct advantage ofROESY over NOESY.

Other parameters such as chemical shifts or temperature coefficients of amide protons

can contribute additional information and can supplement the NOE and coupling data to

confirm the presence of a secondary structure such as a helix or a turn. Deviations of

chemical shifts from the random-coil values [68] are often considered as evidence for

secondary structures in proteins. However, this empirical use of chemical shifts rests on a

large database of proteins with known structure, which is available for a-pcptides in

proteins but not yet for peptides with non proteinogenic residues such as ß-amino acids.

33

As with NOEs and coupling constants, dynamic averaging in multiconformational

ensembles makes such a use of chemical shift data almost impossible.

The reduced amide proton exchange rates, on the other hand, are a general tool to

recognize protons that are involved in "good" hydrogen bonds, regardless of the

particular constitution or size of the peptide.

The standard protocol for the structure determination of proteins as originally developed

by Wüthrich [69] and delineated above fails if the molecular weight of the protein

increases beyond ca. 30 kDa.

For large proteins, apart from the increased overlap in COSY and NOESY spectra,

slower tumbling (increasing correlation time) leads to a large line width, comparable or

even larger than homonuclear scalar coupling constants.

These difficulties can be surmounted by using fully (13C, 15N)-labeled proteins and by

using the larger heteronuclear coupling constants for correlation (3D experiments such as

HNCA, HNCO).

An important further improvement was achieved by the introduction of the TROSY

technique. It explores the fact that the effects of chemical shift anisotropy relaxation and

dipolar relaxation cancel each other under favorable circumstances at high magnetic

fields leading to narrow lines even for very large proteins. Using these advanced

techniques, labeled proteins with molecular weights up to 100 kDa have now been

investigated by high resolution NMR-spectroscopy [70].

For non-natural peptides and proteins containing residues that have to be synthesized

chemically, the use of fully labeled material is often impossible or would be extremely

costly.

34

1.15 Adaption of the Canonical Procedure to Small Peptides

Consisting of Non-Natural Amino Acids

BioNMR

protein structure and dynamics

Canonical procedurefor proteins by Wuthrich

Adaption

Canonical procedureXPLOR-simulated annealing

approach for structure calculation

Small peptidesstructure and dynamics

(6-20 amino acids)

XPLOR approach for

molecules existing in

one dominant conformation

Compared to the proteins consisting of proteinogenic amino acids, the NMR analysis of

non-natural peptides encounters both additional difficulties and opportunities. Apart from

the major obstacle caused by several populated conformers, the additional bonds

introduced in ß-, and y-peptides lead to more complicated spin systems and can make

assignment more difficult (e.g. ot-CH2 vs y-CH2 of side chains). On the other hand, if the

vicinal coupling constants across the additional bond can be determined, additional

constraints defining the backbone conformation may be deduced. Often the size of the

coupling constants can only be qualitatively extracted from cross peaks in DQF-COSY

spectra, but it still allows restraining the dihedral angle.

Except for ß2'3-disubstituted amino acids, the additional CH2 group in the backbone

makes it necessary to diastereospecifically assign the two protons. The usefulness ofboth

NOE and J-coupling restraints depends on this assignment and if not assigned, one has to

use pseudo-atoms instead. This usually leads to a loosely defined local conformation in

the corresponding section ofthe backbone.

35

Geminally disubstituted centers in the backbone (e.g. of Aib), reduce the number of

3J(1H-1H) values available. In this case, only NOEs and heteronuclear cross peak

intensities in HMBC spectra allow to determine possible dihedral constraints and - if the

two ligands are identical diastereospecific assignment.

1.16 Identification of Secondary Structures

A limited number of regular secondary structure motifs such as the a-helix, ß-tum and ß-

sheet are found in proteins. Each has its characteristic NOE cross peak partem and

3J(NH-Ha) coupling constants as summarized in Table 1.

Table 1. Theoretical values of 3J(HN-Ha) typical of secondary structural elements in a-peptides.

secondary structure e 3J(HN-Ha)

a-helix -57 3.9 Hz

310-helix -60 4.2 Hz

antiparallel ß-sheet 439 8.9 Hz

parallel ß-sheet -119 9.7 Hz

The observation of sequential (NH;, NHj+i) NOE cross peaks indicates the presence of a

helical structure. Strong (Ha;, NHi+i) NOE connectivities in the absence of observable

(NHj, NHi+i) NOEs indicate extended backbone conformations. In general, the long-

range NOEs are the best indicators of secondary structure; their pattern often

immediately identifies the type of secondary structure.

To date, four different types of secondary structures have been identified and

characterized for ß-peptides, consisting of either ß3-, ß2-, ß2'2-, and ß2'3-amino acids as

building blocks. The identified secondary structures are the 3i4-helix, the 28-helix, the

12/10-helix, a turn and the pleated sheets (See Table 2).

36

Table 2, Theoretical values of 3J(HN-Hß) typical of secondary structural elements in ß-peptides.

Secondarystructures of

ß-peptide

JJ(HN-Hß) Hz

3JHß-Ha Hz

Typical NOE patterns observed

14-helix 8.5-10

12/10-helix 7-9

9.9-12

8-helix 9-10

ß-peptidic 9-10.5

turn

NHi to NHi+1, aH^ M, aH»,, pHi+2, pHi+3

NHjto pHi+2, ßHi+3, ßHi.2, pHj.3, aHi+2

pHi to aHi+2, aHj+3

NHi to NHi+i, pHi.i, am-i

aHj to NHj+i, presence of interstrand

NOEs

The qualitative inspection of NOEs may indicate the presence of a specific secondary

structure present in p-peptides. For the 14-helix, a continuous series of relatively weak

(NHi, NHi+i) NOEs is accompanied by several medium (i+2/3) NOEs (NH, HNj+2), (Hßis

NHi+3) and (Hßj, Hai+3). For extended conformations, large coupling constants between

3J(NH-Hß) (9-10 Hz) and strong sequential NOEs but no (i+2/3) NOEs are observed. For

example, in a helix, a-H^iai exhibits a large and a-H/atera/ a small coupling to Hp. The

large 3J(1H-1H) -9-10 Hz, corresponds to an antiperiplcmar arrangement of respective

protons.

Structures found in ß-peptides are inherently more stable than those in a-peptides, a 314-

helix is detectable in solution by NMR-spectroscopy, while the 3.6i3-helix of an

analogous a-hexapeptide8 is not observed in methanol [71, 72],

NMR analysis of a-heptapeptide with a central Aib residue showed no secondary structure, but fold wasdetected at the central segment of the peptide due to helix inducing property of an Aib residue.

37

1.17 Structure Calculation Using NMR data

NMR Spectroscopy

Sequential resonance assignment

Qualitative analysis of

3J values and NOE patterns

Consistent with

single structure?

No Not suitable for structure

calculation

Yes

a) Determination ofNOE cross peak volumes by fitting and integration

b) Conversion of cross peak volumes into distance constrints bycalibration with known distances (two spin approximation)

c) Extraction of torsional angles constraints from 3JHH via Karplus equation

Structure Calculation : Simulated annealing or Distance geometry

Topology, Sequence, Force field: (empirical bond lengths,bond angles, improper angles, van der Waals radii)

Structure Refinement

restrained dynamics:energy minimization

Analysis of final structures

Scheme 1. Schematic presentation of the process for detennination of solution structure of peptides fromNMR data.

38

The structure calculation involves the conversion ofNOE cross peaks into distances and

derivation oftorsion-angle constraints from ^('H-1!-!) coupling constants.

There are two different methods available for calculating a peptide or protein solution

structure consistent with NMR data. The first method, distance geometry (DG), is based

on a calculation of matrices of distance constraints for each pair of atoms from all

available distance constraints, bond and torsion angles as well as van der Waals radii.

The second approach is force-field/molecular dynamics simulated annealing (SA) [73].

This method tries to locate the global minimum region of a target function made up ofthe

sum of empirical and experimental NMR-derived terms. A force field incorporates

potentials for classical terms such as covalent bond lengths, bond angles, planarity, and

absolute configuration. The non-bonded term is represented by an empirical energy term

comprising the repulsive part of van der Waals potentials. Therein, additional potentials

representing upper distance and dihedral angle restraints from the NMR experiments are

introduced.

Both established methods for calculating structural bundles (distance geometry [DG] and

MD/simulated annealing [SA]) try to generate structures that are compatible with a given

set of experimental distance and dihedral angle restraints.

If the constraints (target values) were in fact derived from time-averaged values from a

multiconformational ensemble, any attempts at structure calculation must fail, because

there is no single molecular conformation corresponding to these geometric constraints.

It is therefore of paramount importance to first qualitatively inspect the set ofNOEs and

J values for internal consistency with a single dominant conformer. If indications for

superimposed patterns, e.g. two types of helix, or helix and tum are found, no calculation

should even be attempted and the resulting analysis has to remain qualitative.

At the beginning, a randomly folded or extended starting structure is calculated from the

empirical data using only bond lengths, bond angles, configurational restraints and the

known amino acid sequence. The protocol then tries to fold the starting structure in such

a way that the experimentally determined inter-proton distances and dihedral angles are

satisfied by the calculated structure. In order to achieve this, each constraint is assigned

an energy potential, which has its minimum at the target value. Without the

experimentally determined distance and torsion angle constraints from the NMR

experiments, the molecule can adopt a large number of conformations due to the free

39

rotation around its single bonds. If one would use simple energy minimization from a

given starting structure only the nearest minimum would be found. The central task is

therefore to fully sample conformational space, a difficult problem well known in

computational chemistry in general. For few rotable bonds, this can be approached by

systematic variations of torsional angles in the starting structure, whereas stochastic

methods such as Monte-Carlo have to be used for larger systems.

a)

global minimum

b)

,

\ target /distance /

no limit

exceptvander

Waaîs

repulsion

\ . ^

\t^j

\ ilower upper

limit ''mit

dNOE

Figure 15, Diagram a) showing potential energy with global minima separated by several local minima, b)violations of the lower and upper limits of the NOE distances.

The MD/simulated annealing technique tries to do extensive conformational sampling by

first carrying out a molecular dynamics simulation at very high temperature. The kinetic

energy is then slowly taken out of the system ("cooling in silico") and, at the same time,

potentials of the NMR constraints are slowly switched on (increasing force constants

with time). Below a certain kinetic energy, the system will be trapped in a given local

minimum. By repeating this process over and over, each time with a different random

initial distribution of kinetic energy, one hopes to eventually sample all possible local

minima (Figure J5).

40

1.18 Evaluation of Structures

A successful calculation yields structures that represent a good fit of the experimental

data.

An assessment of the completed structure determination should be based on the

following criteria: 1) that the experimental interproton distance and torsion angle

restraints must be satisfied within the errors of the experimental data (no violations); 2)

that the deviation from idealized covalent geometry and van der Waals penalization

should be very small. Out of a set of typically 50-100 calculated structures, those that do

not fulfill critérium 1 are eliminated. The remaining structures are then sorted according

to the calculated energy which reflects critérium 2.

A meaningful representation of solution structures involves superimposing bundles of

conformers calculated with different starting structures with the same NMR constraints.

The precision of the structures can be evaluated. A best-fit (backbone) superposition of

all the calculated structures and a plot ofthe atomic root mean square deviation (RMSDs)

distribution with respect to the mean then allow one to quantitatively assess the precision

associated with different regions of the peptides. A small RMSD value for backbone

atoms is representative of high quality NMR-structure determination, the core of the

polypeptide chain may be comparable to X-ray determined structures, but the side chain

regions of peptides are usually less well defined. It should be made clear at this point that

the "fraying" of certain regions of a structure in the superposition has nothing to do with

flexibility: such segments are simply less well defined by the experimental data.

For most ß-peptides, fraying concerns the terminal residues, which are constrained by

NOEs only from one side, and the side chains, where the main cause for ill-definedness is

the difficulty in stereospecifically assigning diastereotopic CH2 groups. If stereospecific

assignment in the first CH2-group of a proteinogenic side chain is impossible, one has to

resort to a pseudo atom located between the two hydrogens and increase the upper limit.

This usually results in an undefined torsional angle around the branching sigma bond.

The quality of the structures determined by NMR can be improved at both the qualitative

and quantitative levels by making stereospecific assignments at prochiral centers and

collecting as many torsion angle restraints as possible; this also helps to resolve

ambiguities related to the assignment of some NOE cross peaks to diastereotopic protons.

41

1.19 Conclusion

A brief summary of the approach for elucidation of three dimensional solution structures

of small peptides using high resolution 2D-NMR spectroscopy and subsequent structure

calculation has been presented. Detailed studies of conformational preferences of

peptides in solution may eventually provide fundamental insights into the process of

folding and the factors responsible for it.

In fact, NOE constraints are very similarly implemented in the various programs for

NMR structure calculation. NOEs constraints are made to contribute to the total energy

of the molecule. During the annealing process NMR constraints are given much more

weight than the torsion potentials in the force field, although both contribute to the final

structure. So far, attempts to attack this difficulty by using time-averaged NMR

constraints in real, constant temperature MD simulations with explicit solvent have not

met with success. In part, this is due to technical difficulties such as a tendency towards

oscillatory behavior and insufficient sampling in time. On the other hand, this procedure

gives the actual force field much more weight than the SA procedure and therefore

depends much more critically on the correctness ofthe force field [74].

The results of the pure MD method for structure calculation were found to depend on the

type of force field used. Currently, the use of experimentally derived NMR data with MD

simulated annealing calculations is the most efficient method available for the

determination of a representative conformation of a peptide structure in solution. A

maximal number of stereospecific assignments of diastereotopic protons and torsion

angle restraints derived from coupling constants, if possible including heteronuclear ones,

leads to a qualitative and quantitative improvement in the precision of the resulting

structure.

The central barrier to progress in the investigations of short peptides in solution,

however, remains the inadequacy of the presently used methods in dealing with

multiconformational, dynamically exchanging ensembles.

42

1.20 References

[I] D. R. W. Hodgson, J. M. Sanderson, Chem. Soc. Rev. 2004,33,422.

[2] W, G. Grabarse, F. Mahlcrt, S. Shima, R. K. Thauer, U. Ermler, J. Mol. Biol.

2000, 303, 329.

[3] P. Gilleron, R. Millet, J. Domarkas, A. Farce, R. Houssin, J. P. Henichart, J. Pept.Sei. 2006,12, 140.

[4] M. Crisma, G. M. Bonora, C. Toniolo, V. Barone, E. Benedetti, B. Di Blasio, V.

Pavone, C. Pedone, A. Santini, F. Fraternali, A. Bavoso, F. Lelj, Int. J. Biol.

Macromol. 1989,77,345.

[5] I. L. Karle, P. Balaram, Biochemistry 1990, 29, 6747.

[6] C. Toniolo, E. Benedetti, Macromolecules 1991, 24, 4004.

[7] C. Toniolo, M. Crisma, F. Formaggio, E. Benedetti, A. Santini, R. Iacovino, M.

Saviano, B. Di Blasio, C. Pedone, J. Kamphuis, Biopolymers 1996, 40, 519.

[8] S. Benedetti, M. Saviano, R. Iacovino, M. Crisma, F. Formaggio, C. Toniolo, Z

Knstall 1999, 214, 160.

[9] A. K. Thakur, R. Kishore, Biopolymers 2000, 53, 447.

[10] I. L. Karle, J. L. Flippenanderson, S. Agarwalla, P. Balaram, Proc. Natl. Acad.

Sei. U. S. A. 1991, 88, 5307.

[II] H. Wenschuh, M. Beyermann, H. Haber, J. K. Seydel, E. Krause, M. Bienert, L.

A. Carpino, A. Elfaham, F. Albericio,./ Org. Chem. 1995, 60, 405.

[12] C. Toniolo, G. M. Bonora, A. Bavoso, E. Benedetti, B. Di Blasio, V. Pavone, C.

Pedone, Biopolymers 1983, 22, 205.

[13] R. N. Zuckermann, J. M. Kerr, S. B. H. Kent, W. H. Moos, J. Am. Chem. Soc.

1992,114, 10646.

[14] R. M. Wenger, Helv. Chim. Acta 1983, 66, 2672.

[15] M. Nagai, F, Kojima, H. Naganawa, M. Hamada, T. Aoyogi, T. Takeuchi, J.

Antibiot. 1997, 82.

[16] M. Llinares, C. Devin, J. Azay, G. Berge, J. A, Fehrentz, J. Martinez, Eur. J.

Med. Chem. 1997, 32, 767.

[17] S. A. Mantey, D. H. Coy, T. K. Pradhan, H. Igarashi, I. M. Rizo, L. Shen, W.

Hou, S. J. Hocart, R. T. Jensen, J. Biol. Chem. 2001, 276, 9219.

[18] S. Poenaru, J. R. Lamas, G. Folkers, J. A. L. de Castro, D. Seebach, D. Rognan, J.

Med. Chem. 1999,42,2318.

[19] M. Werder, H. Häuser, S. Abele, D. Seebach, Helv. Chim. Acta 1999, 82, 1774.

[20] R. P. Cheng, S. H. Gellman, W. F. DeGrado, Chem. Rev. 2001,101, 3219.

[21] K. Gademann, M. Ernst, D. Seebach, D. Hoyer, Helv. Chim. Acta 2000, 83, 16.

[22] K. Gademann, M. Ernst, D. Hoyer, D. Seebach, Angew. Chem. Int. Ed. 1999, 38,

1223.

[23] D. Seebach, K. Namoto, Y. R. Mahajan, P. Bindschädler, R. Sustmann, M.

Kirsch, N. S. Ryder, M. Weiss, M. Sauer, C. Roth, S. Werner, H.-D. Beer, C.

Munding, P. Walde, M. Voser, Chem. & Biodivers. 2004,1, 65.

43

[24] T. Kimmerlin, K. Namoto, D. Seebach, Hetv. Chim. Acta 2003,86, 2104.

[25] N. Umezawa, M. A. Gelman, M. C. Haigis, R. T. Raines, S. H. Gellman, 2002,

124, 368.

[26] D. Seebach, A. K. Beck, D. J. Bierbaum, Chem. & Biodivers. 2004, 7,1111.

[27] G. Cardillo, L. Gentilucci, A. Tolomelli, Mini-Rev. Med. Chem. 2006, 6, 293.

[28] H. T. Wright, Crit. Rev. Biochem. Mol. Biol. 1991, 26, 1.

[29] I. L. Karle, A. Pramanik, A. Banerjee, S. Bhattacharjya, P. Balaram, J. Am. Chem.

Soc. 1997,119, 9087.

[30] A. Banerjee, A. Pramanik, S. Bhattacharjya, P. Balaram, Biopolymers 1996, 39,

769.

[31] N. G. Bowery, Annu. Rev. Pharmacol. Toxicol. 1993, 33, 109.

[32] H. Morita, S. Nagashima, K. Takeya, H. Itokawa, Y. Iitaka, Tetrahedron 1995,

SI, 1121.

[33] C. Shih, L. S. Gossett, J. M. Gruber, C. S. Grossman, S. L. Andis, R. M. Schultz,J. F. Worzalla, T. H. Corbett, J. T. Metz, Bioorg. Med. Chem. Lett. 1999, 9, 69.

[34] P. Crews, L. V. Mânes, M. Boehler, Tetrahedron Lett. 1986, 27, 2797.

[35] R. Roers, G. L. Verdine, Tetrahedron Lett. 2001, 42, 3563.

[36] J. A. Sturman, J. Nutr. 1988,118, 1169.

[37] F. M. Rossi, E. T. Powers, R. Yoon, L. Rosenberg, J. Meinwald, Tetrahedron

1996, 52, 10279.

[38] P. J. Dunn, R. Haner, H. Rapoport, J. Org. Chem. 1990,55, 5017.

[39] G. Cardillo, C. Tomasini, Chem. Soc. Rev. 1996, 25, 117.

[40] B. Wathelet, Phytochemistry 1988, 27, 607.

[41] G. V. M. Sharma, K. R. Reddy, P. R. Krishna, A. R. Sankar, P. Jayaprakash, B.

Jagannadh, A. C. Kunwar, Angew. Chem. Int. Ed. 2004, 43, 3961.

[42] C. Palomo, M. Oiarbide, A. Landa, M. C. Gonzalez-Rego, J. M. Garcia, A.

Gonzalez, J. M. Odriozola, M. Martin-Pastor, A, Linden, J. Am. Chem. Soc. 2002,

124, 8637.

[43] A. M. Bruckner, H. W. Schmitt, U. Diederichsen, Helv. Chim. Acta 2002, 85,

3855.

[44] A. Cheguillaume, A. Salaun, S. Sinbandhit, M. Potel, P. Gall, M. Baudy-Floch, P.

Le Grel,./. Org. Chem. 2001, 66,4923.

[45] M. C. F. Monnee, M. F. Marijne, A. J. Brouwer, R. M. J. Liskamp, Tetrahedron

Lett. 2000, 41,7991.

[46] E. Schmidt, Angew. Makromol. Chem. 1970,14, 185.

[47] J. M. Fernandezsantin, S. Munozguerra, A. Rodriguezgalan, J. Aymami, J.

Lloveras, J. A. Subirana, E. Giralt, M. Ptak, Macromolecules 1987, 20, 62.

[48] F. Chen, G. Lepore, M. Goodman, Macromolecules 1974, 7, 779.

[49] H. W. Siesler, Polymer 1974,15, 146.

[50] J. Kovacs, R. Ballina, R. L. Rodin, D. Balasubr, J. Applequi, J. Am. Chem. Soc.

1965, 87, 119.

[51] E. J. Gunster, R. C. Schulz, Macromol Chem-Macro. Chem. Phys. 1980,181,

643.

44

[52] D. Seebach, M. G. Fritz, Int. J. Biol. Macromol 1999, 25, 217.

[53] D. Seebach, T. Hoffmann, F. N. M. Kuhnle, U. D. Lengweiler, Helv. Chim. Acta

1994, 77, 2007.

[54] D, Seebach, P. E. Ciceri, M. Overhand, B. Jaun, D. Rigo, L. Oberer, U. Hommel,

R. Amstutz, H. Widmer, Helv. Chim. Acta 1996, 79, 2043.

[55] D. H. Appella, L. A. Christianson, I. L. Karle, D. R. Powell, S. H. Gellman, J.

Am. Chem. Soc. 1996,118, 13071.

[56] D. Seebach, M. Brenner, M. Rueping, B. Schweizer, B, Jaun, Chem. Commun.

2001, 207.

[57] B. R. Huck, J. D. Fisk, S. H. Gellman, Org. Lett. 2000, 2, 2607.

[58] D. Seebach, S. Abele, K. Gademann, B. Jaun, Angew. Chem. Int. Ed. 1999, 38,

1595.

[59] P. I. Arvidsson, J. Frackenpohl, N. S. Ryder, B. Liechty, F. Petersen, H,

Zimmermann, G. P. Camenisch, R. Woessner, D. Seebach, ChemBioChem 2001,

10,111.

[60] A. Chakrabartty, A. J. Doig, R. L. Baldwin, Proc. Natl. Acad. Sei. U.S.A. 1993,

90, 11332.

[61] B. Forood, E. J. Feliciano, K. P. Nambiar, Proc. Natl. Acad. Sei. U.S.A. 1993, 90,

838.

[62] R. Rajan, P. Balaram, Int. J. Pept. Protein Res. 1996, 48, 328.

[63] S. Bhattacharjya, J. Venkatraman, A. Kumar, P. Balaram, J. Pept. Res. 1999, 54,

100.

[64] S. Krauthauser, L. A. Christianson, D. R. Powell, S. H. Gellman, J. Am. Chem.

Soc. 1997,119, 11719.

[65] S. Izquierdo, M. J. Kogan, T. Parella, A. G. Moglioni, V. Branchadell, E. Giralt,

R. M. Ortuno, J. Org. Chem. 2004, 69, 5093.

[66] D. H. Appella, L. A. Christianson, D. A. Klein, M. R. Richards, D. R. Powell, S.

H. Gellman, J. Am. Chem. Soc. 1999,121, 7574.

[67] P. E. Wright, Trends Biochem. Sei 1989,14, 255.

[68] A. Bundi, K. Wüthrich, Biopolymers 1979,18, 285.

[69] K. Wüthrich, J. Biol. Chem. 1990, 265, 22059.

[70] K. Pervushin, R. Riek, G. Wider, K. Wüthrich, Proc. Natl. Acad. Sei. U.S.A.

1997, 94, 12366.

[71] T. Soares, M. Christen, K. Hu, W. F. van Gunsteren, Tetrahedron 2004, 60, 7775.

[72] D. Seebach, R. I. Mathad, T. Kimmerlin, Y. R. Mahajan, P. Bindschädler, M.

Rueping, B. Jaun, C. Hilty, T. Etezady-Esfarjani, Helv. Chim. Acta 2005,88,

1969.

[73] C. D. Schwieters, J. J. Kuszewski, N. Tjandra, G. M. Clore, J. Magn. Reson.

2003,160, 65.

[74] W. F. van Gunsteren, H. J. C. Berendsen, Angew. Chem. Int. Ed. 1990, 29, 992.

45

Structural Aspects of ß-Peptides Containingoc-F and a-OH Substituted ß-Amino Acids

2.1 Introduction

Fluorine substituents have profound polar and stereoelectronic effects on the

conformation of organic molecules. The most pronounced are the gauche effect, e.g. in

1.2 difluoroethane or 7V-ß-fluoroethyl amides and the fluorine anomeric effect in a-

fluoroethers [1, 2]. The size ofthe fluorine atom is intermediate between that ofhydrogen

and oxygen. Many studies show that the size of a fluorine atom is slightly larger than that

of hydrogen. From a comparison of the van der Waals radii of fluorine (1.47 Â),

hydrogen (1.2 Â), and oxygen (1.57 Â) it appears that fluorine is a close isostere of

oxygen. Thus, the substitution of a C-H or C-OH group by C-F in enzyme substrate

analogues has been widely practiced in various areas of biooganic and medicinal

chemistry. Introduction of one of these polar groups in the oc-position of a peptide offers

a means to modify the electronic properties and reactivity without introducing steric

constraints. Such modifications may beneficially influence the biological activity and/or

conformational preferences of compounds. For instance, ß-fluorinated a-amino acids

have been extensively studied as suicide substrates for decarboxylases, trans aminases

and racemases [3]. Nevertheless, it remains difficult to predict the impact of such a

fluorine or hydroxyl substituent and this has led to some frustration in anticipating the

reactivity of fluorinated compounds [4].

Carbon bound fluorine is a poor hydrogen bonding partner1 [5]; as a consequence, the

substitution of hydrogen by fluorine in organic molecules does not significantly increase

the coordinating ability at that centre. However, stereoelectronic and polar effects of a

fluorine atom do affect the conformational preferences, especially when it is placed next

to a functional group: a) next to the C=0 group, it increases the electrophilicity b) next to

a C=C double bond, it decreases nucleophilicity.

1Ab initio calculations and X-ray studies demonstrate that C(sp3)-F is a better hydrogen bond acceptor than

C(sp2)-F [5].

46

a) fluorine-fluorine gauche effect

c)

^ yTyh -HrXrF

F

anti

H

gauche prefered

b) fluorine-amide gauche effects

H (OC)N^"^yH

H

gauche prefered

- «\

R^ R2

FVHN

Figure 16. Stereoelectronic effects of a fluorine atom: a) the preferred gauche arrangement of fluorine

atoms observed in 1,2-difluoroethane, b) the gauche relationship between C-F and N-C(O) bonds, found in

ß-fluoro ethyl amine and 4-fiuoro-L-proline, c) the preferred anti conformation of C-F-C=0 in a-

fmoroamide derivatives.

A fluoro substituent in the a-position next to a carbonyl group, as in a-fluoroamides and

a-fluoroaldehydes, favors the trans conformation around (F-C-C=0, 180°). Therefore, it

was anticipated that the introduction of a fluorine atom into ./V-substituted amides or

peptides would have a pronounced influence on dihedral angle around the bonds geminal

to fluorine {Figure 16). The X-ray determined structures of such a-fluoroamides have

revealed that the C-F bond is uniformly oriented trans to the OO group and syn to NH

in the solid state. Theoretical calculations on a-fluoroamides [6] (in the gas phase)

predicted an energy difference between syn and anti conformations of about 6-7

kcal/mol. It was therefore of interest to study the relative stability of the trans and gauche

conformations in this type of compound in solution. Corresponding results are presented

47

in this chapter and demonstrate the utility of the C-F bond as a tool to influence the

conformation of ß-peptides.

2.2 Peptides Containing Heteroatoms Directly Bound to the Backbone

The consequence of a back-bone-bound heteroatom (NH2, OH, SH, F, CI, Br, etc.) on the

structure of a-peptides or proteins is unknown, because such compounds, being acetal or

ketal derivatives, are hydrolytically unstable. Still, a-hydroxy-glycine-containing

peptides have been identified as intermediates of the enzymatic oxidative removal of C-

terminal glycines with formation of peptide amides and glyoxylic acid, a most important

posttranslational modification [7]: C-terminal amide groups are, for example, present in

mammalian peptide hormones (oxytocin, gastrin, secretin) and in toxins (mellitin, toxin

II, conotoxin, cactulein) containing from 3 to 64 amino-acid residues [8, 9], An (S)-

hydroxy-glycinc-containing tripeptide has been prepared in five steps, under carefully

controlled conditions, by Bogenstätter and Steglich [10]. The hydroxy-glycine moiety is

rather stable at slightly acidic pH, but is cleaved under basic conditions. No information

on the secondary structure of a peptide containing this elusive moiety is available.

Formulae A-E

v^-V R^V0!R3 r'kV02R3 *icooR3 n^VccH

O OHH

HalN^^

R/°

AB C D E

The same is true for halogen-substituted glycine derivatives A (R = H, X = F, CI, Br in

the Formulae A-E). Halogenation with NCS or NBS of glycine-containing peptides

generates a-halo-glycine residues, the reactivity of which has been studied extensively

by the group of Steglich and others [11, 12].

48

In contrast, ß-amino-a-hydroxy-acids and their derivatives B are stable compounds, they

occur in natural products (bestatin, amastatin, phebestin, taxol [13, 14]) and are

components of synthetic drugs. ß-Amino-a-halogenoacid derivatives C should also be

stable. Although they could cyclize to aziridines D or (with TV-acyl protective groups) to

oxazolines E under certain conditions, this process would be expected to be least likely

with fluoro-amino-acids. Thus ß-peptides containing such hetero substituted ß-amino-

acid residues should offer the unique possibility to study the influence of polar and

hydrogen-bond-forming substituents on the folding behavior and stability of secondary-

structures.

2.3 Results and Discussion

A number of fluoro- and hydroxy-substituted ß-peptides (heptapeptides 1-7) have been

synthesized in the Seebach group. The synthesis and CD spectra have been described in a

preliminary report [15].

The parent ß-heptapeptide 1 with a central ßhAla (a-Me) unit of like configuration has

been shown by CD and NMR analysis [16], as well as by molecular dynamics

simulation, to form an especially stable 3i4-helix, with the two methyl groups in lateral

positions (Figure 17b,c). Compound 1 (in MeOH solution) gives rise to an intensive

negative Cotton effect near 215 nm (s = -70,000 units), which is considered to be

characteristic of the ß-peptidic 3i4-helical secondary structure. If we would use the

intensity of this CD band as a measure for the degree of helicity of the compounds 2-7,

we would have to conclude that the monofluoro-4 and the hydroxy-derivative 7 are

"more helical", with an axial disposition of the heteroatom, than the diastereoisomers 3

and 6, with lateral heteroatoms! The difluoro-substituted ß-heptapeptide 5 would wind

up somewhere in between, and 2, the epimer of 1, would definitely not be able to form a

helix with an axial methyl group (Figure 17c).

49

a)

H,N^^N^^N

Ri- CH3 j

R2 = H (A\S)

Ri-H 2

R2 = CH3 (Ä^V)

"Y"'' o I o \ o I o Y o i o \ o

H H H « H H H

^Y^ O O \ Oj

O ""V^ O I o

H H H = H H H

-< 4 ~<^y^ O O \ O . O N^ o o \ o

H H U _/\_ U Ul UH H H p/xp H H H

-J

^Y^ o . o \ o i o N^" o|

o \ o

H2N"^^N"^^N"^^N^V^N^^^N^^N^^OH

H H H • H H H

N Y N

H

ÂHH H H

6

H H HÖH

H H H

50

b) c)

"forbidden"

'V'»v'^tt

< ^ ^f / ^^'7 portions ?

'X

^ré>> i -x"aHowed" Ü4I

o»—-^ / equatorial\V\ pysition.s

Figure 17, ß-Heptapeptides with central a-substituted ß-amino-aeid residues of /-(l, 3, 6) and u-

configuaration (2, 4, 7), the (A/)-3i4-helix of a ß-heptapeptide: a) Formulae of the heptapeptide 1-7 with

ßhAla(aMe), ßhAla(aF), ßhAla(a,aF2), ßhAla(aOH) residues, b) The 3,4-helix (of ca. 5Â pitch with

"allowed" (green) and "forbidden" (black) positions for non hydrogen atoms, and c) the (+)-scconformation around the C(a)-C(ß) bond in its ß-amino-aeid residues (nitrogen atoms: blue, oxygen atoms:

red)

In view of this result, and the surprises encountered earlier with interpretation of CD

spectra, we decided to determine the solution structures of the four a-fluoro substituted

heptapeptides 3-5 as well as that of the a-hydroxy-substituted peptides 6-7 by a

comprehensive NMR-study.

2.4 NMR Analysis of the a-FIuoro and Hydroxy-ß-Heptapeptides 3-7

The complete assignment of all 1H- and :3C resonances for each of the ß3-heptapeptides2

3-7 in CD3OH was achieved from the standard set ofNMR spectra (1-D !H, 13C, and 2D

DQF-COSY, TOCSY, HSQC and HMBC at 500 MHz and 125 MHz, acquired with

presaturation of the solvent OH-signal). The assignment of identical residues within the

sequence was based on CH(a), to CO(i) to NH(j+d correlations in the HMBC spectra.

Coupling constants 3J(NH-Hß) were determined from the 1-D *H spectra. The large

coupling constants suggested an antiperiplanar arrangement of NH and Hß atoms. The

stereoselective assignment ofthe diastereotopic a-CH2 protons were made from TOCSY

2Fluoro-, hydroxy-ß-heptapeptides (3-7,9-11) were synthesized by Dr. Françoise Gessier and Christian

Noti, ETH Zürich.

51

cross peaks. Chemical shifts, assignments, and coupling constants for ß-heptapeptides 3-

7 are compiled in {Tables 3-7).

To obtain three-dimensional structural information of these ß-heptapeptides, ROESY

spectra were measured with mixing times of 150 ms and 300 ms and the cross peak

volumes of the latter were converted into distance restraints (see exp. part). Together

with dihedral angle constraints derived from the coupling constants, the distance

restraints were used in the simulated annealing molecular dynamics calculations (SA) of

structural bundles for 3-7 with XPLOR-NIH. Each calculation produced a set of 20

structures with lowest energy without violations of experimental constraints. The

structure bundles representing the solution structures in MeOH are depicted in Figure 18.

Table 3. 'H-NMR-Chemical shifts for ß-heptapeptide 3 in MeOH

ß-amino NH H2-C(oc) H-C(ß) H-C(y) H-C(Ô) Me-C(e)acid *JHNHß Me-Cfr) Me-C(5)

H2-C(y)

2.00

1.20

1.29 1.61 0.92

1.42

1.17

1.82 0.95

1.13

1.32 1.61 0.92

1.43

ß^hVal1

ß3hAla2 8.092

ß3hLeu3 8.099

aF-ß3hAla4 8.32

ß'hVal5 8.00

ß3hAla6 7.80

ß^eu7 7.85

2.6 3.47

2.54 4.40

2.41 4.36

4.92 4.42

8.56

2.31 4.16

2.34 4.30

8.30

2.49 4.38

9.29

52

Table 4. 'H-NMR-Chemical shifts for ß-heptapeptide 4 in MeOH

ß-aminoacid

NH H2-C(a) H-C(ß)

*JHNHßH-C(Y)

Me-C(Y)

H2-C(Y)

HC-(8)

MeC(8)

Me-C(E)

ßV/al1 7.75 2.65 3.57 2.05 1.071

1.073

ßVvla2 7.99 1.22 4.57

9.20

1.22

ß3hLeu3 8.47 2.75 4.39

9.17

1.27 1.59

<xF-ß3hAla4 8.48 5.075 4.68

9.66

1.25

ß^Val5 7.76 2.48 4.28

9.17

1.81 0.95

ß^la6 7.58 2.33

2.43

4.48

8.56

1.13

ß3hLeu7 7.72 2.49 4.43

8.9

1.31

1.40

0.92

Table 5. 'H-NMR-Chemical shifts for ß-heptapeptide 5 in MeOH

ß-amino NH H2-C(a) H-C(ß) H-C(y) H-C(8) Me-C(s)acid *JHNHß Me-C(Y) Me-C(S)

H2-C(y)

ß3hVal' 7.78 2.64

2.67

3.55 2.04 1.057 0.92

ß3hAla2 7.87 2.68 4.54

9.05

1.22

ß'hLeu3 8.43 2.71 4.41

9.29

1.29

1.38

1.59 0.90

a,a-F2-

ß3hAla4

8.54 4.40

9.78

1.21

ß3hVal5 8.36 2.46

2.55

4.22

8.68

1.81 0.93/

0.96

ß^riAla6 7.64 2.33

2.42

4.45

8.56

1.24 7.64 2.33/2.4

ß'hLeu7 7.76 2.50 4.39

9.29

1.32

1.41

1.61 0.93

53

Table 6. 'H-NMR-Chemical shifts for ß-heptapeptide 6 in MeOH

ß-amino NH H2-C(a) H-C(p) H-C(y) H-C(5) Me-C(e)acid 3JHNHß Me-C(y) Me-C(5)

H2-C(y)

ß3hVal' 2.60

2.68

3.49 2.01

ß3hAla2 8.19 2.40 4.45 1.21

2.61 8.80

ß3hLeu3 8.14 2.41 4.35 1.28

8.40 1.42

cc-OH-

ß^hAla"

8.12 4.07 4.23

7.95

1.11

ß^hVal5 7.73 2.30

2.46

4.16

10.19

1.81

ß3hAla6 7.75 2.34 4.33

9.05

ß^eu7 7.82 2.45 4.36 1.31

2.51 9.17 1.43

1.60 0.92

0.95

1.61 0.91

Table 7. 'H-NMR-Chemical shifts for ß-heptapeptide 7 in McOH

ß-amino NH H2-C(a) H-C(ß) H-C(y) H-C(8) Me-C(e)acid iJHNHß Me-C(y) Me-C(8)

HrC(Y)

1.07ß^Val1 7.78 2.60

2.71

3.56 2.06

ß^Ma2 8.16 2.45

2.78

4.57

9.17

1.23

ß3hLeu3 8.38 2.39

2.65

4.39

9.29

1.28

1.39

a-OH-

ß'hAla4

7.99 4,08 4.57

9.05

1.17

ß3hVal5 7.71 2,47 4.23

9.41

1.80

ß3hAla6 7.64 2.36

2.45

4.43

8.44

1.14

ß3hLeu7 7.83 2.51 AAA 1.31

1.44

1.58 0.92

0.93

1.39 0.91

54

Figure 18. NMR-Derived structures of ß-heptapeptides 3-7, a) Residue 4 = M-ßhAla (a-F) (4); b) Residue 4= ßhAla (cc,cc-F2) (5); c) Residue 4 = «-ßhAla (a-OH) (7); d) Residue 4 = /-ßhAla (a-F) (3); e) Residue 4

= /-ßhAla (a-OH) (6). Shown are bundles of the 15 structures with the lowest energies that do not violate

experimentally derived NOE or dihedral angle constraints. Superposition of bundles by least squares fit of

a, c, d: all backbone atoms ofresidues 2-6; b, e: backbone atoms of residues 3-5.

55

Qualitative inspection of the NOE patterns exhibited by 3-7 immediately revealed that

the two compounds with unlike relative configuration in the central a-substituted ß hAla

residue (4 and 7) as well as the a-difluoro derivative 5 showed the typical pattern

observed for 3i4-helical structures. This was confirmed by the full SA-calculation which

resulted in well defined bundles of helical low energy structures with no NOE violations

for 4, 5, and 7 (Figure 18a,b,c) [17]. As observed earlier for 3i4-helical hexa- and hepta-

ß-peptides, the C-terminal residue (ßhLeu7) is less well defined by NOE constraints and

therefore appears more frayed out in the calculated bundles.

Surprisingly, the mono-fluoro derivative with like configuration (3) assumes a structure

in which the a-fluoro-substituent in the ßhAla(a-F)4 residue is again antiperiplanar to

the carbonyl C=0 bond. As a consequence, the overall structure is not a helix but consists

of two quasi-helical termini (residues 1-3 and 5-6) flanking a central turn with a 10-

membered hydrogen bonded ring between HN of residue 3 and C=0 of residue 4 {Figure

18d). From the NMR structure, the torsion angle (C-F-C(O)) in 3 and 4 is near 180°, the

C-F bond is oriented trans to the C=0 bond, and eis to NHj+i whereby NH;+1 may be

involved in hydrogen bonding with F over a distance of ca. 2.5 Â (NHj+i.,,F), close to

the value commonly observed for such interactions3.

y^4AaA*>

F"

4

F4-Ca4-C(0)4-N5

coplanar

According to NBO analysis [18], the stabilization of the trans conformation (F-C-C=0),

is mainly due to the interaction between lone pairs on fluorine and the NH o* orbital, as

well as between the C-F o and C=0 o* orbitals. Since the dipoles C-F, C=0 are opposed

See [2] for fluorine as hydrogen bonding acceptor.

56

in the trans conformation, one can expect a smaller net dipole moment compared to the

syn conformer, where the two dipoles are parallel.

The structural bundle obtained for the like configured mono-hydroxy heptapeptide 6 is

the least well defined one of the five, with the C-terminal residues ßhAla6 and ßhLeu7

being largely unrestrained through NOEs. However, the 15 structures of lowest energy

{Figure 18e) all show the a-hydroxy substituent of residue 4 nearly syw-co-planar to the

C=0 bond and a 10-membered hydrogen bonded ring between NH of residue 4 and OO

of residue 5. The TV-terminal residues are also quite well defined and have a quasi-7>u-

helical conformation.

Inspection of the structural bundles reveals that, in the ß-heptapeptides studied here, the

stcric demand on an axial substituent in the 3i4-helix might be slightly less in the fourth

residue than it would be earlier in the sequence because the axial a-substituent in residue

4 points towards the C-terminal residue. Since, under the conditions of the NMR

measurements, the terminal carboxyl group is protonated, the hydrogen bond between

HN of residue 5 and the carboxyl oxygen of residue 7 is expected to be weaker than the

hydrogen bonds to amide carbonyls earlier in the sequence. Hence, the rise of the last

half-turn of the helix may well be increased compared to the central turns and offer the

axial a-substituent in residue 4 more room. Although the scarcity of observable NOEs at

the C-terminus did not allow experimental verification of this "loosening" of the helix,

the fact that ß-heptapeptide 2 with a slightly larger methyl group in this position is no

longer able to form a helix, clearly points towards rather strict steric limits for

substituents pointing in the axial direction of a 3i4-helix.

Since the observed trend for axial polar substituents is opposite to the expectations based

on simple steric reasoning, it must be due to stereoelectronic effects, dipole-dipolc

interactions, hydrogen bonding and/or solvation factors. On the basis of X-ray

crystallography and quantum chemical calculations, O 'Hagan and coworkers concluded

that, in a-fluoro amides, the conformer with the a-fluoro substituent oriented

antipehplanar to the C=0 double bond is preferred by ca. 8 kcal/mol over the gauche

conformer (which was not a minimum on the calculated potential energy surface (PES)).

Ritter and coworkers [19] have also investigated the conformations of a-fluoroamides by

IR, NMR and theoretical methods and found that the gauche conformer becomes a true

minimum on the PES if a reaction field solvation model is included in the DFT-

calculations. While the energy of the trans conformer was calculated to be almost

57

insensitive to solvent polarity, the gauche conformer was lowered in energy with

increasing solvent dielectric constant.

2.5 A 3i4-(A/)-Helix Forming ß3-Tridecapeptide with a Central like-a-

F-ßhAla Residue

NMR Structure Determination

From the above studies we conclude that an a-F substituent in short ß-peptides tends to

be anti to the carbonyl group (see peptides 3 and 4), Peptide 4, with an unlike a-fluoro

substituent, forms a well defined 3i4-helix by disposing the F atom axially, which would

preclude helix formation for any slightly bigger a-substituent. On the other hand, the

heptapeptide 3 with like a-fluoro substituent does not form a helical structure. Thus, we

decided to prepare a longer ß-peptide with a central like ßhAla(a-F) residue to probe if it

forms a helix overcoming the energy difference (6-7 kcal/mol) to form a stable helix with

a fluorine atom in a lateral position, which is now known to destabilize the helix in short

peptides. A detailed 2D-NMR study was undertaken to obtain the solution structure of ß-

tridecapeptide4 8 in MeOH.

/ COOH j'S O H2V0 Q

HC\ o^j o r o "Y' o "v^ o r o . o \ o /o "V"' ° i ° | ° o

H H H H H H £ H H H H H H

8

The sequence was selected based on the following design principles: a) the ß3-

tridecapeptidc should be able to form a 3i4-helix with lateral side chains (i, i+3) placed in

juxtaposition b) charged side chains should ensure solubility and possibly stabilize the

helix by charge-charge interaction c) one side of the potential helix was endowed with

aliphatic side chains to allow hydrophobic interaction.

4Synthesis of ß-tridecapeptide 8 was carried out by Dr. Markus Löweneck and Boc protected a-fluoro-ß-

amino acid was prepared by Dr. François Gessier, postdoctoral researchers at ETH Zürich.

58

The spectrum in CD3OH showed good dispersion of signals allowing the complete

assignment of all the *H resonances (Table 8). The individual amino-acid spin systems

were assigned using a combination of DQF-COSY, TOCSY experiments. HSQC, HMBC

techniques were then used for sequential assignment. The 3J(NH-Hp) values were

extracted from the 1-D ]H spectrum and large coupling constants established that NH and

Hß are in antiperiplanar arrangement throughout the sequence. Qualitative inspection of

ROESY revealed NOEs between the NHj proton and Hß of residues i+2, i+3 that are

characteristic ofthe 3i4-helical structure.

Table 8. 'H-Chemical shifts for ß-1ridecapeptide 8 in MeOH

ß-amino NH H2C(a) H-C(ß) H-C(v) H-C(ö) Me-C(e)acid 3JHNHß Me-C(y) Me-C(S)

HrC(Y)

ß3hSer' 2.59/3.06

ß^eu2 8.36 2.48/2.85

ß3hGlu3 8.55 2.36/2.48

ß3hVal4 8.39 2.43/2.72

ß'hlle5 7.92 2.61

ß3hLys6 8.13 2.53/2.83

1-aF- 8.69 5.101

ß3hAla7

ß^hPhe8 8.78 2.33/2.63

ß3hAsn9 8.41 2.44/2.51

ß'hVal10 7.95 2.43/2.73

ß3hAla" 7.56 2.34

ß^Tyr12 7.73 2.29/2.59

ß3hAla13 7.85 2.52

3.71

4.64

9.46

4.34

9.10

4.32

9.10

4.28

9.32

4.46

9.32

4.57

9.68

4.63

9.1

4.80

9.12

4.30

9.4

4.41

8.95

4.57

9.54

4.40

8.50

3.83

1.34/1.53

1.87

1.8

1.6

1.63

1.24

2.75/2.89

2.63/2.77

1.74

1.17

2.61/2.79

1.13

1.63

0.96

1.0

1.44

0.92

59

A ROESY spectrum was acquired with Tm = 300 ms and integration of the cross-peaks

followed by calibration of intensities yielded inter-proton distances. The distances

together with dihedral angles around NH and Hß as derived from the coupling constants,

were used as constraints in MD-simulated annealing calculations following the X-PLOR

protocol. This calculation yielded a set of 30 structures with low violations, they are

shown in Figure 19. The side chains of the individual conformers superimpose fairly

tightly, suggesting that this ß-tridecapeptide adopts a well defined structure in methanol

solution.

Figure 19. The 3j4-helical structure of the tndecapeptide 8. Overlay of the 20 lowest energy structures

obtained by SA-calculation. Fit of backbone atoms (N, Cß, Ca and C), the MOLMOL [20] program was

used.

The bundle shows a well defined 3i4-helix over the full length of the sequence and as a

consequence the fluorine atom occupies a lateral position in the helix! The C-F bond is

not anti but nearly gauche to the OO group (dihedral around F-Ca-C=0 is ca. -90° as

60

shown in Figure 20), a local conformation which must be considered unfavorable based

on both the theoretical results and the experimental structure of the corresponding amides

and of the shorter peptide 3.

The favored local conformation around Ca-Cß is sc which would preclude formation of

an uninterrupted helix. From our earlier observation (destabilizing effect of lateral C-F)

in short peptides, the fluorine atom in a like configuration is expected to destabilize helix

formation due to stereoelectronic effects. Obviously the longer ß3-tridecapeptide folds

into a stable helix by overcoming the small endothermal contribution of this local

conformation (the conformer with the a-fluoro substituent oriented trans to the C=0

bond is expected to be preferred by ca. 7-8 kcal/mol over the gauche conformer)5 and

forcing the fluorine atom into a lateral position.

I ">o

*-

0

-. -90

Nr= c\ \V

r

•'^| 1f\;,' i ; 1

HNH

Ü

Figure 20. Conformation of the central residue /z'fe-fluoro-ßhAla in ß3-tridecapeplide 8. A Newman

projection along the Ca-C(O) bond, the C-F bond is at approximately 90° relative to the C=0 bond.

So far, no evidence has been obtained for a cooperativity effect in the folding of ß3-

peptides to helices. Experimental variable temperature NMR-studies [21], as well as free

dynamics simulations at elevated "/« silico" temperatures [22, 23] always revealed the

gradual increase of the population ofconformers with one or two turns broken when the

temperature was raised. This behavior is expected if the enthalpic contributions

(hydrogen bonds, local conformational preference for gauche at Ca-Cß, side chain

stacking) within each turn are additive along the helix. In contrast, if formation of a

5We have calculated the (gas phase) energies of (2R, 35) and flR, 37?>JV-acteyl-N'-methyl-2-fluoro-3-

methy1-propionamide as a model for the central residue of ß-peptides 3 and 4 and for CD spectra of ß-heptapeptides see [15].

61

helical turn favors the folding in the adjacent turns, cooperative effects and eventually -a

two-state "melting behavior" would be expected. In the framework ofthis interpretation,

the fact that peptide 8 is longer and can form extended helical regions on both sides of

the fluorinated residue does not explain why the latter should be forced against

stereoelectronic preferences to the local conformation corresponding to the helical

structure.

The reason must be sought in favorable interactions extending beyond a single turn, as in

specific interactions within the fluorine-containing tum that are absent in the shorter

peptide 3.

2.6 The Structural Analysis of «//-w-a-Fluoro-, a//-a,a-Difluoro-, and

a//-M~Hydroxy ß-Hexapeptides

a//-w-a-Fluoro- and a//-ot,a-Difluoro-ß-Hexapeptides

Y O I O \ O ^y^ O 0 \ o

H3N^V^N^V^N^V^N^V^N^V^N^^OH

H,N

10

The aim of this project was to study ß-peptides containing oc-fluoro substituents in each

and every residue. A detailed 2D-NMR spectroscopic analysis was undertaken to obtain

conformational information on peptides 9 in CD3OH and 10 in DMSO6, respectively.

The complete assignment of all *H resonances and the sequential assignment were

achieved by using the standard procedure {see exp. part).

The peptide was insoluble in MeOH and in other organic solvents.

62

To study the preferred conformations of the poly-heteroatom substituted backbone of

short ß-peptides, the use of 3J (H-F) or ^-(C-F) coupling constants would be useful to

gain more specific information on dihedral angles. Although the use of 3J(H-H)

couplings in conformational analysis is well established, this is not yet the case for 3J(H-

F) or 3J(C-F) because the corresponding Karplus coefficients are strongly dependent on

the individual substituents on the two carbons and are, as yet, uncalibrated for the case of

oc-fluoro-ß3-peptide units. The distance constraints were derived from the ROESY

spectrum measured with tm = 300 ms. Even though the large 3J(NH-Hß) and observation

of NH-NH NOEs were initially indicative of a helical structure, only intra-residual and

sequential NOEs were observed. Further SA annealing calculations with X-PLOR

confirmed that the conformations of the peptides 9 and 10 are of non-helical nature.

N-terminus

;ooh

C-terminus C-terminus

(/O-a-helix J f (Af)-314helix (M)-3i4helix

Figure 21. The schematic presentations of the direction of the macrodipole acting on a-helix, 14-helix and

14-helix with additional dipoles resulting from axial fluorine atoms.

In the natural a-helix, the positive-pole is at the Af-terminus, which is positively charged

(NH3+), and the negative-pole is at the negatively charged C-terminus, which cause

destabilization. In contrast, the OO dipoles in the ß-peptidic helix are oriented from the

C-towards the Af-terminus, so that this helix has an inherent capping stabilization. Thus,

63

in peptides 9 and 10, the net dipole resulting from all axial fluorines in a helical structure

(ifthe peptide would form a 3i4-helix) would point in the direction of the C-terminus ofa

helix which is analogous to the arrangement of C=0 dipoles in the a-helix. As a

consequence, the dipoles associated with the C-F bonds largely cancel the stabilizing

effect of the C=0 dipoles and may be the reason why 9 and 10 do not form a stable

helical structure (Figure 21)1,

a#-M-a-OH-ß-Hexapeptide

The introduction of a hydroxy group in the a-position might provide additional structural

effects induced through hydrogen-bonding, again without introducing any severe steric

demands. a-Hydroxy-ß-amino acids are present in a variety of biologically active

compounds, including protease inhibitors [24]. A conformational analysis of oligomers

containing w-3-amino-2-hydroxy acid residues has been reported by Grierson et al. who

examined an a-silyloxy oligomer and used NMR spectroscopy and theoretical

computations to propose an extended chain-like conformation [25]. However, this

extended chain-like structure is more likely to be driven by the bulky protection groups

on the oxygens than by the effect of the polar Coc-0 bond. Tromp and coworkers

prepared an oligomer containing /z£e-a-hydroxy-ß-amino acid residues (diastereoisomers

of the 3-amino-2-hydroxy acid, residues described below), but they did not report

conclusive evidence for its conformation [26].

"*Y^ O O \ 0 "N^ o 0 \ 0

;H;h;H;H;H-OH OH OH OH OH OH

11

Previously, the NMR-solution structure of poly-w-a-hydroxy-ß-hexapeptide 11 with

protected termini has been investigated in our group by Gademann. Simulated annealing

calculations with the commercial program QUANTA using NMR-derivcd distance- and

7In the case of peptides in MeOH, the carboxyl group is likely to be protonated.

64

dihedral constraints resulted in a well defined bundle corresponding to a 2rhelix, which

was reported as the fourth type of secondary structure adopted by ß-peptides [27].

The structure was characterized as a right-handed helix with approximately two residues

per turn. The amide C=0 groups are aligned in a sequential up/down arrangement twisted

out from the helix axis. Also, the NMR structure calculation suggested a preferred

arrangement of the carbonyl group and the a-hydroxy group in a syn periplanar

arrangement.

Van Gunsteren and coworkers carried out an unrestrained MD simulation on the same

protected molecule using the thermodynamically calibrated GROMOS force field with

explicit solvent methanol at two different temperatures (298 K and 340 K) [28], In the

298 K trajectory, eight membered hydrogen-bonded rings, characteristic for a 2g-helix,

appeared but only to a very low percentage (< 4 %). At higher temperature (340 K) the

occurrence of 12-membered hydrogen-bonded rings hinted at the formation of a (P)-

2.5i2-helix. In fact, both, the time average of the MD trajectory at 340 K and the NMR-

derived structure calculated with QUANTA were consistent with the observed NOE

distances and J-couplings.

A later careful recalculation using XPLOR-NIH and the original NMR data confirmed

this ambiguity although the originally proposed 2g-structure had slightly less violations

than the 2.5i2-structure proposed by van Gunsteren etal.

In view ofthese contradictory results, it was of interest to clarify the structural preference

of an unprotected poly-w-a-hydroxy-ß-hexapeptide8, expecting it to form a rather stable

secondary structure in the absence of terminal protecting groups (the fact that the 14-

helix of ß-peptides is more stable with unprotected than with protected termini is known

[29]).

8The deprotection of TV-Boc and C-Benzyl groups of hexapeptide 11 was carried out with the help of

Thierry Kimmerlin.

65

The complete assignment of all lH resonances was achieved through a combination of

2D-NMR experiments (DQF-COSY, TOCSY, and HMBC). The 'H-NMR spectrum in

CD3OH showed limited dispersion of the signals from the backbone protons unlike the

protected form of the peptide. Nevertheless, the coupling constants 3J(NH-Hp) and 3J

(Hß-Ha) could be extracted from the *H NMR spectrum. The measured values for

3J(NH-Hß) of 9.0 to 10.2 Hz are in agreement with an antiperiplcmar arrangement ofNH

and Hß atoms. The 3J(Hß-Ha) coupling constants were in the range of 2.8 to 3.0 Hz. For

the conformational analysis, the distance restraints were obtained from integration of

ROESY cross-peaks and calibrated with the known distances. The presence of one weak

NH; to NHi+i cross-peak and non-sequential NOEs suggests the presence of a secondary

structure. The structure was calculated according to the simulated annealing protocol of

X-PLOR using distance restraints, together with six dihedral angle restraints around (NH-

Hß) derived from the J values by using Karplus equations. The calculation produced 30

low energy structures without any NOE or dihedral violations. The structural bundle

consists oftwo families of conformers with helical character only at the TV-terminus with

anti and syn arrangement of (C=0, OH groups), with no detectable specific hydrogen

bonds.

66

2.7 Conclusion

The current study showed that a-fluoro ß3-amino acid residues can be used as a tool to

control the conformation of peptides and compounds of medicinal importance, exploiting

the slightly larger steric impact of fluorine over hydrogen and the stereoelectronic

interactions ofthe C-X (X = F, OH) ) bond when placed next to amide carbonyls.

In conclusion, we have found a preference ofthe C-F bond for an anti orientation relative

to the C=0 group in the a,a-difluoro-ß-heptapeptide 5 and the a-F/OH ß-heptapeptides

4 and 7 with (2R.3S) or unlike configuration of the central amino-acid residue. Both

peptides form a 3i4-helix with an axial F or OH substituent, while the cpimers 3 and 6

with like configuration of these residues are not helical over the full length of the

molecule. This shows that a single F or OH group in the "wrong" position of a ß-

heptapeptide is able to prevent the formation of the full helix. There must be a helix-

stabilizing effect of an axial F or OH heterosubstituent and/or a destabilizing effect of a

lateral one, where as with two geminal fluoro groups the helix stabilizing influence of

the axial F prevails (see 5). The NMR structures also confirm that both substituents, F

and OH, are small enough to occupy a "forbidden" position on the 3]4-helix, where a CH3

group "sterically destroys" the helix (see 2).

Surprisingly, ß-tridecapeptide 8 is able to form a full helix with fluorine in like

configuration despite the lateral position of the F-atom in this 3i4-helix. The poly-w-

fluoro and poly-w-hydroxy ß-hexapeptides do not fold into any preferred secondary

structure, possibly due to the repulsion between net dipole from the polar Co>X bonds

and the terminal charge of the helix.

67

2.8 References

[1] N. C. Craig, A. Chen, K. H. Suh, S. Klee, G. C. Mellau, B. P. Winncwisser, M.

Winnewisser, J. Am. Chem. Soc. 1997,119, 4789.

[2] H. Senderowitz, P. Aped, B. Fuchs, Tetrahedron 1993, 49, 3879.

[3] D. O'Hagan, H. S. Rzepa, Chem. Commun. 1997, 645.

[4] D. Sccbach, Angew. Chem. Int. Ed. 1990, 1320.

[5] J. D. Dunitz, R. Taylor, Chem. Eur. J. 1997, 3, 89.

[6] J. W. Banks, A. S. Batsanov, J. A. K. Howard, D. O'Hagan, H. S. Rzepa, S.

Martin-Santamaria, Chem. Soc. Perk. Trans. 2 Phys. Org. Chem. 1999, 2409.

[7] G. Krcil, 'Methods in Enzymology;Posttranslational modifications, Part A', Vol.

106, Academic Press, Orlando, 1984.

[8] M. 0. Dayhoff, Atlas ofProtein Sequence and Structure', Vol. 5, Silver Spring,

1972.

[91 W. R. Gray, A. Lugue, B. M. Olivera, J. Barrette, L. J. Cruz, J. Biol. Chem. 1981,

4734.

[10] M. Bogenstätter, W. Steglich, Tetrahedron 1997, 7267.

[111 R. M. Williams, D. Zhai, P. J. Sinclair, J. Org. Chem. 1986, 51, 5021.

[12] 2. Ernstein, D. Ben-Ishai, Tetrahedron 1977, 881.

[13] H. Umezawa, T. Aoyagi, H. Suda, M. Hamada, T. Takeuchi, J. Antibiot. 1976,

29, 97.

[14] M. Nagai, F. Kojima, H. Naganawa, M. Hamada, T. Aoyogi, T. Takeuchi, J.

Antibiot. 1997, 82.

[15] F. Gessier, C. Noti, M. Rueping, D. Seebach, Helv. Chim. Acta 2003, 1862.

[16] D. Seebach, P. E. Ciceri, M. Overhand, B. Jaun, D. Rigo, L. Oberer, U. Hommel,

R Amstutz, H. Widmer, Helv. Chim. Acta 1996, 79, 2043.

[17] R I. Mathad, F. Gessier, D. Seebach, B. Jaun, Helv. Chim. Acta 2005, 88, 266.

[18] A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev. 1988, 88, 899.

[19] C. F. Tormena, N. S. Amadeu, R. Ritter, R J. Abraham, J. Chem. Soc. Perkin

Trans. 2 2002,773.

[20] R. Koradi, M. Billeter, K. Wüthrich, J. Mol. Graph. 1996,14,51.

[21] K. Gademann, B. Jaun, D. Seebach, R. Perozzo, L. Scapozza, G. Folkers, Helv.

Chim. Ada 1999, 82, 1.

68

[22] X. Daura, W. F. van Gunsteren, D. Rigo, B. Jaun, D. Seebach, Chem. Eur. J.

1997, 1410.

[23] X. Daura, B. Jaun, D. Seebach, W. F. van Gunsteren, A. E. Mark, J. Mol Biol.

1998, 280, 925.

[24] G. Cardillo, C. Tomasini, Chem. Soc. Rev. 1996, 25, 117.

[25] I. A. Motorina, C. Huel, E. Quiniou, J. Mispelter, E. Adjadj, D. S. Grierson, J.

Am. Chem. Soc. 2001,123, 8.

[26] R. A. Tromp, M. van der Hoeven, A. Amore, J. Brussec, M. Overhand, G. A. van

der Marel, A. van der Gen, Tetrahedron: Asymmetry 2001,12, 1109.

[27] K. Gademann, A. Häne, M. Rueping, B. Jaun, D, Seebach, Angew. Chem. Int. Ed.

2003, 42, 1534.

[28] A. Glättli, W. F. van Gunsteren, Angew. Chem. Int. Ed. 2004, 43, 6312.

[29] M, Rueping, J. V. Schreiber, G. Lelais, B. Jaun, D. Seebach, Helv. Chim. Acta

2002, 85, 2577.

69

Synthesis and Structural Studies on

Peptides Containing a-, ß- and y-AminoAcids: "Mixed"-Helices and Hairpin

Turns

3.1 Introduction

The folding of peptides consisting of ß-amino acids has stimulated considerable recent

interest in the conformation of designed peptides containing the higher homologues of

naturally occurring a-amino acids. Although novel folded structures have been well

characterized in oligomeric sequences of ß-amino acids, only limited information is

available on the accommodation of ß- and y-amino acid residues into the classical

secondary structural motifs formed by a-amino acid sequences. Thus, incorporation of ß-

amino acids into helical and ß-sheet structures is of considerable importance in the

design of analogues of biologically active peptides.

Recently, there has been a growing interest in peptides containing a- and ß-aminoaeid

residues. Such "mixed" peptides can also be proteolytically stable [1], and some tested

negative for hemolytic and antimicrobial activities [2J. a/ß-Mixed peptides were found to

fold to new types of helical secondary structures held together by hydrogen-bonded rings

containing 9 through 15 atoms [3, 4], There are numerous more or less rational ways, in

which sequences of a- and ß-amino acids can be arranged to form "mixed" peptides of

types A-E. Balaram and coworkers have shown that an undecapeptide (of type A),

containing three strongly helix-inducing Aib units and two ß-amino-aeid residues, forms

a helix in the solid state and in solution, with 13-membered H-bonded rings in the

terminal sections, and a 15- and 14-membered ring in the a-ß-ß and in the a-a-ß or ß-a-

a segments, respectively [5], The most favored objects of investigation are a/ß-mixed

peptides with alternating a- and ß-aminoaeid residues (type C).

70

A (a).-(ß)b-(y)0 B (ß).-(a)b-(ß)e C (a-ß). or (ß-a).

D (a-a-ß)aor(ß-ß-a)a E (a).-(ß)b-(a)0-(ß)d-(a)e

With simple proteinogenic side-chains, this type of mixed peptide was reported not to

fold to helices that are stable on the NMR time scale in solution [6, 7], For carbohydrate-

derived ß-amino acids or conformationally restricted cyclic a derivatives, a number of

helical structures were found by X-ray and NMR analysis, e.g. a 13-hclix of oligomers

containing Ala and cw-2-amino-cyclopropane-carboxylic acid and "mixed" helices

comprised of hydrogen-bonded rings of different sizes; a 9/11-helix of a C-linked carbo

ß-amino acid Caa (with Ala); an 11-, a 14/15-, a 10/11/11- and a 11/11/12-helix with

diads or triads (type D) of /rara-2-amino-cyclopcntane-carboxylic acid (ACPC) or trans-

4-amino-pyrrolidine-carboxylic acid (APC) and proteinogenic (Xaa) or non-

proteinogcnic, a-branched {cf. Aib, Cyp, a-MePhe) a-amino acids [8-10],

So far, hairpin-forming mixed a/ß-peptides have only been described in two papers by

Balaram and coworkers. They contain the conventional D-Pro-Gly cc-peptidic turn motif

and carry ß-amino acid residues (ßhGly or ßhPhe) on opposite sides of the antiparallel

sheet part of the hair-pins (type E) [11], Furthermore, tetrapeptides of the sequence ß-a-

ß-ß (tyPe B)> which are agonists of somatostatin1 [12], have been described.

See Chapter 6 for ß-tetra-peptides as somatostatin and RGD analogues and their conformational studies.

71

NH,

O

SMe

o r" o

rnc rv

O0 ( 0

12 [(ßJa)3ß

0 ^^ 0

3„i R3l

N ^^ OHH

NH,

H3N^^ ^N

ÖrY

Y0 ^^ 0 V

13 [(ß3a)3ß3«vk3i

COOH

NH,

XH2N >^ N

-"Y nV^SVrY*1"SMe

2„\ ft2i14 [(ß^)3ßÖ

NH,

HO

\=0/ O / 0 f 0 I 0

VOH

15 [(aßV]

1 1cAoh nA,

0^ OH

0 : 0 \ O \ 0

y

~Ki

° l^~

16 [(aßVl

.OH

H,N

oV=o y—

o ; o"

O^^ 17 [(aß3)4a

OH

0 / O

OH

NH,

72

V P=

H H°

I°

n^rYrv^VayH H

18 [(Ttaß^a]NH,

rVxi

19 [(|l3Aib)3|i3]

HJ*

^Y* sV^/sY;0 0

2a;u\ fi2i20 [(ß2Aib)3ß

P0 \ 0

OH

N ^^ OH

21 [(Aibß3)4]NH,

NH,

H,N

n m

0 \ 0

o NV' o

0\

"0

-OH

22 [(Aibß3)4]OH

73

H2N

rYXn

H3N

^Y^

rvs

COOH

23 [ß3ocß3Aibß3ocß3]

NH2

JUAo

24 [ß;2a(2Meß3)a(2Meßj)aß

o -

o

^qas\n'-~Vy'y

t! H

, .

Ö.

-^H

25 [ß2aß2ß3aß3]

To better understand the preferred conformations of p-amino acids in mixed sequences,

we undertook the synthesis and NMR conformational study of a number of "mixed" oc/ß-

, a/y- and ß/y-peptides 12-25 and 26-28 (below) consisting exclusively of components

with proteinogenic side chains and of the natural, although non-proteinogenic (non-

ribosomal), helix-inducing amino acid residue Aib (for a/ß mixed peptides). We found

two helical and two hairpin structures among them by NMR [13].

74

3.2 Synthesis of a/ß-, oc/y- and ß/y-Mixed Peptides

Mixed peptides were synthesized2 by coupling Fmoc-amino acids on solid support (Wang

resin) with the exception of the last residue for which a Boc-protected amino acid was

used to avoid possible problems with removing the Fmoc group at the end of the

synthesis. The first Fmoc-amino acid was anchored to the resin by the MSNT/Melm

method [14] (Scheme 2).

Wang-linker resin

R1 O

H

DMelm, MSNT, CH2C12

2) Ac20, DMAP, CH2C12

3) 20% piperidine in DMF

or DBU/piperidine/DMF 1:1:48

4) building block, HATU, DIPEA

repeal steps 3) and 4)

Fmoc-N^kA0^Wang

R/ O R1 0

H-N4\^N^\^0H -

ITA/TIS/H20

95:2.5:2.5 R2 O

Fmoc. JL J\ Jn. JL ^Wang

Scheme 2. General procedure for the solid-phase synthesis of ß-peptides on a Wang resin. "Mixed a/ß.-a/y- and ß/y-peptides were synthesized accordingly.

The loading was determined by treating the resin with 20% piperidine in DMF and

measuring the absorbance of the dibenzofulvene-piperidine adduct at 290 nm. A typical

loading efficiency of 70-80% was obtained for all peptides. The free hydroxyl groups on

the resin were capped using acetic anhydride and a catalytic amount ofDMAP. The chain

elongation was carried out as shown in Scheme 2.

Synthesis of a/ß-nona, a/y- and ß/y-mixed peptides was carried out in collaboration with Dr. Oliver

Flöegel. All o/ß-hepta-peptides were synthesized by Dr. Sebesta Rodavan and a/ß-hexa-peptide 25 was

prepared by Dr. Michael Limbach.

75

The Fmoc-protecting group3 was removed using a combination of 20% piperidine in

DMF and/or DBU/piperidine/DMF 1:1:48. The use of DBU was essential in order to

achieve efficient Fmoc deprotcction when the growing peptide chain contained more than

six p-amino acid residues. Fmoc- and Boc-ß-homoamino acids (1.5-3 equiv. with respect

to the resin loading) were activated using HATU/DIPEA and coupling reactions were

performed in DMF at room temperature for 2-4 h. Completion of the coupling was

confirmed by the TNBS test [15]. The resins were then treated with TFA/H20/TIS

(95:2.5:2.5) to give crude free peptides.

3.3 CD Spectroscopy

The CD-spectra of the a/p-mixed peptides were measured in methanol. The CD spectra

of those peptides 15-18 for which we have not found an NMR solution structure and

peptides 21-22 containing Aib residues which were found to have helical or turn

structures by NMR, are shown in Figure 22. A superficial glance at these spectra does

not really show a general trend or a fundamental difference between the CD spectra of

those compounds for which a secondary structure was detected (14, 19, 22, 25) and

76

JIMS* i

? i /«S US««»-! /^ * /S? ! J* i i

X

/

\V.

fXtXI v

«i-

-1 !i.">tX.<i> .

i—,

IM

15 16

JÏSKIfl -i .'

'3vCvC -î <

JS« i*î î:« «3 ?50 &«0 3«0

i«Wtf

% »fi

ISO ?:W

S,...

V,

17 18

i ,<'

sw« -i J

i i

0 j

V

A

o «; ^

i ./

......»».»..•••.tt.t...-.!!)iK

üä 2ïf: Jrftt ;«.• ï.%:

21 22

Figure 22. Non normalized CD-spectra of a/ß-mixed peptides 15-18 and 21-22 measured at 20° in MeOH

(c = 0.2 mM).

77

3.4 The NMR-Solution Structure of Mixed a/ß Peptides

The conformational properties of peptides derived from a-amino acid residues arc well

established. Our goal was to elucidate the conformational preferences of mixed a/ß-

peptides in the hope of detecting new secondary structures in these novel hybrid peptides.

A wide variety of backbone substitution patterns might be useful for a rational design of

mixed helix structures.

The conformational behavior of all mixed peptides in CD3OH solution was investigated

with 2D-NMR (DQF-COSY, TOCSY, HMBC, ROESY experiments) by analyzing and

interpreting the data with slow-cooling-simulated annealing calculations using XPLOR-

NIH. The mixed peptides 12-13,16-18, 23 and 24 were designed to contain ß-homo- and

a-amino acids with proteinogenic side chains in alternating positions, except in peptide

18, D-a-amino acids have been used because helical model consideration suggested that

D-amino acid side chain could be accommodated better then the L-amino acid side chain

with ß-/y-amino acid homologues derived from L-amino acids. Analysis of their ROESY

spectra showed no NOE typical of one of the known secondary structures for a or ß-

peptides. The 3J(NH-aH) values were about (6-7 Hz) close to the random coil values.

Though a large number of sequential (i, i+1) NOEs was present in all of these peptides,

crucial non-sequential (i, i+2 or i + 3) NOEs were scarce suggesting poor propensity for

helix formation. More detailed data on the backbone conformations were deduced from

distance restrained Simulated Annealing calculations which confirmed that all the above

mentioned peptides showed no regular secondary structure. A similar result had been

observed for a mixed (ß-ß-a)3 peptide6 in MeOH. The chemical shifts and 3J-coupling

constants for mixed-peptides7 forming secondary structures have been compiled in

Tables 9-12.

The heptapeptide 23 was designed to have a/ß residues with a central Aib-unit as helix

inducer and in heptamcr 24, ß2,3-substituted homoamino acids were introduced in

positions 3 and 4 expecting them to stabilize the resulting structure due to the restricted

torsional angles in the backbone. NMR-analysis showed that neither peptide 23 nor 24,

6An attempt was made to detect helical structure in solution for a (ß-ß-cx)3 peptide by Magnus Rucping

during his PhD thesis (Diss. Nr, 14677, ETH-Zürich).See exp. part and [13J for the characterization of remaining mixed-peptides.

78

formed any regular structure, but the structural bundles contained bends around the ß2'3-

homoamino acid and the Aib, respectively, without any detectable hydrogen bonds.

Table 9. 'H-Chemical shifts for o/ß-heptapeptide 14 in CD3OH

ß-amiiioacid

NH Hj-C(a)3JHNHa

H-C(ß) H-C(Y)Me-C(Y)

H2-C(Y)

H-C(8)

Mc-C(5)

Me-C(e)

ß2hLeu' 2.21 3.77 n/a 1.57 0.99

alle2

ß2hMet3

8.3

8.5

4.2

6.7

2.68

1.85

3.12/3.60

1.55

1.90

1.10

2.46

aLys4 8.27 4.32

7.0

1.68/1.80 1.45 1.27/1.64 2.87

ß2hVal5 8.21 2.21 3.0/3.77 1.83 0.92/1.00

aAla6

ß2hPhe7

8.29

8.19

4.33

6.6

2.91

1.36

3.26/3.6 2.85

Table 10. 'H-Chemical shifts for a/ß-heptapeptide 19 in CD3OH

ß-amino NH H2-C(a) H-C(ß) H-C(y) H-C(ô) Me-C(s)acid 3JHNHß Me-C(y) Me-C(8)

HrC(Y)

ß'hLeu1 2.60 3.58 1.51 1.7 0 93

Aib2 8.34 1.46/149

ß^Val3 7.65 2.34/2.41 3.91 1.77 0.90/0.94

9.4

Aib4 8.08 1.43/1.44

ß3hAla5 7.63 2.27/2.44 4.20 1.18

8.8

Aib6 8.06 1.33/1.34

ß'hPhe7 7.70 2.51 4.45 2.86

8.8

79

Table 11. 'H-Chcmical shifts for a/ß-octapeptide 22 in CD3OH

ß-amino NH HrC(a) H-C(ß) H-C(y) H-C(Ô) Me-C(s)acid 3JHNHß Me-C(y) Me-C(ö)

HrC(T)

2.66/2.8

Aib1 1.35/1.53

ß^Tyr2 7.8 2.42/2.5 4.4

8.8

Aib3 8.14 1.48

9.1

ßUys« 7.6 2.35 4.1

n/a

Aib5 8.12 1.4

ß3hAsp6 7.85 2.4/2.53 4.5

8.8

Aib7 8.0 1.44

ß3hVal8 7.59 2.52 4.0

9.0

•61 1.58 1.68

2.6

1.8 0.91

Table 12. 'll-Chemical shifts for a/ß-hexapeptide 25 in CD3OH

ß-amino NH H2-C(a) H-C(ß) H-C(y) H-C(5) Me-C(E)acid 3JHNHa 3JHNHß Me-C(y)

H2-C(y)

Mc-C(ö)

ß2hPhe' n/a 3.0/3.2 2.77/2.9

aHis2 8.5 4.65

7.6

2.9/3.0

ß2hLeu3 8.1 2.57

6.0

3.23/3.43 1.23/1.55 1.57 0.90

ß3hVal4 7.84 2.47/2.56 4.0

8.7

1.81 0.92

alle5 8.0 4.1

8.3

1.7 1.12

ß'hTyr6 7.9 2.39/2.47 4.39

8.5

2.68/2.74

80

a) b)

Figure 23 Overlay of lowest energy NMR-solution structures of the a,ß-mixed peptides 14 (a) and 25 (b)

The hexapeptide 25 consists of ß3-homo, ß2-homo and a-amino acids with a central ß2/ß3

segment8 which is known to induce a turn stabilized by H-bondmg, whereas heptapeptide

14 is composed of ß2-homoammo and a-amino acids in alternating positions. From NOE

derived distance restraints combined with simulated annealing, we deduced the presence

of a well defined turn induced by the ß2/ß3 unit with a 10-membered H-bonded ring in

peptide 25 Also, for the mixed peptide 14 the NOEs across the strands provided

evidence for a 9-membered tum involving the ß2hMet and Lys units (Figure 23)

For mixed peptides 19 and 22, the integration ofNOE cross-peak volumes from the 300-

ms ROESY spectrum yielded distance restraints that were used in simulated annealing

calculations Each calculation started with an extended conformer, and the resulting

bundles without NOE or dihedral-angle constraint violations are shown in Figure 24-25

They reveal that both peptides 19 and 22 fold into a right handed 14/15-hehx, involving

ß-homoamino acid and Aib-residues in 14 and 15-membered H-bonded rings

respectively

8The use of ß2/ß3umt as a turn inducing element has been demonstrated previously m the Seebach group

9See Exp Part for the detailed Simulated Anneahng protocol

81

Figure 24 The NMR-solution structures of the a/ß-mixed peptide 19 in MeOH represented in bundles

obtained by bimulated annealing calculations

Figure 25 The NMR-solution structures of the a/ß-mixed peptide 22 in MeOH represented m bundles

obtained by simulated annealing calculations

It should be noted that the diastereospecific assignment of the geminal methyl groups of

Aib residues was crucial for the structure calculation The diastereotopic methyl groups

of Aib were assigned through an HMBC experiment in which one of the Me groups

strongly coupled to its own NH, which results in a higher intensity cross peak and allows

82

assignment as (Me)RS. We attribute the 14/15-helix formation in these peptides mainly to

the influence of the Aib residue since it has been established that Aib, as well as many

other dialkyl amino acids, are strong promoters of helical local conformation due to

restricted torsional angles.

3.5 Conformational Analysis of Mixed a/y- and ß/y-Peptides

Based on our findings and studies of others on hybrid peptides consisting of a- and ß-

amino acid monomers, we expanded our investigation to two new classes of mixed

peptides [16-18]. One is composed of a- and y-amino acid10 [19] constituents, and the

other consists of ß- and y-amino acid constituents with proteinogenic side chains in

alternating positions, expecting them to form a new class of structures. The synthesis of

mixed peptides 26-28 was successfully achieved via the SPPS procedure11.

H0^*0

OH

O V^ H O /\H o v^

H o y^H

26 [(<xy4)4a]

NH2

I H° ( H

°I H °f^H ° (~~*

o >k Ho k/ R

o /k H ö xk Ho

27 [(ay4)4a

HO,Q V h h °

H,N "^ NH

n n h

O v O

NH2 28 KßVteP3

10

y-amino acids with side chains in the 2-, 3-, and 4-positions, are shown to form (M)-2.6m helices in the

crystal state and in MeOH solution.1 'See experimental section for the detailed protocol.

83

The sequence-specific assignment of all resonances in the 'H-NMR spectra was achieved

by DQF-COSY, TOCSY, HMBC and ROESY experiments, as described previously for

p-peptides. For chemical shifts and coupling constants of peptides 26-28, see exp. part.

The NH protons of ß, y- amino acid residues show coupling constants 3J(NH-Hß and

NH-Hy) in the range of 9-10 Hz, which correspond to an antiperiplanar arrangement of

NH to Hß and Hy in ß and y-amino acids, respectively. The 3J(NH-Ha) values of the a-

amino acid residues were in the range of 6-7 Hz, which might be the result of averaging

of several local conformations. Due to the spectral overlap, it was not possible to get

individual chemical shifts and coupling constants for many of the Ha and Hß protons of

the y-amino acid residues. In order to establish the three-dimensional solution structure of

these mixed peptides, ROESY spectra were acquired, and inter-proton distances were

deduced from ROESY cross peak volumes by calibration with known inter-proton

distances. Qualitative analysis of NOEs for a/y-mixed peptides showed a number of (i,

i+2) and two (NH, NH) NOE cross peaks which, at first sight, pointed towards a helical

structure. However these NOEs were absent in the case of the ß/y-peptide. Furthermore,

slow-cooling simulated annealing calculations using the NMR-derived dihedral angles

and distance constraints resulted in a set of different conformational clusters, all of them

non-helical. Homogeneous peptides made from ß-amino acids that are synthetically

derived from L-amino acids have been shown to form left-handed helices, whereas

corresponding y-peptides form right-handed helices. In mixed ß/y-peptides such as those

studied here, the two propensities oppose each other and this may be the major factor

preventing the formation ofwell defined preferred secondary structure.

Therefore, it would be logical to study mixed ß/y-peptides in which either one end of the

amino acids must have opposite absolute configuration (derived from D-amino acids). In

this case, the inherent local conformational preferences of the two building blocks are

expected to enforce rather than oppose each other.

84

3.6 Conclusion

The synthesis of mixed peptides containing a/ß-, a/y- and ß/y-amino acid residues in

alternating positions was accomplished via the SPPS method. The series of mixed-

peptides was investigated by 2D-NMR spectroscopy in order to determine their

conformational preference and search for stable secondary structures. The results show

that mixed a/ß-peptides 19 and 22 with Aib residues in alternating positions fold into a

right handed 14/15-helix. Thus introduction of ß-amino acids lead to secondary structure,

i.e. closely related to the a-helix provided the a-amino acid is a helix inducing a,a-gem-

dialkyl unit.

For mixed peptides 12-13, 16-18, 23-24, 26-27 and 28 unstructured or partially folded

conformations with several different families of conformers that rapidly interchange were

found.

Helices and turns described herein are more due to backbone conformational preferences

than to hydrogen bonding. Mixed peptides 14 (containing a central ß2/a segment) and 25

(with a central ß2/ß3 segment) fold into hairpin-turn structures stabilized by 9 and 10-

membered hydrogen bonded rings, suggesting that a-amino acids can be inserted in the

i+2 position of the turn in combination with ß2-amino acids in (i+1) position to stabilize

turn like structures.

These results confirm the earlier proposed tum like of structures for mixed a/ß3-

tctrapeptides, including somatostatin agonists which achieved high binding affinities for

the SRIF receptor. They suggest that an a/ß3-segment induces a turn structure similar to

that ofthe ß2/a-segment mainly due to backbone preferences.

The use of ß-amino acid residues in peptide chemistry can provide an interesting design

alternative to conventional a-peptides. Sequences containing a- and ß-amino acid

residues should be useful in generating new polypeptide scaffolds for arraying functional

side chains necessary for desired biological activity. The designed mixed peptides

provide a direct route to access proteolytically stable analogues retaining the overall

folding properties observed in natural proteins. The structure space of mixed peptides

appears to be considerably greater than that of their a-amino acid counterparts and may

be useful for the construction of 'mixed' proteins with higher ordered structures of

enhanced stability and modified functions.

85

3.7 References

[I] D. F. Hook, P. Bindschädler, Y. R. Mahajan, R. Sebesta, P. Kast, D. Seebach,

Chem. & Biodivers. 2005, 2, 591.

[2] P. I. Arvidsson, N. S. Ryder, H. M. Weiss, D. F. Hook, J. Escalante, D. Seebach,

Chem. & Biodivers. 2005, 2, 401.

[3] A. Hayen, M. A. Schmitt, F. N. Ngassa, K. A. Thomasson, S. H. Gcllman,

Angew. Chem. Int. Ed. 2004, 43, 505.

[4] K. Ananda, P. G. Vasudev, A. Sengupta, K. M. P. Raja, N. Shamala, P. Balaram,

J. Am. Chem. Soc. 2005,127, 16668.

[5] R. S. Roy, I. L. Karle, S. Raghothama, P. Balaram, Proc. Nail Acad. Sei. U.S.A.

2004,101, 16478.

[6] F. Rossi, E. Bucci, C. Isernia, M. Saviano, R. Iacovino, A. Romanelli, P. Di

Lcllo, M. Grimaldi, D. Montesarchio, L. De Napoli, G Piccialli, E. Benedetti,

Bwpolymers 2000, 53, 140.

[7] M. A. Schmitt, S. H. Choi, I. A. Guzei, S. H. Gellman, J. Am. Chem. Soc. 2006,

128, 4538.

[8] M. A. Schmitt, S. H. Choi, I. A. Guzei, S. H. Gellman, J. Am. Chem. Soc. 2005,

127, 13130.

[9] S. De Pol, C. Zorn, C. D. Klein, O. Zerbe, O. Reiser, Angew. Chem. Int. Ed. 2004,

43,511.

[10] G. V. M. Sharma, P. Nagcndar, P. Jayaprakash, P. R. Krishna, K. V. S.

Ramakrishna, A. C. Kunwar, Angew. Chem. Int. Ed. 2005, 44, 5878.

[II] I. L. Karle, H. N. Gopi, P. Balaram, Proc. Natl. Acad. Sei. U.S.A. 2001, 98, 3716.

[12] C. Nunn, M. Rueping, D. Langenegger, E. Schuepbach, T. Kimmerlin, P. Micuch,

K. Hurth, D. Seebach, D. Hoyer, Naunyn-Schmiedebergs Arch. Pharmacol 2003,

367, 95.

[13] D. Seebach, B. Jaun, R. Sebesta, R I. Mathad, O. Flögel, M. Limbach, H. Seilner,

S. Cottens, Helv. Chim. Acta 2006, 89, 1801.

[14] W. C. Chan., P. D. White, 'Fmoc SolidPhase Peptide Synthesis; A Practical

Approach', Vol. 222, Oxford University Press, Oxford, 2000.

[15] W. S. Hancock, J. E. Battersby, Anal. Biochem. 1976, 71, 260.

86

[16] A. Romanelli, I. Garclla, V. Menchise, R. Iacovino, M. Saviano, D.

Montesarchio, C. Didierjean, P. Di Lello, F. Rossi, E. Benedetti, J. Pept. Sei.

2001, 7, 15.

[17] G. V. M, Shamia, K. R. Reddy, P. R. Krishna, A. R. Sankar, P. Jayaprakash, B.

Jagannadh, A. C. Kunwar, Angew, Chem. Int. Ed. 2004, 43, 3961.

[18] R. S. Roy, H. N. Gopi, S. Raghothama, R. D. Gilardi, I. L. Karle, P. Balaram,

Biopolymers 2005, 80, 787.

[19] D. Seebach, M. Brenner, M. Rueping, B. Schweizer, B. Jaun, Chem. Commun.

2001, 207.

87

Solution Structures of a ß3-IcosapeptideContaining Homologues of all 20 Common

Proteinogenic Amino Acids and

of an a-Heptapeptide with Central Aib

4.1 Introduction

Helices are the most abundant secondary structural elements in proteins. a-Helices

constitute about a third of the residues in globular proteins. Increasing helix stability is

important for pharmacological applications. However, a rational method is still lacking.

There has been significant progress in our understanding of the interactions responsible

for helix stability in peptides. In order to from a stable a-helix, at least 12-15 a-amino

acid residues are necessary and such a helix is generally is not stable in water. On the

other hand, ß-peptides with as few as six ß-amino acids (derived from the corresponding

L-amino acids) fold into 3i4-helices in methanol.

In the case of an a-peptidic helix, a macrodipoles of increasing size results chain length1,

which points from the N- to the C-terminus [1]. This macrodipole destabilizes the a-

helix. In contrast, the 3i4-helix has a macrodipole2 in the opposite direction, i.e. from the

C- to the TV-terminus, leading to a charge-macrodipole stabilizing effect ("intrinsic

capping"). Thus, unlike a-peptides, ß-peptides with longer chain lengths can be expected

to form stable helical structures in solution even without salt-bridges or 'capping'

stabilization by charged side chains. The fact that the 3i4-helix of ß-peptides is more

stable with unprotected than with protected termini corroborates this interpretation (See

Figure 26) [2, 3J.

1In proteins, the charges imparted to the ends of the a-helix by the dipole are often counteracted by

terminal residues bearing polar and charged side chains. This is termed "capping".The mixed ß2/ß3-peptide has no net dipole, as alternating C=0 dipoles are oriented in opposite directions

along the helix axis.

C-Terminus

54A

Af-Terminus

[—47 A—j

(P)-5.6,3-Helix

C-Terminus

A/'-Terminus

1—54 A —I

(M)-3U-Htlix

Figure 26. Comparison of a-helix and ß-peptidic 3]4-helix showing the differences of macrodipole and the

helicity.

The lengths of helices in proteins and enzymes rarely exceed five turns, i.e.

approximately 18 amino-acid residues in a 3.6i3-helix. The recent NMR-solution

structure of an oc-pcptidic 25-mer in the strongly helix-inducing solvent CF3CH2OH/H2O

showed a helix mainly in the region of residues 7 to 23. Similar structural properties of

a-peptides3 containing about twenty residues have been reported, as determined by

NMR-spectroscopy in solution [4, 5].

No solution or solid-state structural information about longer ß-peptides containing more

than twelve residues was available at the beginning of this thesis. So far, helical

structures of long-chain ß-peptides have only been inferred from CD spectra, which can

provide an indication of the presence of a helix in solution [6]. However, peptides that

adopt small populations of helical conformations may be difficult to detect through CD

spectra. Such structures may be of importance for initiation of folding, and can be studied

by NMR-spectroscopy. The longest ß-peptide, for which an NMR-solution structure has

been determined, is a ß-dodecamer, which forms a 14-helical structure in methanol

solution [7].

An a-peptide consisting of residues 13-33 of HIV-1 Vpr in micelles. It forms an amphophilic, leucine-

zipper-like a-helix, which serves as a basis for interactions with a variety of viral and cellular factors.

89

In order to check the propensity and stability of the 14-helix with increasing peptide

chain length the ß3-icosapeptide 29 was designed. It was synthesized on solid phase and

its structure in solution was investigated by high-resolution NMR-spectroscopy.

Me.

H2N

H2N^NHNH

A^ANAvANAN^kN

29

HO'^O

The ß-peptide sequence was designed such that it contains ß-amino acids with all 20

proteionogenic side chains. Careful planning of the sequence was essential. In making

helices, it is important to consider factors such as hydrophobic interactions or salt-

bridges. The helix propensities of individual amino acids and sequence-dependent

phenomena had to be considered. Based on our group's earlier observations with shorter

ß-peptides, the sequence was designed in such way that a possible 3i4-helical

conformation would be stabilized by one series of hydrophobic side chains being stacked

above each other on one face of the helix surface, and two strands of side chains

contained polar groups and one salt-bridge each on the other, as illustrated in Figure 27a.

The synthesis of ß-icosapeptide4 29 and its structural analysis by CD measurement have

been reported previously [8],

4See the PhD thesis of Thierry Kimmerlin for the synthesis of ß-icosapeptide 29 (ETH Diss. No. 15800).

90

a)

b)

Salt Bridge

Hydrophllic

Hydropfcofrfe

"normal" ß3-amino acid

in a 3i4-helixhelix-breaking

ß-'hPro in a ß-peptide

Figure 27. a) Schematic presentation of an (Af)-3i4-helix of ß-icosapeptide 29 and side cham arrangementof each amino acid on the helical wheel, b) Comparison of a non-cyclic ß^aa with ß3hPro in the

conformation required for the folding to an (A/)-3i4-helix.

The TV-terminal ßhCys side-chain was protected by disulfide formation. The ßhPro was

placed at the C-terminus because, of all the residues, ßhPro has the least preference for

occurring within the helix and is known to disrupt the helix formation due to the lack of

an amide proton for hydrogen bonding in both a- and ß-peptidic helices. Furthermore the

rigidity ofthe of ßhPro ring prevents the preceding residue in the sequence from adopting

a helical conformation. Although ßhGly lacks a side chain, depriving it of helix-

promoting side-chain interactions and increasing its conformational flexibility, it can still

be included both in helices and in turns. It was decided not to use 'capping' by charged

side chains near the two termini, so that we could study the helix's inherent stability.

91

4.2 Assignment and NMR Structural Analysis of ß -Icosapeptide 29

Firstly, all the spin systems of amino acids present in the sequence needed to be

identified. These differ only in the side-chains with their characteristic coupling network.

The assignment of all the protons of the ß-icosapeptide5 was nontrivial, though the

sequence contains 20 amino acids which are unique in their constitution. Their almost

identical chemical shift values imposed a limitation on the identification of the particular

residues. The N-tcrminal ammonium group was not observed due to rapid exchange with

the solvent.

It was possible to identify the various residues through DQF-COSY correlation typical

for a given amino acid, enabling the assignment of groups of all coupled protons

belonging to the same residue. In the DQF-COSY spectrum, the NH/Hß cross peak

region clearly showed 18 cross peaks corresponding to 18 amino acid residues of the

sequence. The absence of an NH/Hß cross peak led to the identification of ßhPro which

lacks an amide proton and displays a characteristic pattern in the aliphatic region. The

ßhGly residue was quickly identified in this region, as the NH signal was split to a triplet

by coupling with the two Hß protons. Amino acids ßhAla, ßhVal, ßhLeu, ßhlle and

ßhThr with their unique spin connectivités and methyl group containing side chains could

be easily distinguished from each other based on their COSY connectivities. The

remaining amino acids ßhAsp, ßhAsn, ßhSer, ßhCys, ßhHis, ßhPhe, and ßhTyr each

contain two methylene groups (a, y) CH2. However, in combination with TOCSY, which

established the connectivity from each NH through to the corresponding side-chain

protons, all residues could eventually be assigned. The sequential assignment (i.e.

position of each residue) was confirmed by the backbone NOEs with medium intensity

da,N(i,i+l) which connects sequential residues. The sequential weak dN,N(i,i+l) cross

peaks assisted the assignment process. The sequence was further confirmed through the

HMBC correlations across NH;-C=Oi-i amide units. The chemical shifts and coupling

constants of ßMcosapeptide 29 are summarized in Table 13.

The NMR measurement was carried out by Christian Hilty and Touraj Etezady-Esfarjani of the Wülhrich

group (Department of Biology, ETH Zürich), using a 750 MHz instrument.

92

Table 13. 'H-NMR-Chemical shifts for ß3-icosapeptide 29 in MeOH

ß-aminoacid

NH H2-C(a) H-C(ß)*JHNHß

H-C(y)

Me-C(Y)H2-C(y)

H-C(ö)

Me-C(Ô)

Me-C(e)

ß3hCys' 2.86/2.97 4.00 3.07/2.11

ß3hAla2 8.21 2.84 4.57

9.5

1.25

ß^Ser3 8.62 2.48/2.93 4.31 3.56

ß3hHis4 8.54 2.57/2.76 4.66

8.5

2.82/2.95

ß3hAsn5 8.33 2.59/2.89 4.75

9.2

2.15/2.40

ß3hGlu6 8.23 2.47/2.87 4.56

9.5

1.85

ß3hGIy7 8.68 2.38/3.09 2.88/4.05

ß3hTrps 8.63 2.61/3.02 4.70

8.4

3.0/3.1

ß'hArg9 8.53 2.54/2.92 4.50 1.76 1.64 3.15/3.21

ß3hVal10 8.39 2.53/2.83 4.32

9.2

1.84 0.99/1.02

ß3liAspu 8.47 2.75 5.14

8.8

2.85/2.98

ß3hGln12 8.60 2.53/3.03 4.55

9.2

1.82/2.00

ß3hlle13 8.03 2.74/2.75 4.29

9.5

1.48 0.94/1.09

ß3hLys14 8.24 2.61/2.93 4.55

9.6

1.70 1.19 1.77

ß^Thr15 8.14 2.87 4.51

8.8

3.85 1.19

ß-hLeu16 8.16 2.39/2.53 4.56

9.5

1.37/1.51 1.69 0.95/0.97

ß'hTyr17 7.92 2.34/2.47 4.59

8.7

2.69

ß3hMet18 7.53 2.25/2.47 4.59

9.5

ß'liPhe19 7.74 2.48/2.63 4.68

8.3

2.81/2.91

ß3hPro20 2.88 4.37 2.07 1.91/1.95 3.41/3.50

93

The large coupling constants {ca, 9 Hz for 3J(NH-Hß) which were extracted from the 1-D

*H NMR spectrum correspond to an antiperiplanar arrangement of NH and Hß (Figure

27b). The diastereotopic H2C(a) protons were assigned by assuming that the axial

protons exhibit a large coupling, and the lateral protons a small coupling with H-C(ß).

This is in agreement with stronger NOEs being observed from H-C(ß) to the lateral

HiaiC(a) protons, compared to the axial H^Cfa) protons, and stronger NOEs from Hax-

C(a); to the NHj+i protons.

To determine the three-dimensional structure of the peptide, NOESY spectra at different

mixing times (150 ms and 400 ms) were acquired. Examination of the NOE patterns

showed all the cross peaks characteristic of 3i4-helical conformation. The cross peak

volumes of the NOESY spectrum with tm = 400 ms were converted into distance

restraints and calibrated with known distances according to the two spin approximation.

Together with dihedral-angle constraints (derived from the coupling constants) the

distance restraints were used in simulated-annealing (SA) molecular-dynamics

calculations of structural bundles with XPLOR-NIH. Starting with a randomized

conformer, all 30 structures calculated by SA converged to a single cluster with no NOE

and dihedral angle violations. The bundle of 10 structures with the lowest energy is

displayed in Figure 28 and reveals that the molecule adopts a 3i4-helical conformation

over its whole length, although the backbone is less well defined for the two terminal

residues (because of the smaller number ofNOEs at the ends of the sequence). Compared

with shorter 3i4-helical ß-hexa- and ß-heptapeptides observed earlier, the bundle

obtained for ß3-icosapcptidc 29 has a larger rmsd value for the backbone. In fact, the

individual structures in the bundle vary in total length (as measured from the TV-terminal

nitrogen atom to the C-terminal carboxyl carbon atom) from 25 to 34 Â. This variation is

a consequence of the limited range of the detectable NOEs (i to max. i+3) which leads to

an accumulation of small errors along the full length of the 20-mer. Although the NOE

data leave no doubt that the secondary structure of this ß-icosapeptide is an uninterrupted

3i4-helix, the local conformation in the side chains is not well defined by the

experimental data. In particular, only one of the two salt bridges built into the sequence

by design is corroborated by a weak NOE between the side chains of ß3hAspn and

ß3hLys14. A view along the axis of the (A/)-helix forthe section containing residues 8-14,

reveals that the side chains C(y) shown in i and (i+3) positions are offset from an ideal

3i4-hclix by 10° to 20° per turn in a right handed direction.

94

a) b)

c) d)

\\ ^

Figure 28 Solution-NMR structure of ß3-icosapepüdc m MeOH, bundle showing lowest energy structuresfrom simulated annealing calculations a) Structural bundle with all side chains b) Superposition ofbackbone atoms from residues 2-19 c) Single structure from the bundle showmg six helical turns, with sidechain interactions d) '1 op view along the helix axis

95

There are 3.1 to 3.4 residues per turn (six turns) with an average pitch of 4.8 Â, and the

cylinders least-square fit through the C(O) and C(ß) atoms have diameters of 4.0 and 5.4

Ä, respectively. The corresponding values from an X-ray structure of an a-peptidic 3.613-

helix are 3.6 and 4.7 Â as depicted in Figure 26.

4.3 NMR-Solution Structure of an a-Heptapeptide

It is known that in proteins a-peptidic stretches in the range of 10-25 amino acids can

adopt secondary structures such as a-helices or ß-sheets, but simple a-peptides of this

length generally do not form helices in protic solvents. Exceptions are the fluorinated

alcohols CF3CH2OH and (CF3)2CHOH. A GROMOS MD-simulation of a short o>

peptide had predicted that no particular secondary structure would be populated in MeOH

[9]. However, the solid state structures of short peptides containing Aib residues show

helical conformations. Synthetic a-tri-, a-tetra-, and cc-pentapeptides containing at least

one Aib residue they all adopt conformations corresponding to the 3io-helix [10],

whereas longer (6 to 20 residues) Aib-containing peptides predominantly fold into o>

3.613-helices [11 J. In contrast, oligopeptides made up of ß-amino acids are known to

adopt rather stable helical or ß-sheet structures even for very short chain lengths down to

4-7 residues in an organic solvent such as methanol.

We have always claimed that comparable a-peptides consisting of less than ten amino-

acid residues do not fold into a helix under the conditions we used to detect the folding of

their counterparts containing homologated amino acids (NMR analysis in MeOH). When

looking at the huge body of work by Baldwin et al, ('single domain proteins') [12], this

claim appears to be reasonable, but we have actually not found any report on an NMR-

structural study of a short a-peptide in MeOH, which is considered to be a helix-

stabilizing solvent.

This lack a of "reference" prompted us to investigate the NMR solution structure of a

seven-residue a-peptide with a central Aib residue in order to verify that, in contrast to ß-

peptides, short a-peptides do not form a helical structure in MeOH. Thus, the a-

heptapeptide 30 was synthesized. In order to facilitate the interpretation ofNMR spectra

to be measured, all residues were chosen to be different and Aib, the helix-inducing

residue was placed in the central position. The structure was analyzed by NMR

spectroscopy in MeOH.

96

o :n 0/Nn 0 \

30 130 „

S\

A set of Standard NMR experiments DQF-COSY, TOCSY, HSQC, HMBC of a-peptide

30 in CD3OH at 500 MHz/125 MHz permitted complete assignment of all 1-D lH- and

13C-resonances {Table 14). The coupling constants 3J(NH-Ha) were extracted from the

'H-NMR spectrum. Analysis of the ROESY spectrum showed no NOEs typical of one of

the known secondary structures of a-peptides. Integration of NOE cross-peak volumes

from the 300-ms ROESY spectrum yielded distance restraints, which were used in slow-

cooling-simulated annealing calculations with XPLOR-NIH.

Table 14. 'H-NMR-Chemical shifts for a-heptapeptide 30 in MeOII

a-Amino

acid

NH H-C(a)^JHNHa

H-C(p) H-C(y)

Me-C(y)

CH2(y)

H-C(8)Me-(S)

C(a) C(ß) C(Y) C(8)

Val1 3.66 2.18 1.04 1.04 59.85 31.73 19.00

Ala2 8.54 4.45

6.85

1.38 50.53 17.92

Leu3 8.13 4.22

6.24

1.59 1.73 0.95 54.47 41.69 25,80 22.20

Aib4 8.17 1.45 25.84/

25.20

lie5 7 29 4.18

7.70

1.96 1.22/

1.70

0.88 60.01 37.39 26.10 11.60

Met6 8.01 4,43

7.83

1.94/

2.00

2.48 54.12 32.40 31.20

Phe7 7.90 4.65

8.19

2.99/

3.21

55.35 38.59

Each calculation was started with an extended conformer, and a bundle of all calculated

structures with neither NOE nor dihedral-angle constraint violations is shown in Figure

29. All calculated structures do show a bend induced by the Aib residue in the middle of

the sequence but otherwise, this a-heptapeptide does not assume any one of the known

canonical secondary structures. Although the bundle of calculated structures appears well

defined and does not show any NOE violations, this may be the result of the

97

predominance of sequential NOEs. In contrast to long-range NOEs, such as those

determining the central bend; time-averaged sequential NOE constraints from a

multiconformational ensemble are more likely to appear consistent with a single (unreal)

conformer without violations of the error bounds given by the constraints.

Figure 29 Solution-NMR structure of the cc-peptide 30 m CD3OH Low-energy bundle structures with no

NOK or dihedral-angle constraint violations obtained from simulated annealing calculations. Superpositionbased on the backbone atoms of the central Aib residue.

4.4 Conclusion

For the first-time, the structural propensity of a longer ß3-icosapeptide containing ß3-

amino acids with 20 different proteinogenic side chains has been investigated by 2D-

NMR-spectroscopy. High resolution NMR measurments showed that the ß3-icosapeptide

forms a (/W)-3i4-helix in MeOH solution, which is, so far, the longest ß-peptide shown to

adopt a helical conformation in an organic solvent. The fact that the 20-mer helix does

not appear to suffer from macrodipole destabilization may be taken as yet another piece

of evidence for the importance of the staggered ethane bond in the backbone of ß-

peptides The conformation-stabilizing effect of this element, which is present in 18 of

the 20 residues, seems to outweigh the destabilization by the macrodipole. Thus, we can

conclude that both very short and very long ß3-peptidic 3i4-helices are much more stable

than their natural counterparts under the same solvents, and temperature conditions.

The observed helix may indeed be an attractive model for the trans-membrane ion

channels (length ca. 40 Â), since only a few more ß-amino acids would have to be added

to the icosapeptide, in order to obtain a sequence that would span a bilayer. Such

98

molecules would be useful to study the potential of ß-peptides to mimic the interaction

between antimicrobial peptides (AMPs) and model lipid systems, and, thus, to better our

understanding of the mode of action of AMPs, as well as how peptide properties such as

sequence, charge and charge distribution, conformation, hydrophobicity, and peptide

length affect the interaction of AMPs with lipid membranes in bacteria and eukaryotic

cells.

Although the a-heptapeptide was designed to maximize its tendency to adopt an a-

helical conformation, the NMR investigation confirms that short a-peptides do not

assume any particular secondary structure in MeOH. This points to an essential role of

the additional methylene moiety in the backbone of ß-peptides for helix stability.

4.5 References

LI] R. Aurora, G. D. Rose, Protein Sei. 1998, 7, 21.

[21 D. Seebach, S. Abele, K. Gademann, G. Guichard, T. Hintermann, B. Jaun, J. L.

Matthews, J. V. Schreiber, L. Oberer, U. Hommel, H. Widmer, Helv. Chim. Acta

1998, 81, 932.

[3] M. Rueping, J. V. Schreiber, G. Lelais, B. Jaun, D. Seebach, Helv. Chim. Acta

2002, 85, 2577.

[4] A. De Capua, A. Del Gatto, L. Zaccaro, G. Saviano, A. Carlucci, A. Livigni, C.

Gedressi, T. Tancredi, C. Pedone, M. Saviano, Biopolymers 2004, 76, 459.

[5] A. Engler, T. Stangler, D. Willbold, Eur. J. Biochem. 2001, 268, 389.

[6] D. Seebach, J. V. Schreiber, P. I. Arvidsson, J. Frackenpohl, Helv. Chim. Acta

2001,54,271.

[7] T. E. Esfarjani, C. Hilty, K. Wiithrich, M. Rueping, J. V. Schreiber, D. Seebach,

2002,55,1197.

[8] T. Kimmerlin, D. Seebach, Helv. Chim. Acta 2003,86, 2098.

[9] T. Soares, M. Christen, K. Hu, W. F. van Gunsteren, Tetrahedron 2004, 60, 7775.

[10] C. Toniolo, G. M. Bonora, A. Bavoso, E. Benedetti, B. Di Blasio, V. Pavone, C.

Pedone, Biopolymers 1983, 22, 205.

[11] I. L. Karle, J. L. Flippenanderson, K. Uma, P. Balaram, Proteins 1990, 7, 62.

[12] R. L. Baldwin, G. D. Rose, TIBS 1999, 24, 26.

99

ß-Peptidic Secondary Structures Enforced

5 by Zn2+-Complexation: SpectroscopicEvidence

5.1 Introduction

Molecular assembly in nature is a fundamental process and is achieved through the

control of non-covalcnt interactions that include hydrogen bonding, electrostatic,

hydrophobic, aromatic ji interactions and metal coordination. Proteins are known to fold

into specific, compact tertiary structures that carry out many of the specific binding,

catalytic, or regulatory functions within a cell. As such, they are unique in their ability to

recognize and interact with a diverse pool of molecules [1]. In fact, the overall structure

of a protein is derived from the assembly of a limited number of regular secondary

structures such as helices, turns and ß-shects; these structures serve as modules for

construction of tertiary and quaternary structures.

The ability of peptides to adopt various secondary structures has been a subject of

extensive study over the years. The stability of specific folding is determined by the type

of interaction between residues that are remote in a linear polypeptide chain. In order to

understand the fundamental principles govern the structure, stability and function of

peptides, de novo design offers a way to study the basic interactions that cause protein

folding and protein-protein interactions. In addition, the increasing rate of protein-

structure determination provides insight into the features that stabilize the three-

dimensional structures of proteins.

The creation of a specific tertiary structure with ß-amino acid oligomers has been

achieved by Seebach and coworkers [2] who have described the aggregation behavior of

a ß-heneicosapeptide that could, in principle, also fold into an amphiphilic 3i4-helical

structure1. Similarly, Gellman and co-workers [3], have synthesized an amphiphilic 3m-

helix that undergoes self-association in aqueous solution to form tetrameric and

hexameric aggregates. Recently, the structure of a ß-dodecapeptide that self assembles

1To date, a ß-peptidc containing 20 amino acids has been structurally characterized in solution.

100

into a thermostable quaternary structure was determined by X-ray crystallography [4]. In

a parallel approach, Cheng and DeGrado [5] synthesized a two-helix bundle derived

from ß-amino acids, in which the helices are held together through a disulfide bond and

by long range contacts (i.e., hydrophobic interactions and salt bridges). As an extension

of this concept, Diederichsen and co-workers synthesized ß-peptides designed to fold

into a 3 i4-helical structure containing nucleobasc recognition units in the side chains [6,

7]. The nucleobase pairing can organize these ß-peptide helices, gaining more control

over the geometry, stoichiometry, and specificity of self-association [8, 9].

Among the numerous possible approaches that can be used to induce and reinforce intra-

or inter molecular interactions, the use of a metal ion capable of binding to certain

functional groups in a peptide has been adopted. Metal ion coordination to the peptide

predominantly organizes the local structure of the active site in most metallocnzymcs. A

number of attempts have been made to rationalize the interaction of metal ions with

peptides to tune the secondary structures in peptides.

Proteins make use of metal ions as intrinsic parts of their structures. Metal ions serve a

variety of functions in proteins, the most important of which is to enhance the structural

stability of the protein in the conformation required for biological function and/or to take

part in the catalytic processes of enzymes. The use of Zn2+ chelation by Cys and His side

chains to stabilize the conformational rigidity of proteins is a recurring feature found in

Zn-finger protein domains.

Figure 30. The diagram shows the coordination of a Zn+

ion by characteristically spaced Cys and His

residues in a zinc finger motif.

101

Only the structural rigidity imparted by metal ion chelation allows these protein domains

to selectively bind DNA by making specific contacts with base pairs (major groove)

through the side chains of Arg, Lys, Asp, etc. The most common Zn-finger protein motif

is made up of an a-helix and a hairpin turn. These two secondary structural elements are

held together by a Zn2+ ion that binds to two His residues, located in the i and i+4

positions of the a-helix, and to two Cys residues located at equidistant positions in the

hairpin turn. The overall arrangement creates a hydrophobic pocket in which the Zn2+ is

bound with its optimal tetrahedral geometry (Figure 30). Thus, in order to mimic such a

complex three-dimensional arrangement of atoms, we must assess the folding propensity

of p-peptides upon chelation with Zn2+ ions.

Metal binding sites in proteins and short a-peptides have been previously designed,

through the incorporation of both natural and non-natural amino acids, to study the effect

of a metal ion on the stability of secondary structures [10, 11]. In a-peptides, two

histidine residues in i and (i+4) positions at the C-terminus, as well as in the middle of a

short peptide, can nucleate and stabilize a helical conformation upon addition of various

metal ions such as Zn2+, Cd2+, Cu2+, Pd2+, and Ni2+ [12-14]. A similar approach has been

used to stabilize secondary turn structures. Searle and coworkers reported short a-

peptides with a proline residue to induce a reverse turn, and two or three histidines for

metal binding on both strands near the N- and C-termini [15]. They observed a

substantial stabilizing effect of the hairpin structure by cordination with Zn2+ ions.

102

5.2 Design of ß-Peptidic Hairpins

In a-peptides, a hairpin (strand-loop-strand segment) is the smallest structural unit

responsible for folding to a ß-sheet secondary structure. The hairpin motif has been

widely employed to generate parallel and antiparallel ß-sheet structures.

ß-turn nucleating segments

., /=° HN U0 UN] NH""

(D)-Pro-Gly Nipecotic acids ß2/ß3-Amino acid

Figure 31, The hair-pin-turn inducing segments employed for the generation of antiparallel sheet

structures in a- and ß-peptides.

In earlier work, Balaram and coworkers used a typical motif D-Pro-Gly found mainly in

a-peptides, to nucleate the turn [16, 17], and characterized mixed a/ß-peptides by

incorporation of the ß-amino acid residues into the strands. Gellman and co-workers used

a D-Pro-Xxx linker, where Xxx was l,2-diamino-l,l-dimethylethyl or a-hydroxy acetate,

for the construction of parallel and antiparallel hairpin2 structures, respectively [18-20].

Also, a template containing a nipecotic acid (ß2-homoproline) unit was used to enforce

the formation of a hairpin structure. Finally, the Seebach group [21, 22] showed that the

insertion of a (5)-ß2hVal/(S)-(>S)-ß3hLys reverse tum segment in the middle of a ß-

hexapeptide, with two sheet inducing w-ß2,3-amino acid residues on either strands, folded

to a hairpin in methanol and probably also in aqueous solution.

2The segments depicted in Figure 31 cause reversal of a peptide chain. The ß-amino acids form polar ß-

shects due to an array of unidirectional hydrogen bonds; in contrast, a-amino acid residues produce non-

polar ß-sheets, with alternating directions ofhydrogen bonds.

103

Although it has been well established that ß-pcptides can adopt water-stable helical

conformations by interactions (covalent and non-covalent) such as disulfide bridges, salt

bridges and macrodipole-"capping", [23] no example of a ß-peptidic hairpin secondary

structure, stabilized by metal-ion complexation is known. So far, ß-peptide structures

have been extensively investigated in organic solvents3 and in the solid state. Thus, we

decided to study the influence of a Zn2+ ion in stabilizing secondary structures of ß-

peptides in aqueous solution and to investigate a possible assembly of ß-peptides

sequences into tertiary structures such as Zn-finger motifs. Using NMR spectroscopy, a

series of ß-octapeptides was examined that allowed us to explore the relationship

between the absolute configurations of amino acids in a sequence and formation of a

helix or a hairpin structure in solution.

Recently, a ß-peptidic 3i4-helix capable of chelating Zn2+ with two ßhHis residues has

been studied by Seebach et al. [24]. The helix was designed to complex Zn2+ through the

two His residues located at positions i and i+3 of a peptide made entirely of ß3-amino

acids {Figure 32). The ion chelation by the two ß3hHis residues was expected to give

higher stability to the helical structure in aqueous media, which was confirmed by CD

measurements. However, it is still unknown how much stability is imparted to the

structure by this metal cordinatin and whether the 3i4-helical structures offers a template

in which Zn2+ ion chelation by the two ß3hHis residues is optimal.

3The solvent methanol is known to promote secondary structure formation in both a- and ß-peptides.

104

^v

> Salt Bridge

j

NH?

OH

HO H02C (^ —/ Ä-7 Ct°2H J M f

HHHHHHHH

ß3-decapeptide carrying ß3hOrn4, ß3hGlu7, and ß3hLys10 for salt-bridge formation and p3hHis6and ß3hHis9 for Zn2+ complexation

Figure 32. Schematic presentation of a ß -decapeptide designed to have a salt-bridge-stabilized andZn2+-

complexing helical structure to promote folding in aqueous solution [241

In order to mimic a structure resembling a Zn-finger domain, a peptide having proper

sequence of ß-amino acids was designed. The sequence can be understood simply by

considering the two separate structural elements of the Zn-finger domain: a helix and a

hairpin turn. The task is simplified even further after realizing that, contrary to a-

peptides, ß-peptidic secondary structures are not much determined by the amino-acid

residues themselves. Only few ß-amino acids (ca. 14-16) should be enough to

mimic/stabilize tertiary structure of a Zinc-finger. Therefore it is important to study the

structural propensity of ß-peptides in the presence of Zn2+ ions. The design principles of

stabilizing or destabilizing the idealized 14-helix and the ß-peptidic hairpin structures are

summarized in Figure 33.

105

a)

l|

*

P fA f /

p\ r®\

$ J'-^zf f

b*r

b)

c)

H2N

«2.3

R3 O

"Y^" „,NW0H H!NV*°»S2 "2"

allowed in 314-helix

R3 O

allowed in ß-sheet

i-îin T OH, ^ , „

52 H2N ^ OH

RJ O

1 1

Figure 33. Design principles: (a) (A/)-14-Helix and (b) hairpin turn of ß-peptides with allowed and

forbidden ß-homoamino-acid constitutions and configurations. All positions labeled with red H are

forbidden for non-H-atoms (exceptions may be F and OH in the ß-peptidic 3i4-hclix and pleated sheet,

c) Allowed configurations of ß-amino acids in a 14-helix and ß-sheet structure.

106

The design of a hairpin tum capable of chelating 2n2+ ions is based on a peptide reported

by Seebach et al. Figure 34), which was modified by simply adding ß hCys at the C-

terminus and ß3hHis at the /»/-terminus. Thus, one can hypothesize that the two termini of

the peptide might be brought together through metal ion complexation thus stabilizing the

hairpin structure. It was judged that the greater conformational flexibility of the terminal

residues ß3hHis and ß3hCys should significantly facilitate Zn2+ coordination, with the

remaining tetrahedral coordination sites of zinc being occupied by either the C- terminal

carboxylate, the TV-terminal amino group, or a solvent molecule.

To study Zn2+ complexation with the structural framework of a ß-peptidic hairpin turn

and pleated sheet, we designed and synthesized4 three ß-oetapeptides. They all contain a

central segment of (5}-ß2hVal-(/S)-ß3hLys that is known to induce formation of a ten-

membered H-bonded ring, which also fits into a 14-helix. ß3hHis and a ß3hCys were

placed at the N- and C-termini of the ß-oetapeptides respectively, so that Zn2+ ions would

be able to force these termini into close proximity by complexation. The remaining pairs

of amino acids ßhXaa2/ßhXaa3 and ßhXaa6/ßXaa7 were chosen such that there would be

a natural hairpin turn in the absence of Zn2+ ion (ß-oetapeptide 31). A preferred 14-

helical structure of a ß-oetapeptide 32 which would also be able to form a hairpin [25].

Finally, the ß-oetapeptide 33 a diastereoisomer of 32, contains two violating ß3-homo-

amino acids at positions 2 and 7 for both secondary structures which is neither capable of

folding to a 14-helix nor to a hairpin turn.

4

ß-Peptides for zinc-complexation studies were synthesized by Gerald Lelais and Dr. Oliver Flögel.

107

o "N^" o o S o I o / o

OHH I

-J V^

J Ky y y

strand turn segment strand

HN^N

OiOiO oNoiOio/o

H = H = H H HEHEH

, y -S. y

chelation site 31 chelation site

hairpin structure, turn with antiparallel sheet; with four helix breaking residues in strands

NH2

>— OH

^ o I o N^ o oNo/o/o/o

H2N^^N^^N^^N^^N^^^N^^N^^N^^OHH H HJLH H H H

32

helix structure, no violation of hairpin structure with antiparallel elements

HN^N

o i o "V"'' ° °

H2N ^ N ^ N ^^ N ^-T NH H H I H

33

diastereoisomer of 32; can not fold to a helix or a hairpin structure

Figure 34, ß-Octapeptidcs with terminal Zn2+-binding amino acids ß3hHis and ß^Cys. Peptides 31-32 can

possibly fonn a hairpin structure (turn with antiparallel sheet). Peptide 32 can fold to a 314-helix. ß-Octapeptide 33 is violating both, the rules for 14-helix- and for hairpin-structure formation.

108

5.3 The Structural Analysis by NMR-Spectroscopy: Zn Complexation

of Hairpins (31-33)

Earlier MS spectra and CD spectral5 [26] analysis of these peptides revealing that Zn2+

coordinates with ß-peptides to form 1:1 and 2:1 zinc-peptide complexes provided

valuable hints to the existence of secondary structures. A detailed secondary structure

determination of ß-peptides 31-33 in solution by high-resolution 2D-NMR spectroscopy

of ß-peptides 31 (in H2O, with and without ZnCb), 32 and 33 (in MeOH and in H2O,

with and without ZnCU) was carried out in order to understand the above observation.

The detailed NMR analyses involved DQF-COSY, TOCSY, HMBC, ROESY

measurements and interpretation of the data according to the XPLOR protocol The

corresponding chemical shifts and 3J(NH-Hß) coupling constants extracted from 'H-

spectra arc given in Tables 15-21.

Hairpin-Structure of ß-Octapeptide 31

Peptide 31 was designed to fold to a hairpin turn with a central ß2-ß3 segment, just like

the previously reported ß-hexapeptide with sheet-enforcing ß2'3-residues of u-

configuration in both strands. So far, no ß-peptidic turn has been characterized in

aqueous solution. The four ß2'3-residues in compound 31 were expected to stabilize such

a turn sufficiently to be observed in H2O. Thus, we determined the NMR-solution

structure of this ß-oetapeptide in H2O in buffered solution at pH 5.4 (with and without

added ZnCl2).

In both solutions (in H2O, with and without ZnCh), the large coupling constants for w-ß2,3

residues on either strand indicated an antiperiplanar arrangement of the NH and Hß

protons. The three-dimensional structure was obtained by ROESY experiments with a

mixing xm = 300 ms. The extended conformation of residues in both strands was

established by NOEs observed between the NHj and Ha; and Hom. Further, strong

evidence for a hairpin structure in aqueous solution was supported by diagnostic NOEs

between residues 3 and 6 (Ha-Hß and NH-Hß).

5The detailed CD spectra analysis of peptides 31-33 and designed Zn-finger with or without ZnCl2 in

aqueous solution and ESI-MS characterization of peptide complexes can be found in [26].

109

Table 15. 'H-NMR-Chemical shifts for ß-octapeptide 31 in H20

ß-amino NH H2~C(a) H-C(ß) H-C(y), H-C(8) Me-C(s)acid 3JHNHß Me-C(y) Me-C(8)

H;-C(y)

ß3hHis' 2.67 3.94 3.15

ß3hAla2 8.04 2.36 3.9

9.54

1.02

ß3hAla3 8.06 2.3 3.95

9.32

1.01

ß3hVal4 7.89 2.1 3.29/3.41 1.76

ßUys5 8.07 2.35 4.1 1.49

0.93

1.4 1.62

ß3hAla6 8.0 2.34 3.96 1.08

9.17

ß3hAla7 8.09 2.33 3.98 1.1

9.24

ß3hCys8 8-1 2.49 4.25 2.74

8.95

Table 16. 'H-NMR-Chemical shifts for ß-octapeptide 31 in II20 + ZnCl2

ß-amino NH H2-C(a) H-C(ß) H-C(y), H-C(S) M^"acid 3JHNHß Me-C(y) Me-C(8) C(s)

-^-Yïk£ÛÛ

ß3hHis' 2.67 3.92 3.14

ß^la2 8.06 2.35 3.94

9.17

1.02

ß3hAla3 8.058 2.34 3.95

9.32

1.03

ß^hVal4 7.9 2.17 3.28/3.43 1.76

ß^Lys5 8.07 2.36 4.1

8.66

1.3/1.49

ß3hAla6 8.04 2.35 3.96

9.32

1.05

ß3hAla7 8.0 2.37 3.98

9.46

1.11

ß3hCyss 8.09 2.4/2.54 4.22

8.44

2.63/2.74

0.87/0.93

1.3/1.49

1.33/1.41 J61

110

Table 17. 'H-NMR-Chemical shifts for ß-octapeptide 32 in MeOH

ß-aminoacid

NH H2-C(a) H-C(ß)

3JHNHßH-C(y),

Me-C(y)

H2-C(y)

H-C(S)

Me-C(S)

Me-C(e)

ß3hHis' 2.58/2.87 3.95 3.03

ß3hAla2 8.5 2.49/2.82 4.57

9.1

1.2

ß'hVal3 8.2 2.43/2.81 4.04

9.54

1.74 0.92

ß2hVal4 8.29 2.76 2.9/3.9

9.1

1.75 0.95/1.0

ß3hLys5 8.36 2.44/2.56 4.42

8.8

1.64 1.49 1.66

ß'hLeu6 7.6 2.28 4.56

9.46

1.31/1.42 1.59 0.91

ß3hSer7 7.69 2.38/2.64 4.39

8.6

3.41/3.54

ß3hCyss 7.92 2.46/2.69 4.62

9.32

2.85/2.96

Table 18. 'H-NMR-Chemical shifts for ß-octapeptide 32 in H20 + ZnCl2

ß-amiiio NH H2-C(a) H-C(ß) H-C(y), H-C(Ô) Me-C(s)acid 3JHNHß Me-C(y) Me-C(ô)

H2-C(y)

ß3hHis1 3.56 2.74/2.89

ß3hAla2 7.95 2.38 4.19 1.14

ß'hVal3 7.89 2.26/2.44 3.99 1.69 0.82

ß2hVal4 7.92 2.11 3.12/3.49 1.740.85/0.92

ß3hLys5 8.03 2.36 4.14

8.95

1.48 1.25/1.331.62

ß3hLeu6 7.99 2.3 4.2

9.1

1.25/1.411.53 0.81/0.86

ß^Ser7 7.97 2.45 4.18 3.52/3.59

ß3hCys8 7.85 2.52 4.05

9.54

2.61

Ill

Table 19. 'H-NMR-Chemical shifts for ß-octapeptide 33 in MeOH

ß-aminoacid

NH H2-C(a) H-C(ß)

3JHNHßH-C(y),

Me-C(y)H2-C(y)

H-C(S)

Me-C(5)

Me-C(s)

ß'hffis1 2.59 3.90 3.15

ß'hAla2 8.18 2.40 4.27

8.1

1.0 1.20

ß^Val3 7.83 2.50/2.33 4.15

9.7

1.74 0.90

ß2hVal4 8.09 2.04 3,61/3.15 1.82 0.99/0.93

ß3hLys5 8.01 2.42/2.31 4.24

8.9

1.56 1.44 2.92

ß3hLeu6 7.91 2.36 4.32

9.2

1.49/1.28 1.678 0.92

ß3hSer7 7.86 2.48 4.30

8.8

3.56

ß3hCys8 8.10 2.55 4.27

8.9

2.70

Table 20. 'H-NMR-Chemical shifts for ß-octapeptidc 33 in H20

ß-aminoacid

NH H2-C(a) H-C(ß)

3JHNIIßH-C(y),

Me-C(Y)

HrC(T)

H-C(S)

Me-C(8)

Me-C(e)

ß3hHis' 2.55 3.78 2.93

ß3hAla2 8.10 2.37 4.22

9.1

1.14

ß3hVal3 7.86 2.46/2.24 4.00 1.65 0.78

ß2hVal4 7.89 2.09 3.47/3.140 1.75 0.92/0.86

ß^ys5 8.02 2.36 4.13

8.6

1.51 1.63

ß'hLeu6 7.98 2.35 4.18

8.3

1.41/1.26 1.54 0.87/0.83

ß3hSer7 7.97 2.49/2.38 4.23 3.55

10.0

ß3hCyss 8.00 2.45/2.40 4.19 2.72/2.60

8.4

112

Table 21. 'H-NMR-Chemical shifts for ß-octapeptide 33 in H20 + ZnCl2

ß-amino NH H2-C(a) H-C(ß) H-C(y), H-C(5) Me-C(e)

acid 3JHNHß Me-C(y) Me-C(S)Ha-C(T)

ß3hHis' 2.56 3.80 3 03

ß3hAla2 8.12 2.39 4.22 1.14

8.9

ß3hVal3 7.84 2.47/2.27 4.00 1.66 0.786

ß2hVal4 7.93 2.09 3.53/3.09 1.75 0.93/0.86

ß3hLys5 8.031 2.37 4.14 1.50 1.42 1.64

9.2

ß'hLcu6 7.98 2.50 4.18 1.42/1.25 1.55 0.87/0.82

8.6

ß3hSer7 7.96 2.50 4.25 3.55

8.68

ß3hCys8 2.45 4.17 2.68/2.64

9.5

The NOE cross-peak volumes were subsequently converted into distance restraints (in

H20 and in H20 + ZnCl2), together with dihedral constraints derived from coupling

constants, and were used in simulated annealing calculations (X-PLOR). The calculation

yielded a bundle of 30 low energy conformers, with no violations of experimentally

derived restraints, which is presented in Figure 35.

It is evident from these measurements and calculations that peptide 31 is present as a

hairpin turn under these conditions (Figure 35a), with the terminal residues somewhat

less well-ordered than the rest. The central segment contains a 10-membered H-bonded

ring, which resembles a ß II tum0 of a-peptides and the first trans-catenary H-bond

which is part of a 14-membered ring of an antiparallel sheet-structural element. As

expected, addition of ZnCl2 at this pH does not cause significant changes of the structure

(Figure 35b) since the imidazole group is protonated.

The most common two residue turns in globular proteins are type I and type II turns. In contrast, the

preferred turns in ß-hairpins are their mirror images, types Ir and IF.

113

Even if Zn+

were to bind the termini of the octapeptide 31, we would not expect a

dramatic structural change, since the turn is intrinsically preformed and does not require

stabilization by complexation (Figure 35b).

a) b)

Figure 35. The NMR-solution structure of the ß-octapeptide 31 with four sheet-inducing ß2,3-ammo acid

residues detcimined in aqueous solution, a) ß-peptidic hairpin structure in (H20, pH 5.4 (Dn)-Tris-buffer).b) Peptide 31 ((Dn)-Tns-buffered in H20 atpH 5.4) in the presence of 1 equiv. of ZnCl2.

The Helix Structure of ß-Octapeptide 32

The situation is fundamentally different with the ß-octapeptide 32: in MeOH it should

fold to a 14-helix, into which its one and only ß2-amino acid residue of (^-configuration

fits (Figure 33). In H20, we expected the loss of helical structure upon Zn2+

complexation: if the side chains of the terminal ß3hCys and ß3hHis provide enough

binding enthalpy, a hairpin turn would be enforced. There is no turn-violating residue in

this ß-peptide. To prevent complications due to disulfide formation of 32 (R = H) during

NMR sample preparation and during NMR measurements, we converted the cysteine SH

group to a mixed disulfide group (32, R = MeS). All the 'H resonances were assigned

using previously described procedures. All the coupling constants 3J(NH-Hß) are in the

114

range of - 9.2 Hz, which corresponds to an antperiplanar arrangement. Analysis of the

ROESY spectrum with xm = 300 ms, showed characteristic NOE cross peaks for a 3h-

helical structure, which included cross peaks between the (Hai, Hßi+3), (NHi, Hßi+3),

(Ha;, Hßi+3) and (NH,, NHi+i) protons.

The helical structure of 32 in MeOH, which was suggested by the typical CD spectrum ,

was confirmed by NMR structure analysis (Figure 36). A well-ordered 3i4-(A4)-helix is

seen in MeOH, with the N-terminal imidazole and MeSS groups separated by a distance

ofca. 20Â.

Zn2+

$

R = MeS R = H

Figure 36. Bundles of low-energy conformers showing the solution structure of peptide 32 as 3i4-Helix in

MeOH and a hairpin-tum structure in aqueous solution in the presence Zn2+ ions as determined by NMRand SA calculations.

When peptide 32 with a free SH group was dissolved in H2O, (R = H) it lost (most of) its

helical structure according to the CD spectra (H20 and Tris buffer pH 7.3), but we do not

know its structure under these conditions. By NMR analysis it is known that this ß-

peptide exists in a hairpin-turn secondary structure in H2O in the presence of 1 equiv. of

ZnCh (Figure 36). The imidazolyl and the SH groups are close to each other, and

diagnostic NOEs across the two strands confirm the hairpin structure bound to Zn2+ ions.

The structure is stabilized by several H-bonds between the two strands.

7CD measurement of a ß-oetapeptide 32 in MeOH showed a negative cotton effect between 210 and 215

nm, which is considered to be characteristic for a 14-helix structure. Upon addition of ZnCl2 in Tris-buffer,a decrease in intensity of the positive band between 200 and 205 nm was observed.

115

The 'Impossible' ß-Octapeptide 33

Finally, it was especially important to look at the NMR structure of the "doubly

unsuitable" ß-octapeptide8 [27] 33, in which two stereo centers in 32 are inverted at

amino acid residue positions 2 and 7, and which is, therefore, predicted not to be capable

of folding either to a helix or a tum (see formulae in Figure 34) Three detailed NMR

analyses of 33 were performed m CD3OH, in H20 (KH2P04 buffer + 10% D20, pH 7 6),

and m H20 in the presence of Zn2+ ions ((Dn)-Tns buffer, 1 equiv of ZnCl2, 10% D20,

pH 7 2) For structural information about peptide 33 in MeOH, a total of 107 NOEs were

collected and used in simulated annealing calculations The calculated structures having

minimum energy with no violation of either distance restraints or dihedrals are depicted

m Figure 37

a) b) c)

Figure 37 NMR solution structure of ß-octapeptide 33 (The central ß2/ß3 segment represented by purpleand violet The terminal ß3hHis and ß^Cys are represented m orange) a) in MeOH, bundle shows no

central ten-membered H-bonded ring with undefined structures, b) in H2O solution, structure shows some

similarity with structures found in MeoH, c) in H20 in the presence of ZnCl2, the bundle of conformations

showing terminals His and Cys relatively closer in space suggesting Zn2+ complexation

The resulting conformers are rather poorly defined and do not represent any of the known

ß-peptidic secondary structures It is clear that ß-peptide 33 containing the two residues

((#)-ß3hAla2 and (5)-ß3hSer7) in 'wrong' configuration resulted into a different family of

8flic allowed and forbidden positions for ß-peptides to construct either a helix or a hairpin have been

discussed m [27]

116

conformers (compare Figure 36 with Figure 37a). The four C-terminal residues fold into

a vague helix, which is broken by the ß2/ß3 fragment. It is worth noting that this fragment

does not have a turn structure with a ten-membered H-bonded ring, but displays a

conformation in which the two "turn side chains" protrude in opposite directions. In

addition, the phHis and ßhCys side chains are far away from each other, precluding any

Zn2+ chelation in this conformation. For the peptide 33 measured in aqueous solutions, a

total of 101 NOEs were used for the simulation. In this calculation, two sets of

conformations with low restraint violations were obtained, of which 15 structures

correspond to conformation as shown in Figure 37b. Again, the SH and the imidazole

groups are remote from one another.

For peptide 33 in H20 in the presence of ZnCb, a total of 71 NOEs were collected for the

simulation. The calculation yielded 17 structures with low restraint violation and

minimum energy {Figure 37c). From the overlay of these structures, again, a

noncanonical conformation results. The ß-peptide backbone has changed its shape by the

addition of Zn2+ ions, which position the terminal ß3hCys and ß3hHis residues in relative

proximity. It contains a hairpin structure, in which the ß2hVal-ß3hLys fragment (purple

and violet in Figure 37c) do not display the ten-membered H-bonded ring, as seen in the

typical ß-peptidic turns in Figure 35 and 36. Comparison of the bundles in Figure 36 the

hairpin and Figure 37c, provide a nice demonstration of the correctness of the rules for

allowed and forbidden positions of side chains on ß-peptidic hairpin turns. In ß-peptide

32 there arc no turn-violating residues, and Zn2+ creates a perfect hairpin tum with an

antiparallel sheet attached. In the case of the isomeric ß-peptide 33, the opposite

configuration of two amino-acid residues prevents proper folding to the ß-peptidic turn.

Still, the Zn2+ ion is powerful enough to pull the terminal residues together, in what one

would call a highly distorted, twisted, and crooked backbone conformation.

117

Based on the results described on the previous pages, the Seebach group has designed a

Zn-finger motif with 16 residues and their analogues and studied their structural

propensities by CD analysis. The CD spectrum of a designed Zn-finger motif in H2O

showed an increased intensity for the negative Cotton effect, shifted towards a

wavelength typical of 14-helical structures upon addition of Zn2+ ions at (pH 7.5). The

titration with ZnCl2 is compatible with the formation of a 1:1 complex (Figure 38), as are

ESI-MS measurements. However, a solution NMR structure is not available yet.

a) b)

H,N

1:1 peptide-Zn2+-complexes

.C02H

H02C

2:1 complex

Figure 38. The possible modes of 1:1 (a and b) and 2:1 (c) Zn2+ ion coniplexation with terminal ßhHis and

ßhCys side chains of ß-peptide.

118

5.4 Conclusion

We have shown that Zn2+ ions arc able to stabilize both helical and hairpin-turn structures

of ß-peptides in H20, even at the low concentrations typical for CD and NMR

measurements, if Cys and His side chains are positioned properly in the peptide

sequence. The free energy of peptide-Zn2+ complex formation, including the strong

affinity of Zn2+ for the S of ß3hCys and for the imidazole JV-atom of ß3hHis, is large

enough to overcome enthalpic and entropie losses of the ß-peptidic backbone and of

solvation. This is not trivial in view of the fact that similar effects have been observed

with a-peptides and proteins, since the stability of ß-peptidic secondary structures is

much larger than that of their a-peptidic counterparts. The local conformational

preferences of the sp3-sp3 C-C-bonds in the ß-peptidie residues have been shown to be

major contributing factors to the stabilization of the corresponding secondary structures,

so that the forces involved in their distortions by Zn2+ ions must be larger than with a-

peptides. With the Zn-stabilized structures in H2O, described herein, are useful for the

construction of tertiary structures of ß-peptides and "ß-proteins" with low molecular

weights. In the future peptides that bind metal ions might also be constructed in a similar

way by exploring coordinating donor atoms present in the side chains of Asp and Glu.

119

5.5 References

[I] G. M. Whitesides, B. Grzybowski, Science 2002, 295, 2418.

[2] T. Kimmerlin, D. Seebach, J. Pept, Res. 2005, 65, 229.

[3] T. L. Raguse, J. R. Lai, P. R. LePlae, S. H. Gellman, Org. Lett. 2001, 3, 3963.

[4] D. S. Daniels, E James Petersson, J. X. Qiu, A. Schepartz, J. Am. Chem. Soc.

2007,129, 1532.

[5] R. P. Cheng, W. F. DeGrado, J. Am. Chem. Soc. 2002,124, 11564.

[6] U. Diederichsen, H. W. Schmitt, Eur. J. Org. Chem. 1998, 827.

[7] U. Diederichsen, H. W. Schmitt, Angew. Chem. Int. Ed. 1998, 37, 302.

[8] P. Chakraborty, U. Diederichsen, Chem. Eur. J. 2005,11, 3207.

[9] A. M. Brueckner, P. Chakraborty, S. H. Gellman, U. Diederichsen, Angew. Chem.

Int. Ed. 2003,42,4395.

[10] F. Q. Ruan, Y. Q. Chen, P. B. Hopkins, J. Am. Chem. Soc. 1990,112, 9403.

[II] B. Imperiali, T. M. Kapoor, Tetrahedron 1993, 49, 3501.

[12] M. R. Ghadiri, A. K. Fernholz, J. Am. Chem. Soc. 1990,112, 9633.

[13] M. R. Ghadiri, C. Choi, J. Am. Chem. Soc. 1990,112, 1630.

[14] M. Gelinsky, H. Vahrenkamp, Eur. J. Inorg. Chem. 2002, 2458.

[15] G. Platt, C. W. Chung, M. S. Searlc, Chem. Commun. 2001, 1162.

[16] I. L. Karle, H. N. Gopi, P. Balaram, Proc. Natl. Acad. Set U.S.A. 2001, 98, 3716.

[17] H. N. Gopi, R. S. Roy, S. R. Raghothama, I. L. Karle, P. Balaram, Helv. Chim.

Acta 2002, 85, 3313.

[18] J. M. Langenhan, I. A. Guzei, S. H. Gellman, Angew. Chem. Int. Ed. 2003, 42,

2402.

[19] Y. J. Chung, B. R. Huck, L. A. Christianson, H. E. Stanger, S. Krauthauser, D. R.

Powell, S. H. Gellman, J. Am. Chem. Soc. 2000,122, 3995.

[20] S. Krauthauser, L. A. Christianson, D. R. Powell, S. H. Gellman, J. Am. Chem.

Soc. 1997,779,11719.

[21] D. Seebach, S. Abele, K. Gademann, B. Jaun, Angew. Chem. Int. Ed. 1999, 38,

1595.

[22] X. Daura, K. Gademann, H. Schaefer, B. Jaun, D. Seebach, W. F. van Gunsteren,

J. Am. Chem. Soc. 2001,123, 2393.

120

[23] S. A. Hart, A. B. F. Bahadoor, E. E. Matthews, X. J. Qiu, A. Schepartz, J. Am.

Chem. Soc. 2003, 125, 4022.

[24] F. Rossi, G. Lelais, D. Seebach, Helv. Chim. Acta 2003, 86, 2653.

[25] D. Seebach, K. Gademann, J. V. Schreiber, J. L. Matthews, T. Hinteimann, B.

Jaun, L. Oberer, U. Hommel, H. Widmer,Helv. Chim. Acta 1997, 80, 2033.

[26] G. Lelais, D. Seebach, B. Jaun, R. I. Mathad, O. Flögel, F. Rossi, M. Campo, A.

Wortmann, Helv. Chim. Acta 2006, SP, 361.

[27] D. Seebach, A. K. Beck, D. J. Bierbaum, Chem. & Biodivers. 2004, 7,1111.

121

NMR-Solution Structure Investigation of

g Cyclic ß-Tetrapeptides as Possible RGD

and Somatostatin Analogues

6.1 Introduction

RGD-Peptides

The Arg-Gly-Asp (RGD) motif present in many extracellular matrix proteins such as

fibronectin, vitronectin, osteopontin, collagens, thrombospondin and fibrinogen serves as

an essential recognition site for integrins. Integrins are the main cell surface receptors,

mediating cell adhesion to extracellular proteins and bind specifically to the Arg-Gly-Asp

(RGD) motif. Interaction between the RGD sequence motif and receptors is responsible

for many biological functions such as cell adhesion, signaling, platelet aggregation,

apoptosis (natural cell death), inhibition of angiogenesis and fertilization [1-5]. These

properties render them attractive targets for drugs, especially for treating cancer.

In fact, short synthetic peptides containing an RGD sequence can effectively block

receptor-ligand interactions. Based on this knowledge, many RGD peptides and

peptidomimetics have been studied to develop therapeutic agents for treating thrombosis

and tumor metastasis [6]. The mechanism of action of RGD compounds is to inhibit the

cell adehesion to the extracellular matrix proteins via binding of the RGD sequence to the

receptors on the cell surface (Figure 39) [7]. Thus, the development of the RGD-

dependent ligand recognition by receptors has been the major focus of many research

groups.

A wide range of short amino acid sequences have been investigated in order to explore

the relationship between structure and biological activity, as well as the selectivity for

various receptors [8, 9]. Structural studies of RGD-containing peptides and proteins

suggest that the RGD motif usually occurs at the apex of a solvent exposed loop, a

characteristic of prime importance for its biological activity. The ability of integrins to

122

distinguish between different RGD containing extracellular proteins is partially due to the

variety ofRGD conformations.

Figure 39. The schematic presentation of RGD motif mediated activation of receptors in the extracellular

environment.

Recently, the solid state structure of the extracellular segment of integrin avß3 has been

reported for the free form, as well as for a complex with a cyclic pentapeptide containing

the RGD sequence flO] {Figure 40). The crystal structure of avß3 in complex with a

ligand-mimetic peptide provided a first glimpse as to how integrins recognize the RGD

motif Arginine and aspartic acid side chains interact by forming salt bridges with the

integrin core, one of the aspartic acid carboxylate oxygens coordinates with Mn2+. The

groove in which the ligand resides is rather shallow, and large parts of the ligand make

virtually no contact with the protein. It seems as if the Asp and Arg residues of the RGD

ligand act as an electrostatic clamp, attaching themselves to charged regions of the

protein; usually ligands tend to attach to hydrophobic pockets with large interfaces

determined by van der Waals interactions. An unusual feature of the ligand-binding site

is that it contains little hydrophobic space. The central glycine residue lies directly on the

integrin surface and might also add to the stability of the complex. The main interactions

123

are between the positively charged arginine and negatively charged side chains in the a

subunit and between the anionic aspartic acid and the cation in the metal ion-dependent

adhesion site (MIDAS) ofthe ß subunit.

These crystal structures, in combination with NMR data and molecular dynamics studies,

provide a deeper insight into the mechanism of integrin-mediated signal transduction [11,

12]. However, the crystal structures offer little insight into how integrins achieve

specificity and high affinity in recognizing physiological protein ligands. For example,

the bound mimetic cyclic peptide ligand contacts a limited set of residues in aV and ß3

subunits, burying only ca, 45% ofthe total surface area ofthe peptide [13]. The contact is

made exclusively through the essential RGD sequence.

Figure 40. The crystal structure of avß3 receptor in complex with a cyclic RGD-pentapeptide analogue.The guanidine group of Arg is stabilized by salt-bridge interactions and one of the carboxylate oxygenatoms is m contact with Mn2+cation, whereas second carboxyl oxygen atom is involved in H-bonding.

However, in physiological integrin-ligand interactions, many more residues of the a and

ß subunits are involved in ligand recognition and residues outside the RGD motif in the

ligand make essential contributions to the complexation.

The docking studies on avp3 integrin ligands have shown that the main interactions are

between the positively charged arginine and the a-subunit and between the anionic

aspartic acid and the ß-subunit and that selectivity between different subunits is achieved

by the local conformation within the RGD sequence [14]. Despite numerous studies

reported in the literature, achieving ligand selectivity toward different integrins is still a

challenging task. Therefore, it is important to study the conformational behavior of

124

peptides containing the RGD sequence with different substitution patterns as ligands in

order to better understand binding specificity and affinity to receptors.

6.2 Cyclic RGD-Peptides

To date, numerous peptides (linear or cyclic) and peptidomimetics that contain or mimic

RGD have been designed to modulate functions of receptors [15]. Constraining highly

flexible linear peptides by cyclization is one of the mostly widely used approaches to

define the desired conformation of peptides for particular applications. The cyclic

systems offer the possibility of maintain in the RGD sequence in a specific conformation,

and equilibria between different conformations are reduced. Moreover, cyclic peptides

are more specific than their linear counterparts, and have the added advantage that they

are resistant to proteolysis. The metabolic stability of both linear and cyclic RGD

peptides as a function ofpH and buffer concentration has been compared [16]. The latter

study revealed that the cyclic peptide is 30-fold more stable than its linear counterpart at

pH 7. It has been clearly demonstrated that the increase in stability of the cyclic peptide

compared with the linear one is due to decreased structural flexibility imposed by the ring

[17].

Cyclic RGD peptides have been developed for various purposes: fibrinogen receptor

antagonists, selective avß3 integrin antagonists for treatments ofhuman tumor metastasis

and tumor-induced angiogenesis, phagocytosis of cells undergoing apoptosis as well as

osteoporosis, diabetic retinopathy and acute renal failure. A number of therapeutic

candidates such as antibodies [18], peptidomimetics, and cyclic peptides [19], have been

clinically evaluated and shown to successfully modulate avp3-related processes. The

absolute configuration of specific amino acids in the cyclic structures can have a great

influence on the biological activity. For example the difference between l- and D-

phenylalanine in the cyclic peptides c(RGDFV) and c(RGDfV) was shown to be 1000-

fold. [20]. Each backbone amide bond of c(RGDfV) was Af-methylated giving a series of

five monomethylated cyclic pentapeptides, represents five classes of cyclic peptides with

representative RGD arrangements and different integrin binding specificities. The effect

of./V-methylation on biological activity revealed that the pentapeptide (RGDf-N(Me)V) is

one of the most active and selective avß3 integrin antagonists known so far [21]. Studies

on retro-inverse peptides, showed that the backbone conformation, the side chain

125

topology of the peptides and the amide bond direction lead to drastically different

inhibitory activities with regard to the <xvß3 receptor [22]. Disulfide bonds at cysteine

residues in the structure of cyclic RGD peptides can also alter the binding activity of the

peptides [23]. A potent and selective peptide antagonist of the avß3 receptor, a

afunctional chimeric molecule containing a cyclic RGD motif and a sequence

corresponding to the echistatin C-terminal tail connected by a linker, has been reported.

Several RGD-based antithrombic agents have been developed. For example, Aggrastat,

an RGD peptidomimetic, has been used clinically for treating thrombosis [24, 25],

6.3 Design of ß-Peptidic RGD-Peptides

Many structural features are thought to be important for the ligand-integrin complexation.

A number of conformations including ß-turns for Arg-Gly or Gly-Asp sequences have

been suggested. In particular, emphasis has been given to the presence of a well defined

Gly-Asp ß-turn as a prerequisite for integrin binding, while the selectivity of the ligands

has been correlated with the distances between either the Cß atoms or the oppositely

charged centers of Arg and Asp residues. However, little is known about the influence of

residues flanking the RGD sequence on either the affinity or specificity of the

interactions.

Recently, peptides composed exclusively of ß-amino acids have been shown to adopt a

variety of stable secondary structures such as helices, sheets and turns, and to be stable

towards proteolytic enzymes [26]. These properties prompted us to investigate small ß-

peptides containing RGD sequence as mimics of RGD motifs, and the influence of ß-

amino acids on the structure of cyclic tetrapeptides as well as their physiological activity.

126

a) somatostatin analogue b)

RGD-turn mimic

ß-turn mimic

Figure 4L a) The NMR-solution structure of a ß3-tetrapeptide. b) The designed cyclic ß-tetrapeptidc with RGD side chains.

The objective of this project was to synthesize cyclic RGD mimetics by incorporating ß2-

and ß -amino acids into small cyclic peptides and to analyze conformations in solution

by NMR-spectroscopy, in order to be able to correlate peptide backbone conformation

and side chain orientation with biological activity. The solution structure of cyclic ß3-

tetrapeptides is known to contain intra-molecular H-bonds bisecting the cycle into 10-

and 12-membercd hydrogen bonded rings (Figure 41a). As demonstrated previously by

our group [27], the use of the ß2/ß3 unit in a ß-peptide leads to a 10-membered H-bonded

ring and the ß3/ß2 motif to formation of a 12-membered ring. The 10-membered turn is

comparable to a ß-turn of an a-peptide both in size and orientation ofthe side chains. We

thus designed cyclic ß-tetrapeptides1 attaching the side chains of Arg and Asp to this

ß3/ß2 motif, in order to have the correct number of carbon atoms to mimic the Gly

residue between Asp and Arg, The other two amino acids were chosen as ß2- and ß3-

amino acids in order to be able to evaluate the influence of the side chains on the

interaction with the active site (Figure 42).

1All the cyclic ß-tetrapeptides, and mixed cyclic a/ß-peptides discussed in this chapter have been

synthesized by Dr. Estelle Dubost, a postdoctoral coworker at ETH Zürich.

127

V—NH HN—<

0=\ ^—NH HN--/OH I I V_

34

N

^-NH2H2N

0=^OH

>—-NH HN

^^NH HN-

H2N^ ^N

H,N^ ^N

0-^^OH

36 37 HO

Figure 42. Cyclic ß-tetrapeptides as RGD analogues with various substitutions synthesized on SPPS.

6.4 Results and Discussion

Since small peptides possess conformational flexibility in solution, the determination of

the bioactive conformation necessary for structure-based rational drug design is often

difficult. On the other hand, cyclic peptides are constrained and somewhat less flexible

than their linear counterparts. Thus, for cyclic peptides the structures in solution are more

likely to correspond to the bioactive conformations and may lead to better insight into

structure-activity relationships (SAR) [28].

The cyclic ß-peptides 34-37 have been subjected to detailed NMR-structural analysis in

CD3OH. The complete assignment of proton and carbon resonances of cyclic peptides

was accomplished by the standard strategy, combining 2D ('H-'H) COSY, TOCSY and

ROESY spectra. The information deduced from these experiments was interpreted in

terms of peptide conformations in solution. Sequential assignment was accomplished by

through-bond connectivities from heteronuclear multi-bond correlation (HMBC) spectra.

128

Carbonyl resonances were also assigned using through bond long-range correlations.

Additional information regarding the three-dimensional structure of the cyclic ß-peptides

was obtained by ROESY experiments with xm = 150 and 300 ms.

The assigned ROESY cross peaks were integrated and subsequently converted into

qualitative inter-proton distances (distance constraints). The proton distances were

obtained according to the isolated two-spin approximation from volume integrals of

(ROESY) spectra. The 3J(NH-Hß) coupling constants extracted from one-dimensional

spectra were converted into dihedral restraints. Chemical shift data and determined

coupling constants are listed in (Tables 22-25),

Table 22. 'H-NMR-Chemical shifts for ß-tetrapeptide 34 in CD3OII

ß-amino NH H2-C(a) H-C(ß) H-C(y) H-C(S) Me-C(s)acid 3JHNHfl Me-C(y) Me-C(8)

H2-C(y)

1.65 3.18/3.27

ß3hAla' 7.88 2.43/2.26 4.30

8,44

1.19

ß3hArg2 7.91 2.48/2.21 4.01

7.74

1.59

ß2hAsp3 8.27 2.79 3.43/3.08

ß2hAal4 8.05 2.42 3.42/3.15 1.09

Table 23. 'H-NMR-Chemical shifts for ß-tetrapeptide 35 in CD3OH

ß-aminoacid

NH H2-C(cx) H-C(ß)^JHNHß

H-C(v)

Mc-C(y)

H2-C(Y)

H-C(ô)

Me-C(S)

Me-C(e)

ß3hVal' 7.61 2.43/2.41 3.99

9.61

1.47 0.58

ß3hArg2 7.92 2.47/2.12 4.02

7,41

1.56 1.66 3.18/3.29

ß2hAsp3 8.40 2.74 3.43/2.99 2.50

ß2hPhe4 8.14 2.66 3.62/3.14 2.84

129

Table 24. 'H-NMR-Chemical shifts for ß-tetrapeptide 36 in CD3OH

ß-amino NH H2-C(a) H-C(ß) H-C(y) HC(8) Me-C(s)acid 3JHNHß Me-C(y) Me-C(5)

H2-C(Y)

ß3hAsp' 7.79 2.18/2.50 4.47 2.52

ß3hArg2 8.22 2.13/2.43 3.43 1.54 1.65 3.23

ß3hPhe3 8.29 2.32 4 3 2.73/2.94

7.26

ß3hPhe4 7.74 2.21 4.4 2.64.2.82

8.44

Table 25. 'H-NMR-Chemical shifts for ß-tetrapeptide 37 in CD3OH

ß-aminoacid

NH HrC(a) H-C(ß)

*JHNHßH-C(y)

Me-C(y)

HrC(Y)

H-C(S)

Me-C(ö)

Me-C(e)

ß^he1 7.88 2.21/2.48 4.46

8.66

2.81

ß3hGlu2 7.84 2.17/2.47 3.91

7.48

1.8 2.37

ß2hArg3 8.20 2.56 3.10/3.45 1.47 1.59 3.1

ß2hAla4 8.12 2.46 3.19/3.32 0.97

The molecular dynamics based simulated annealing calculations were performed with

NMR derived distances and dihedral restraints. The calculations revealed that all the

cyclic peptides adopt well defined backbone conformations as depicted in Figure 43. The

conformation of the cyclo-ß-tetrapeptidcs is characterized by an intramolecular H-bond

dividing the 16-atom backbone into a 10-membered and a 12-membered H-bonded ring.

The 10-membered ring is mimicking a natural a-peptidic ß-turn, and is very similar to

the ring observed in the 12/10/12-helix and in a hairpin of a ß2,2-peptide [29, 30].

130

36 37

Figure 43. An overlay ofNMR-solution structures of cyclic ß-tetrapcptides showing the relative

orientation ofArg and Asp side chains in MeOH.

The Arg and Asp or Glu ß-peptidic backbones are part of thel2-membered hydrogen

bonded ring, whereas the flanking residues constitute the 10-membcred ring in the ß-

peptides 34-35 and 37. The side chains of Arg and Asp point in opposite directions. The

distance (Cy, Cy) between these two side chains within the RGD sequence is ca. 7 Â,

comparable to the values (ca. 6 Â) found in other studies. All the side chains of amino

acids occupy nearly lateral positions on the macrocyclic ring, which results in nearly a

flat ring shape. In addition, the other two amide groups are pointing up and down with

respect to the ring plane, reducing the net dipole moment ofthe molecule.

131

6.5 Cyclic ß-Tetrapeptides as Somtostatin Analogues

Introduction

Somatostatin consists of a family of cyclopeptides that are mainly produced by normal

endocrine, gastrointestinal, immune and neural cells, as well as by certain tumors [31].

Highly active somatostatin peptide hormones such as SRIF-14 and SRIF-28 have been

isolated and characterized. In mammals, these peptides originate from a pro-hormone

called pro-somatosatin (proSS), which can inhibit secretion. Somatotropin-release

inhibitory factors (SRIF) inhibit the release of many hormones including the growth

hormone, glucagon, insulin, and gastrin. Furthermore, SRIF acts as a neurotransmitter in

the central nervous system and peripheral tissue, where it modulates several processes,

such as smooth muscle motor activity and the release of other neurotransmitters [32].

Consequently, analogues of somatostatin emerged as interesting tools for the treatment of

disorders linked to those physiological functions including diabetes, cancer, rheumatoid

arthritis, and Alzheimer's disease.

There are five different somatostatin receptors (sst].5) identified and characterized as G-

protein-coupled transmembrane receptors, which trigger multiple trans-membrane

signaling processes [33]. Binding of SRIF to receptors leads to an activation of the G-

protein which then modulates the activity of several key enzymes such as adenylyl

cyclase or phosphotyrosine phosphatases. In particular, sst2-selective agonists have

emerged as useful candidates for such treatments [34-36],

To date, the physiological functions of only two receptors sst2 (mediation of the release

of GH) and sst5 (inhibition of the release of insulin), are known in detail. The SRIF

receptor subtype ssti can mediate antiproliferative effects. The main receptor subtype is

sst2, which mediates both antisecretory and antiproliferative action. The ligation of sst2

inhibits secretion of GH, glucagon, gastrin, and gastric acid and inhibits ion secretion in

the colon. The sst2 type is expressed in a number of cancer cell types including small-cell

lung and gastro-enteropancreatic tumors and mediates tumor cell growth inhibition.

Subtype sst3 was reported to mediate antiproliferative and proapoptotic effects. The role

of sst4, which is expressed at high levels in the lung and brain, is not well understood,

while ssts has been shown to mediate inhibition of GH and cell proliferation.

132

Because of its wide range of physiological functions, somatostatin may play an important

role in the treatment of numerous human diseases. However, the clinical use of

somatostatin has been hampered by some disadvantages of the native hormone, such as

its very short half-life in blood (< 3 min) and its lack of selectivity. To overcome these

problems, several peptidic and nonpeptidic somatostatin analogues have been reported in

the literature to increase receptor subtype selectivity and metabolic stability. The first

potent cyclic hexapeptide somatostatin analogue, cyclo(Prol-Phe2-D-Trp3-Lys4-Thr5-

Phe6), was synthesized by Veber et al [37], This analogue showed activity in inhibiting

growth hormone, insulin, and glucagons and high selectivity for the sst2 receptor subtype.

Later, a number of other active cyclic peptides have been reported. However, their

clinical development has been unsuccessful due to poor metabolic stability and side-

effects. To date, only three octapeptide SRIF analogues, octreotide (Sandostatin®),

lanreotide, somatuline and octastatin, are in clinical studies and/or use [38, 39]. All three

octapeptides bind most effectively to only two ofthe five somatostatin receptor subtypes,

sst2 and ssts. Octreotide shows slightly higher affinity for sst2 than for ssts, whereas both

lanreotide and octastatin, which are Tyr3/Val6-containing somatostatin analogues, bind

most effectively to ssts. They are used in the treatment of acromegaly and certain gastro-

entero-pancreatic tumors. Their elimination half-life from the blood serum (90 min) is

still rather short. It is therefore of great interest to find non a-peptidic analogues that

mimic natural SRIF or octreotide. For these reasons, we chose somatostatin as the target

to demonstrate the potential of ß-peptides as mimetics of natural a-peptide hormones.

133

Trp8 Lys9I I

Phc7 Thr10

Phe6 Phe11

Asn5 Thr12

Cys4 Ser13

Cys3 Cys14l2

Gly2I,

R—Ala1

Somatostatin

H-.N

)iini, N..

H / / \

Octreotide

OH

CH2OH

ÖH

ß-tetrapeptide

SRIF-14 R=H

SRIF-28 R=H-Ser-Ala-Asn-Ser-Asn-Pro-Ala-

Met-Ala-Pro-Arg-GIu-Arg-Lys

Figure 44. Naturally occurring somatostatin SRIF-14, SRIF-28 and somatostatin derived peptide

analogoues: octreotide and a cyclic ß-tetrapeptide.

Given that the ß-peptidic secondary structures can be readily designed, and ß-peptides

are highly stable to proteolytic degradation, this class of peptides has great potential in

medicinal chemistry. Our group has demonstrated that simple, designed, low-molecular

weight ß-peptides, such as linear ß2/ß3-dipeptides, y4-dipeptides, a/ß3-tetrapeptides, as

well as a cyclic ß3-tctrapeptide arc capable of binding to SRIF receptors with high

affinity and selectivity [40]. Earlier, the solid-state conformation of three cyclo-ß-

tetrapeptides was determined by powder X-ray diffraction [41, 42]. They were found to

form tubular stacks with an infinite network of pleated-sheet-type H-bonds (so called

peptide nanotubes). In addition, the first solution structure of a ß-peptidic somatostatin

analogue, a cyclic ß-tetrapeptide c(ß3Phe-ß3Trp-ß3Lys-ß3Thr) has been reported by our

group [43]. Unlike the solid state conformation of ß3-tctrapeptides, the solution structure

of this compound revealed that there is an intramolecular-hydrogen bond between the

Trp C=0 and the Lys NH groups, placing the corresponding side chains close to each

other as required for the optimum interaction with receptors. The affinity of this peptide

to the somatostatin receptors was, however, in the micromolar range and not in the

nanomolar range as is the case with octreotide.

134

6.6 Design and Structural Aspects of Somatostatin Analogues

Extensive structure-activity relationship studies have established that the primary

pharmacophore consists of a ß-turn spanning the Trp-Lys fragment which is necessary

for biological activity. The tetradecapeptide SRIF-14, one ofthe widely distributed active

forms of somatostatin, is believed to adopt a two-stranded ß-sheet conformation induced

by a ß-tum encompassing Phe7-Trp8-Lys9-Thr10, and the disulfide bridge between Cys3

and Cys14, respectively (Figure 44). The conformation is further stabilized by the

transannular H-bonding pattern typical for antiparallel sheet structures. Furthermore, it

was shown that the Trp and Lys residues are essential for high affinity binding to ssti-

ssts, whereas the Phe and Thr residues can undergo minor substitutions provided that

lipophilic residues are next to Trp and hydrophilic side chains next to Lys. However, it

was shown that the entire native hormone is not necessary for expression of the full

activity.

The structure of octreotide has been studied both in the solid state and in solution. The

NMR studies in water and in DMSO-cfö solutions showed that octreotide adopts a

predominant antiparallel ß-sheet conformation characterized by a ß-turn comprising D-

Trp and Lys residues [44]. Similar secondary structural features were also found for

octreotide in a water/MeOH mixture [45]. In the solid state, sandostatin has a ß-shect

structure. The crystal also contained two other conformers, in which the arrangement of

the residues D-Phe'-Lys5 is similar but the C-terminal tripeptide segment showed a

helical fold [46]. Computer-aided modeling studies revealed similarities between the

powder X-ray structure of cyclo-ß-tetrapeptides and type ß II' rums of a-peptides. Thus,

we have attached the side chains necessary for biological activity to the ß-peptidic

backbone to imitate the natural a-peptidic structure (Figure 44).

It is known that incorporating a ß2/ß3 motif into a ß-peptide induces a turn stabilized by

10 membered H-bonded ring which is comparable both in size and orientation to a ß-turn

in a-peptides, except that the central amide bond is reversed. The strong folding

preference of ß-peptides should, in principle, enforce a specific type of backbone

conformation which should lead to high affinity for receptors. Thus, the use of a ß2/ß3-

135

turn segment with Tip and Lys side chains must lead to a better mimic of the natural a-

peptide hormone. Further, its introduction into a cyclic peptide can lead to reduced

flexibility of the "hgand".

ß3-amino acid

NH2

N-tcrminus *

folding E/^9° ^NH

0 NH

C-terminus>c

ß2-amino acid

Figure 45. ß2/ß3-Segment folding to a ß-peptidic tum is stabilized by a 10-membered H-bonded ring.

In search of potential SRIF-type ligands, based on the NMR-solution structure of a cyclic

ß3-tetrapeptidc which showed an intramolecular-hydrogen bond, we have designed,

synthesized, and investigated the solution structures of cyclic ß-tetrapeptides 38-40. They

contain the ß2hTrp/ß3hLys motif, to stabilize a turn structure {Figure 45), and ß3hThr and

ß hPhe residues at the two other positions. We also examined the effect of replacing

ß3hThr and ß2hPhe by other ß-amino acids, in order to correlate structure with the

activity of somatostatin analogues. Furthermore, a cyclic mixed peptide 41 containing a

single a-amino acid in the ß-peptide sequence has been examined to study the influence

of a-amino acid on the conformation ofa cyclic tetrapeptide (Figure 41a).

136

V-NH HN 7—NH HN

^NH HN ^NH HN

"°-ZXS Xy o->jai\>NH,

40 41

NH,

Figure 46. Cyclic ß-lctrapeptides as potential somatostatin analogues with various substitution patterns.

6.7 Results and Discussion

A detailed 2D-NMR spectroscopic study was undertaken to obtain high-resolution data

on the conformational preferences of the cyclic ß-tetrapeptides 38-41 in methanol. The

complete assignment of all !H resonances was accomplished by a combination of DQF-

COSY and TOCSY with the standard procedure described earlier. The sequence specific

assignment was further confirmed by heteronuclear correlation (HMBC) experiments. All

the chemical shifts and 3J(NH-Hß) values, as extracted from 1-D spectra, are complied in

{Tables 26-28), Analysis of ^-spectra provided an indication that the Trp and Lys side

chains are in close proximity, a phenomenon which is commonly observed for

somatostatin and analogues. This was established by the upfield chemical shift observed

for the ÔCH2 resonances of the Lys side chains caused by the aromatic anisotropy of the

indole ring of the ßhTrp side chain. Such shifts have previously been correlated to the

activity of SRIF analogues [47].

137

Table 26. 'H-NMR-Chemical shifts for ß-tetrapeptide 38 in CD3OH

ß-amino NH HrC(cc) H-C(ß) H-C(y) H-C(S) Me-C(s)acid 3JHNHß Me-C(y) Me-C(5)

H2-C(T)

ß2hTrp' 7.99 2.36 3.32/3.55 2.88/2.98

ß3hLys2 7.41 2.21/2.38 3.92 1.12/1.27 0.82/0.89 1.37

8.6

ßhGly3 7.75 2.35 3.36/3.44

ßhGly4 7.95 2.34 3.43/3.48

Table 27. 'H-NMR-Chemical shifts for ß-tetrapeptide 39 in CD3OH

ß-amino NH H2-C(a) H-C(ß) H-C(y) H-C(8) Me-C(s)acid 3JHNHß Me-C(y) Me-C(5)

HrC(Y)

ß2hlrp' 8.1 2.8 3.28/3.59 2.84/3.00

ß'nLys2 7.66 2.18/2.29 4.03 1.11/1.28 0.76 1.37

ßhGly3 7.96 2.36 3.37

ßhGly4 8.09 2.41 3.42

Table 28. 'H-NMR-Chemical shifts for ß-tetrapeptide 41 in CD3OH

ß-amino NH H2-C(a) H-C(ß) H-C(y) H-C(5) Me-C(s)acid 3JHNHß Me-C(y) Me-C(ö)

H2-C(y)

ß2hPhe' 7.70 2.35/2.49 4.27 2.80/2.84

ß'hAla2 7.73 2.36 3.12/3.48 1.02

Lys3 8.03 4.23 1.57/1.85 1.37 1.60 2.87

ßhTrp4 7.80 2.50 4.48 2.94/3.05

138

Although the conformation of a ligand bound to a receptor could be different from the

conformation in solution, such shifts often provide the first clue of the affinity.

Additional structural information was obtained from ROESY experiments with xm = 300

ms. The NOE cross peaks were integrated and subsequently converted into distance

restraints with the two-spin approximation. The structures of the cyclic peptides were

calculated by MD-simulated annealing employing the XPLOR protocol with NOE-

derived distance restraints and dihedral constraints derived from 3J coupling constants.

Each calculation was started with an energy-minimized starting structure. The calculation

produced 30 structures of lowest energy without violation of experimental constraints are

shown in Figure 47.

38 39

40 41

Figure 47. The NMR-solution structures of cyclic tetrapeptides 38-41 analyzed in MeOH. Overlay of low

energy structures calculated from simulated annealing calculations.

139

The conformation of cyclic ß-tctrapeptides 39 and 402 showed well defined backbone

conformations with 10- and 12-membered intra-molecular H-bonded rings, dividing the

16-atom-backbone. All the amide bonds adopt trans arrangements. The two amide bonds

not involved in hydrogen-bonding are pointing in opposite directions, perpendicular to

the average plane of the ring. As expected, the ß2/ß3-unit forms a 10-membered H-

bonded ring, which mimics the natural opeptidc hormone ß-tum and which is similar to

the ring observed in the 12/10-helix. The position of ô-CH2 (y-CH2 in Lys) for each

peptide shows the charctcristic upfield shift, indicating the proximity of Trp and Lys side

chains. These side chains occupy lateral positions on the ring separated by a distance of

ca. 5.5 Â. The peptide 38 containing a (i?)-ß2hTrp residue does not show any intra

molecular H-bonds and is structurally less well defined. The backbone-conformations of

peptides 38 and 39 are not rigid when compared to the other cyclic ß-peptides. This

flexibility can be attributed to the presence of two ßhGly moieties in the cyclic peptides

as unsubstituted ßhGly (no significant conformational constraints) is known to adopt two

conformations3. The mixed peptide 41 with (/?)-ß2-Ala and Lys residues forms a well

defined backbone conformation, again with no observable intra molecular H-bonds, but a

turn around the ß2/a-segment. The overall shape of this molecule is not flat but boat-

shaped, with all the amide planes in the trans geometry. The two adjacent amide groups

point in one direction and the other two in nearly the opposite direction, similar to what

was observed in the solid state structure of (i?,/?,,S,,S)-cyclo-tetra-ß-homoalanine

determined from powder X-ray diffraction data. Also, the characteristic upfield shift for

yCH2 of the Lys side chain was not observed in this mixed peptide, implying that the side

chains of Trp and Lys are not in close proximity.

The NMR-structural investigation of 40 was cairied out at Novartis, Basel.3The search for ßhGly in linear and cyclic peptides in X-ray crystal structures revealed the ßhGly residue

exists in folded and extended forms see Diss. No, 14677 ETH Zürich.

140

6.8 Conclusion

The amino acid residues of Phe, Trp, Lys, and Thr, which comprise a ß-turn in

somatostatin, are necessary for biological activity, with residues Tyr, Lys being essential

to maintain the activity by means of receptor-ligand interactions. The NMR

investigations have shown that cyclic ß-tetrapeptides 38-41 adopt well defined structures

in MeOH solution. The configuration of the ß2-amino acid influences the stability of the

conformation. The use of a ß2/ß3-segment, a turn inducing unit, is capable of stabilizing

cyclic tetrapeptides into one main conformation in solution when compared to an a//-ß3-

cyclotetrapeptide. The study also demonstrates that the insertion of single a a-amino acid

residue into small cyclic ß-peptides can lead to higher conformational homogeneity.

Peptide hssti hssti hsst3 hsst.4 hssts

SRIF-149.08+/-0.07

0.83 nM

10.06+/-0.07

0.087 nM

9.67+/-10.07

0.21 nM

8.39+/-0.28

4.07 nM

9.01+/-0.24

0.97 nM

Octreotide6.65

224 nM

9.19

0.645 nM

7.88

13.2 nM

6.44

398 nM

7.17

67.6 nM

38 <5 <5 <54.94+/-0.07

11.5 uM<5

39 <5

5.05

8.9 pM<5

5.66+/-0.03

2.2 uM<5

40 <5 <5 <5

5.34+/-0.06

4.6 uM<5

414.95

11.92uM

Table4 29. Affinity of peptides for human recombinant SRIF receptors, expressed in CCL39

cells. The pKo and Kd values +/- SEM (standard error method) determined from at least three

experiments. The data for SRIF-14 and octreotide taken from [48].

4Affinity data is reproduced from a report of Dr. Estelle Dubost, a postdoctoral researcher, ETH Zurich.

141

For the design of cyclic mixed a/ß-peptides, the a-amino-acid residue may be a useful

complement to the ß3-amino acids in the design of conformationally homogeneous cyclic

peptides with predictable structures. Thus, the higher degree of conformational rigidity of

mixed cyclic a/ß-peptides may lead to promising new scaffolds for modulation of the

properties and functions of peptides.

The binding data in Table 29 shows that somatostatin-14 binds with good affinity for

receptors ssti-s, whereas the octreotide is relatively selective and has affinity for ssti and

sstt with lower affinity for sst2. The binding affinity data establishes that cylic ß-

tetrapeptides presented here are highly selective for receptor sst4 with although modest

affinity (in the uM range) compared to the nM range of octreotide. The modest affinity

values for 38 and 39 suggest that Phe and Thr residues do not play an important role in

binding affinity. However, peptide 41 containing a single a-amino acid is selective and

shows modest affinity for the receptor sst5 which shows the importance of

conformational stability of a peptide for efficient binding.

142

6.9 References

[I] A. E. Aplin, A. K. Howe, R. L. Juliano, Curr. Opin. Cell Biol. 1999,11, 737.

[2] E. Ruoslahti, Annu. Rev. Cell Dev. Biol. 1996,12, 697.

[3] R F. Nicosia, E. Bonanno, Am. J. Pathol. 1991,138, 829.

[4] R. A. Bronson, F. Fusi, Biol. Reprod. 1990, 43, 1019.

[5] M. Ginsberg, M. D. Pierschbacher, E. Ruoslahti, G. Margueric, E. Plow, J. Biol.

Chem. 1985,260,3931.

[6] Y. Tsuchiya, S. Sawada, K. Tsukada, I. Saik, Int. J. Oncol. 2002, 20, 319.

[7] P. Schaffner, M. M. Dard, Cell. Mol. Life Sei. 2003, 60, 119.

[8] S. Oishi, K. Miyamoto, A. Niida, M. Yamamoto, K. Ajito, H. Tamamura, A.

Otaka, Y. Kuroda, A. Asai, N. Fujii, Tetrahedron 2006, 62, 1416.

[9] R. Haubner, R. Gratias, B. Diefenbach, S. L. Goodman, A. Jonczyk, H. Kessler,

J. Am. Chem. Soc. 1996,118, 7461.

[10] J. P. Xiong, T. Stehle, R. G. Zhang, A. Joachimiak, M. Frech, S. L. Goodman, M.

A. Aranout, Science 2002, 296, 151.

[II] S. J. Bogdanowich-Knipp, D. S. S. Jois, T. J. Siahaan, J. Pept. Res. 1999, 53, 523.

[12] G. V. Nikiforovieh, K. E. Kover, W. J. Zhang, G. R. Marshall, J. Am. Chem. Soc.

2000,122, 3262.

[13] J. Takagi, Biochem. Soc. Trans. 2004, 32, 403.

[14] L. Marinelli, A. Lavecchia, K. E. Gottschalk, E, Novellino, H. Kessler, J. Med.

Chem. 2003, 46, 4393.

[15] K. E. Gottschalk, H. Kessler, Angew. Chem. Int. Ed. 2002, 41, 3161.

[16] S. J. Bogdanowich-Knipp, S. Chakrabarti, T. D. Williams, R. K. Dillmall, T. J.

Siahaan, J. Pept. Res. 1999, 53, 530.

[17] R. Haubner, W. Schmitt, G. Hoelzemann, S. L. Goodman, A. Jonczyk, H.

Kessler, J. ,4ml Chem. Soc. 1996,118, 7881.

[18] J. C. Gutheil, T. N. Campbell, P. R. Pierce, J. D. Watkins, W. D. Huse, D. J.

Bodkin, D. A. Cheresh, Clin. Cancer Res. 2000, 6, 3056.

[19] F. Mitjans, T. Meyer, C. Fittschen, S. Goodman, A, Jonczyk, J. F. Marshall, G.

Reyes, J. Piulats, Int. J. Cancer 2000, 87, 716.

143

[20] K. Mueller, D, Obrecht, A. Knierzinger, C. Stankovic, C. Spiegler, W.Bannwarth,

Approches in Lead Finding andLead Optimization, Hoffmann-La Roche, Basel,

1992.

[21] M. A. Dechantsreiter, E. Planker, B. Matha, E. Lohof, G. Holzemann, A.

Jonczyk, S. L. Goodman, H. Kessler, J. Med. Chem. 1999, 42, 3033.

[22] J. Wermuth, S. L. Goodman, A. Jonczyk, H. Kessler, J. Am. Chem. Soc. 1997,

119, 1328.

[23] N. Assa-Munt, X. Jia, P. Laakkonen, E. Ruoslahti, Biochemistry 2001, 40, 2373.

[24] N. S. Nicholson, N. A. Abood, S. G. Panzer-Knodle, L. G. Frederick, J. D. Page,

A. K. Salyers, 0. D. Suleymanov, J. A. Szalony, B. B. Taite, R. J. Anders, Med.

Res. Rev. 2001,27,211.

[25] W. B. Batchelor, T. R. Tolleson, Y. Huang, R. L. Larsen, R. M. Mantell, P.

Dillard, M. Davidian, D. W. Zhang, W. J. Cantor, M. H. Sketch, E. M. Ohman, J.

P. Zidar, D. Gretler, P. M. DiBattiste, J. E. Tcheng, R. M. Califf, R. A.

Harrington, Circulation 2002,106, 1470.

[26] H. Wiegand, B. Wirz, A. Schweitzer, G. P. Camenisch, M. I. R. Perez, G. Gross,

R. Woessner, R. Voges, P. I. Arvidsson, J. Frackenpohl, D. Seebach, Biopharm,

Drug Dispos. 2002, 23, 251.

[27] D. Seebach, S. Abele, K. Gademann, G. Guichard, T, Hintermann, B. Jaun, J. L.

Matthews, J. V. Schreiber, L. Oberer, U. Hommel, H. Widmer, Helv. Chim. Acta

1998, 81, 932.

[28] M. Koppitz, M. Huenges, R. Gratias, H. Kessler, S. L. Goodman, A. Jonczyk,

Helv. Chim. Acta 1997, 80, 1280.

[29] M. Rueping, J. V. Schreiber, G. Lelais, B. Jaun, D. Seebach, Helv. Chim. Acta

2002, 85, 2577.

[30] S. Abele, D. Seebach, Eur. J. Org. Chem. 2000, 1.

[31] S. Reichlin, N. Engl. J. Med. 1983, 309, 1495.

[32] M. F. Chesselet, T. D. Reisine, J. Neurosci. 1983, 3, 232.

[33] D. Hoyer, G. I. Bell, M. Berelowitz, J. Epelbaum, W. Feniuk, P. P. A. Humphrey,

A. M. Ocarroll, Y. C. Patel, A. Schonbrunn, J. E. Taylor, T. Reisine, Trends

Pharmacol. Sei. 1995,16, 86.

[34] W. Bauer, U. Briner, W. Doepmer, R. Haller, R. Huguenin, P. Marbach, T. J.

Petcher, J. Pless, Life Sei. 1982, 31, 1133.

144

[35] L. E. H. Smith, J. J, Kopchick, W. Chen, J. Knapp, F. Kinose, D. Daley, E. Foley,

R. G. Smith, J. M. Schaeffer, Science 1997, 276, 1706.

[36] R. R. Davies, S. J. Turner, K. Alberti, D. G. Johnston, Diabetic Med 1989, 6,

103.

[37] D. F. Vcber, R. M. Freidinger, D. S. Perlow, W. J. Paleveda, F. W. Holly, R. G.

Strachan, R. F. Nutt, B. H. Arison, C. Homnick, W. C. Randall, M. S. Glitzer, R.

Saperstein, R. Hirschmann, Nature 1981, 292, 55.

[38] D. Hoyer, H. Lubbert, C. Bruns, Naunyn-SchmiedebergsArch. Pharmacol. 1994,

350,441.

[39] S. Siehler, K. Seuwen, D. Hoyer, Naunyn-Schmiedebergs Arch. Pharmacol. 1999,

360, 488.

[40] D. J. Hill, M. J. Mio, R. B. Prince, T. S. Hughes, J. S. Moore, Chem. Rev. 2001,

101, 3893.

[41] D. Seebach, J. L. Matthews, A. Meden, T. Wessels, C. Baerlocher, L. B.

McCusker, Helv. Chim. Acta 1997, 80, 173.

[42] H. C. Le, T. Hintermann, T. Wessels, Z. H. Gan, D. Seebach, R. R. Ernst, Helv.

Chim. Acta 2001, 84, 208.

[43] K. Gademann, M. Ernst, D. Hoyer, D. Seebach, Angew. Chem. Int. Ed. 1999, 38,

1223.

[44] G. Melacini, Q. Zhu, M. Goodman, Biochemistry 1997, 36, 1233.

[45] H. Widmer, A. Widmer, W. Braun, J. Biomol. NMR 1993, 3, 307.

[46] E. Pohl, A. Heine, G. M. Sheldrick, Z. Dauter, K. S. Wilson, J. Kallen, W. Huber,

P. J. Pfafïli, ,4cta Crystallogr. Sect. D-Biol. Crystallogr. 1995, 51, 48.

[47] K. Gademann, T. Kimmcrlin, D. Hoyer, D. Seebach, J. Med. Chem. 2001, 44,

2460.

[48] S. Siehler, K. Seuwen, D, Hoyer, Naunyn-Schmiedebergs Arch. Pharmacol. 1998,

357, 483.

145

Experimental Part

7.1 Abbreviations

aq. Aqueous HR High resolution

arom. Aromatic M Molecular peak (MS)

Bn Benzyl MALDI Matrix assisted laser desorption

Boc te/t-Butoxycarbonyl Melm 1 -Methylimidazole

calc. Calculated M.p. Melting point

cone. Concentrated MSNT 1 -(Mesitylene-2sulfonyl)-3-1H-

CD Circular Dichroism 1,2,4-triazole

DMF Dimethylformamide NMM N-Methylmorpholine

DBU 1,8-Diazabicyclo[5.4.0] NMR Nuclear Magnetic Resonance

undec-7-ene org. Organic

DIPEA Diisopropylethylamine Pd/C Palladium on chalcoal

DMAP 4-Dimethylaminopyridine Prep. Preparative

equiv. Equivalent(s) PyBop (Benzotriazl-1 -yl)oxytris

ESI Electrospray ionisation (pyrrolidinophosphonium

Fmoc 9-Fluoroenylmethoxycarbony hexafluorophosphate)

h Hour(s) r.t. Room temperature

h.v. High vacuum (0.01-0.1 Torr) sat. Saturated

HATU 0-(7-Azabenzotriazol-1 -yl) soin. Solution

1,1,3,3-Tetramethyluronium TFA Trifluoroacetic acid

hcxafluorophosphate THF Tetrahydrofuran

HOBt 1-Hydroxy-1 H-benzotriazole TIS Triisopropylsilane

HPLC High performance liquid TLC Thin layer chromatography

chromatography TNBS 2,4,6-Trinitrobenzosulfonic acid

7

146

7.2 General Methods and Materials

Lyophilization: Hetosic cooling condenser with ah.v. pump. Solvents are specified.

Analytical reversed phase (RP) HPLC: Analytical HPLC: Analysis of prepared

compounds was performed on a Knauer HPLC system (WellChrom K-1000 Maxi-Star

pump, degasser, UV detector (variable-wavelength monitor)) or on a. Merk HPLC system

(LaChrom, pump type L-7150, UV detector L-7400, Interface D-7000, HPLC Manager

D-7000), the following columns were used: Macherey-Nagel Q-columns Nucleosil 100-

5 C8 (250 x 4 mm) (Nucleosil 100-5 C1S (250 x 4 mm). TFA for prep. HPLC was used as

UV-grade quality (>99% GC).

Analyses were performed using a linear gradient ofA; 0.1% TFA in H2O and B: MeCN

at a flow rate of 1 ml/min with UV detection at 220 nm. Retention time (tiO in min.

Preparative reversed phase (RP) HPLC: Merck/Hitachi HPLC system (LaChrom,

pump type L-7150, UV detector L-7400, Interface D-7000, HPLC Manager D-7000).

Column: Macherey-Nagel Q-columns Nucleosil 100-5 C8 (250 x 4 mm) (Nucleosil 100-

5 Ci8 (250 x 4 mm). Crude compounds were purified using a gradient ofA (0.1% TFA in

H20) and B (MeCN) at a flow rate of 18 ml/min (unless otherwise stated) with UV

detection at 215 or 220 nm and then lyophilized. UV grade TFA (> 99% GC) was used

forRP-HPLC.

Peptides with free amino groups: The free amino groups on the peptides formed TFA

salts even after lyophilization. The molecular mass (MS) corresponds to the peptide

without TFA.

Mass spectra (MS): Finnigan MAT TSQ 7000 (ESI) spectrometer, lonSpec Ultima 4.7

TFT Ion Cyclotron Resonance (MALDI or HRMS, in a 2,5-dihydroxybenzoic acid

matrix) mass spectrometer; in m/z (% of basis peak).

Circular dichroism (CD) spectra: CD spectra were recorded on a Jasco J-710

spectropolarimeter from 190 to 250 nm at r.t. in 1-mm rectangular cells. The optical

system was flushed with N2 at a flow rate of ca. 10 1/min. Parameters: band width 1.0

nm, resolution 0.2-1 nm, sensitivity 100 mdeg, response 0.5 s, speed 50 nm/min, 5-15

147

accumulations. All spectra were corrected for the corresponding solvent spectrum.

Peptide concentration: 0.2 mM. The molar ellipticity 0 in 10 deg'cm2mol"1 (k in nm) is

calculated for the corresponding peptide (not normalized), taking into account the mass

of TFA for each free amino group. Smoothing was done by Jasco software. Solvents:

MeOH (HPLC grade).

NMR spectra: !H-NMR spectra were recorded on a Bruker Avance 600 (600 MHz),

DRX 500 (500 MHz). 13C-NMR spectra were recorded on a Bruker Avance 600 (150

MHz), DRX 500 (125 MHz). Chemical shifts in ppm with Me4Si as internal standard; J

values in Hz. The multiplicities were classified by the following symbols: s (singlet), d

(doublet), t (triplet), q (quadruplet), m (multiplet or more overlapping signals), br. (broad

signal).

Solvents: THF was freshly distilled over Na/benzophenone under Ar before use. CH2CI2

employed for the coupling reactions was filtered through AI2O3 (Alumina Woelm N,

activity I). Solvents for chromatography and workup procedures were distilled over anhy.

CaS04 (pentane, hexane, AcOEt, MeOH) or P205 (CH2C12).

Reagents Et3N, (/-Pr)2EtN were distilled over CaH2 and stored over KOH and under Ar

(Et3N) or only Ar [(?'-Pr)2EtN]. a-Amino acid derivatives were purchased from Senn,

Novabiochem, or Fluka, Rink amide, Rink amide AM, HMBA and Wang resins were

purchased from Novabiochem.

TNBS test: A little amount of resin was brought in contact on object slide and

subsequently treated with a drop ofTNBS (10 mg /1.5 ml DMF) and one drop EtN(/-Pr)2

(10% in DMF). In case of colour changing of the resin beads from white to red, the

coupling was incomplete. The reaction was prolonged accordingly.

Mixed disulfide formation: Peptide containing free SH group was treated with

methanthiosulfonate (MMTS) stirring in aqueous solution containing 10% DMSO. In

order to prevent the formation unintended disulfide formation, reaction was carried out

avoiding contact with air.

148

General Procedures

Anchoring of ./V-Fmoc-protected amino acids on Wang resin and determination of

loading and initial capping: GeneralProcedure 1 (GP1)

Esterification of the Fmoc-protected amino acid with Wang resin was performed

according to [1], the MSNT/Melm method. The resin was transferred into a dried manual

SPS reactor, swelled in CH2C12 (20 ml/g resin) for 60 min and in DMF (20 ml/g resin)

for other 30 min, and washed with CH2C12. In a separate dry round-bottomed flask

equipped with magnetic stirrer, the Fmoc-protected amino acid (3 equiv.) was dissolved

in dry CH2C12 (3 ml/mmol), then Melm (3.75 equiv.) and MSNT (5 equiv.) were added

under Ar. Stirring was continued until the MSNT was dissolved. Thereafter, the solution

was transferred using a syringe to the reaction vessel containing the resin and mixed by

Ar bubbling for 2-3 h. Subsequently, the resin was filtered, washed with CH2CI2 (20 ml/g

resin, 5 x 1 min), and dried under h.v. for 24 h. The resin substitution was determined by

measuring the absorbance of the dibenzofulvenepiperidine adduct: two samples of the

Fmoc-amino acid resin were weighed exactly (ml(resin) and m2(resin)) and suspended

in pipcridine (20%) in DMF, in a volumetric flasks (VI = V2 = 10T0"31). After 30-40

min the mixtures were transferred in a UV cell and piperidine (20%) to another UV cell

(blank), and the absorbance (A) was measured at 290 nm. The concentrations (ci and c2,

[mM]) of the benzofulvenepiperidine adduct in solution were determined using a

calibration curve. The loading was then calculated according to equation (1):

Substn [mmol/g resin] = cn Vn/{mn(resin) - [cn Vn • (MW- 18)/1000]} (1)

(MW = molecular weight of the Fmoc-protected ß-amino acid)

The yield for the attachment to the resin (loading yield) was determined by equation (2):

Loading yield = [(Substi + Subst2)/2]/Substtheor. (2)

The peptide-rcsin was covered with DMF (20 ml/g resin) and the unreacted OH groups

were then capped using Ac20 (10 equiv.) and DMAP (0.1 equiv.) dissolved in DMF (0.1

ml/mmol Ac20) for 1-2 h under Ar bubbling. The resin was then washed with DMF (20

149

ml/g resin, 5 x 1 min) and with CH2C12 (20 ml/g resin, 5 x 1 min). After capping the resin

was dried under h.v. for 12 h. Thereafter, the resin was used for coupling amino acids.

Capping: GeneralProcedure 2 (GP2)

The peptide-resin was covered with DMF (20 ml/g resin), and unreacted OH groups

were capped using AC2O (10 equiv.) and DMAP (0.1 equiv.) dissolved in DMF (0.1

ml/mmol Ac20) for 1-2 h under Ar bubbling. The resin was then washed with DMF (20

ml/g resin, 5 x 1 min) and with CH2C12 (20 ml/g resin, 5 x 1 min).

The Fmoc-deprotection. Generalprocedure 3 (GP3)

The Fmoc was removed using 20% piperidine in DMF (20 ml/g resin, 2 x 10 min),

DBU/piperidine/DMF 1:1:48 (20 ml/g resin, 3 x 10 min), and 20% piperidine in DMF

(20 ml/g resin, 1x10 min) under Ar bubbling. After filtration, the resin was washed with

DMF (20 ml/g resin, 5 x 1 min) and CH2C12 (20 ml/g resin, 5x1 min).

Coupling of amino acids on Wang resin: GeneralProcedure 4 (GP4)

a) The Fmoc-deprotection was carried out according to (GP 3). Solid phase synthesis was

continued by sequential incorporation of Fmoc-protected amino acids. For each coupling

step, the resin was treated with a soin, of the Fmoc-protected amino acid (3 equiv.),

HATU (2.9 equiv.) and Hünig base (6 equiv.) in DMF (20 ml/g resin) for 60 min.

Monitoring of the coupling reaction was performed with the TNBS test [2], In the case of

a positive TNBS test (indicating incomplete coupling), the suspension was allowed to

react further for 0.5-4 hr or, after filtration, the peptide-resin was treated again with the

same Fmoc-protected amino acid (1-3 equiv.), and with the coupling reagents. The resin

was then filtered and washed with DMF (20 ml/g resin, 5x1 min) prior to the following

deprotection step. For the last coupling either Boc- or Fmoc-protected amino acids were

used. In the second case, the Fmoc group of the last amino acid was removed as

previously described (GP3), and the resin was washed with DMF (20 ml/g resin, 5 x 1

min) and with CH2C12 (20 ml/g resin, 5 x 1 min), and dried under h. v. for 2 h.

150

b) As described in (GP 4a), but coupling with Fmoc-amino acid (5 eq.), HATU (5 eq.)

and Hünig base (10 eq.) in DMF (5 ml) and capping [Ac20 (10 eq.), DMAP (0.2 eq.) in

DMF for 30 min under N2 bubbling] after every coupling step.

Wang resin cleavage and final side chain deprotection: General Procedure 5 (GP5)

The resin cleavage and the final peptide deprotection were performed the dry peptide-

rcsin was suspended in a soin, of TFA/H20/TIS 95:2.5:2.5 (10 ml) for 2 h. The resin was

removed by filtration, washed with TFA (3x2 ml), and the org. phase concentrated

under reduced pressure. The resulting oily residue was slowly treated with cold Et2Û, and

the formed precipitate separated. The crude ß-peptide was dried under h.v. and stored at -

20° before purification.

HPLC Analysis and Purification of the Peptides: GeneralProcedure 6 (GP 6)

6a) Macherey-Nagel C8 column (Nudeosil 100-5 C8 (250 x 4 mm)) by using a linear

gradient of A: 0.1% TFA in H20 and B: MeCN at a flow rate of 1 ml/min. Crude

products were purified by prep. RP-HPLC on a Macherey-Nagel C8 column (Nudeosil

100-7 C8 (250 x 21 mm)) using gradient of A and B at a flow rate of 18 ml/min and then

lyophilized.

6b) Same as in GP 6a, but with a gradient of A: 0.1% TFA in H20 and B: MeOH.

6c). Same as in GP 6a, but with an anal. Merck Lichrospher-100 column (Q, 100 x 4.6

mm), flow rate 1.2 ml/min.

6d) Macherey-Nagel Ci# column (Nudeosil 100-5 Cjs (250 x 4 mm)) by using a linear

gradient of A: 0.1% TFA in H20 and B: MeCN at a flow rate of 1 ml/min. Crude

productswere purified by prep. RP-HPLC on a Macherey-Nagel Cjs column (Nudeosil

100-7 Cjs (250 x 21 mm)) using gradient ofA and B at a flow rate of 12 ml/min and then

lyophilized.

151

7.3 Chemical Shifts of Peptides:

H-(2i?,3,S)-ß2'3hVa!(a-F)-(2^,31S>p2'3hAla(a-F)-(2^,31S>p2'3hLeu(a-F)-(2Ä,35)-ß2'3hVal(a-F)-(2Ä,3Ä)-ß2'3hA]a(a-F)-(2Ä,35)-ß2'3hLeu(a-F)-OH(9).

'H-NMR (600 MHz, CD30H): 0.96 (d, J = 6.68, Me, ß3hLeu); 0.98 (d, J = 6.82, Me,

ß3hVal); 1.04 (d, J = 6.75, Mc, ß3hVal); 1.12 (d, J = 6.75, 2Me, ß3hVal); 1.33 (d, J =

7.04, Mc, ß3hAla); 1.43 (m, 1H, CH2, ß3hLeu); 1.46 (m, IH, CH2, ß3hAla); 1.61 (m, CH,

ß3hLeu); 1.62 (m, CH, ß3hLeu); 1.65 (m, CH, ß3hLeu); 1.66 (m, IH, CH2, ß3hLeu); 1.99

(m, CH, ß3hVal); 2.15 (m, CH, ß3hVal); 3.49 (m, CH, ß3hVal); 4.18 (m, CH, ß3hVal);

4.49 (tw, CH, ß3hLeu); 4.56 (m, CH, ß3hVal); 4.59 (m, CH, ß3hAla); 4.92 (br, CH,

ß3hAla); 4.95 (d, J - 3.10, CH, ß3hLcu); 5.02 (d, J = 2.64, CH, ß3hVal); 5.15 (d, J =

2.35, CH, ß3hVal); 5.34 (d, J = 2.05, CH, ß3hVal); 7.98 (d, J = 9.1, NH, ß3hAla); 8.08

(d, J = 9.90, NH, ß3hVal); 8.24 (d, J = 9.24, NH, ß3hLeu); 8.31 (d, J = 9.17, NH,

ß3hAla). I3C-NMR (150 MHz, CD30H): 18.95 (Me, ß3hVal); 18.98 (Me, ß3hAla); 19.07

(Me, ß3hAla); 19.94 (Me, ß3hVal); 22.32 (Me, ß3hLeu); 23.56 (Me, ß3hLeu); 25.71 (CH,

ß3hLeu); 25.76 (CH, ß3hLeu); 30.14 (CH, ß3hVal); 30.81 (CH, ß3hVal); 40.61(CH2,

ß3hLeu); 40.74 (CH2, ß3hLcu); 47.16 (CH, ß3hAla); 47.88 (CH, ß3hAla); 49.85 (CH,

ß3hLeu); 50.01 (CH, ß3hLeu); 56.99 (CH, ß3hVal); 59.05 (CH, ß3hVal); 89.48 (CH2,

ß3hVal); 91.20 (CH, ß3hLeu); 91.39 (CH2, ß3hVal); 92.3 (CH2, ß3hAla); 92.54 (CH2,

ß3hLeu); 92.65 (CH2, ß3hAla); 168.37 (CO); 169.75 (C=0); 169.97 (C=0); 170.15

(C=0); 170.40 (C=0).

H-(31S)-ß2,2,3hVaI(a,a-F2)-(31S)-ß2'2'3hAla(a,a-F2)-(35)-ß2,2,3hLeu(a,a-F2)-(31S)-ß2'2'3hVal(a,a-F2)-(3Ä>ß2'2'3hAla(a,a-F2)-(35)-ß2'2'3hLeu(a,a-F2)-OH(10).

!H-NMR (600 MHz, DMSO): 0.78 (d, J = 6.46, 2Me, ß3hLeu); 0.81 (d, J = 6.46, 2Me,

ß3hLcu); 0.92 (d, J = 6.60, 2Me, ß3hVal); 0.93 {d, J = 6.90, Me, ß3hVal); 0.98 (d, J =

6.97, Me, ß3hVal); 1.12 (d,J= 7.10, Me, ß3hAla); 1.17 (d, J = 7.04, Me, ß3hAla); 1.20

(m, CH2, ß3hLeu); 1.50 (m, CH, ß3hLeu); 1.52 (m, CH, ß3hLeu); 1.70 (m, CH2, ß3hLeu);

1.96 (m, CH, ß3hVal); 2.07 (m, CH, ß3hVal); 3.63 (m, CH, ß3hVal); 4.30 (m, CH,

ß3hLeu); 4.50 (m, CH, ß3hVal); 4.61 (m, CH, ß3hAla); 4.66 (m, CH, ß3hLeu); 4.69 (m,

CH, ß3hAla); 9.09 (d, J = 8.36, NH, ß3hAla); 9.03 {d, J = 9.32, NH, ß3hLeu); 8.95 (d, J

= 9.76, NH, ß3hVal); 9.10 (d, J = 8.10, NH, ß3hLeu); 9.28 (br, NH, ß3hAla). 13C-NMR

152

(150 MHz, CD3OH): 12.75 (Me, ß3hAla); 12.82 (Me, ß3hAla); 18.43 (Me, ß3hVal);

19.88 (Me, ß3hVal); 20.41 (Mc, ß3hLeu); 20.72 (Me, ß3hAla); 23.58 (CH, ß3hLeu);

23.83 (CH, ß3hAla); 26.47 (CH, ß3hVal); 26.83 (CH, ß3hVal); 35.04 (CH2, ß3hLeu);

36.35 (CH2, ß3hAla); 46.51 (CH, ß3hAla); 46.26 (CH, ß3hAla); 48.98 (CH, ß3hLeu);

49.86 (CH, ß3hAla); 55.12 (CH, ß3hVal); 56.58 (CH, ß3hVal).

H-(2/?,3*)-ß2,3hVal(a-OH)-(2i?,31S)-ß2'3hAla(a-OH)-(2Ä,3Ä)-ß2'3hLeu(a-OH)-(2/?,3Ä)-ß2'3hVal(a-OH)-(2i?,31S)-ß2'3hAla(a-OH)-(2i[?,31S)-ß2'3hLeu(a-OH)-OH(ll).

0.93 (d, J - 6.82, Me, ß3hVal); 0.94 (d, J = 6.46, 2Me, ß3hLeu); 0.97 (d, J = 6.46, 2Me,

ß3hLeu); 1.01 (d, J - 6.68, Me, ß3hVal); 1.10 (d, J = 6.82, 2Me, ß3hVal); 1.15 (d, J =

6.75, Me, ß3hAla); 1.23 (d, J = 7.04, Mc, ß3hAla); 1.47 (m, CH2, ß3hLeu); 1.62 (m, CH,

ß3hLeu); 1.43 (m, 1H, CH2, ß3hLeu); 1.50 (m, 1H, CH2, ß3hLeu); 1.61 (m, 1H, CH2,

ß3hLeu); 1.94 (m, CH, ß3hVal); 2.1 (m, CH, ß3hVal); 3.20 (m, CH, ß3hVal); 4.02 (d, J =

2.92, CH, ß3hAla); 4.05 (m, CH, ß3hVal); 4.09 (br, CH, ß3hLeu); 4.U (d,J = 3.08. CH,

ß3hLeu); 4.14 (d, J = 2.64, ß3hAla); 4.26 (d, J = 2.80, CH, ß3hVal); 4.34 (m, CH,

ß3hAla); 4.39 (m, CH, ß3hAla); 4.40 (m, CH, ß3hLcu); 4.42 (m, CH, ß3hLeu); 4.43 (d, J

= 1.83, CH, ß3hVal); 7.71 (d, J - 9.61, NH, ß3hLeu); 7.81 (d, J = 9.61, ß3hLeu); 7.84

(br, NH, ß3hAla); 7.97 (d, J = 9.24, ß3hAla); 7.99 (d, J - 9.68, NH, ß3hVal). I3C-NMR

(150 MHz, CD3OH): 16.87 (Mc ß3hAla); 17.65 (Me, ß3hAla); 19.36 (Me, ß3hVal); 23.52

(Me, ß3hLeu); 20.29 (Me, ß3hVal); 25.88 (CH, ß3hLeu); 25.89 (CH, ß3hLeu); 30.01 (CH,

ß3hVal); 31.16 (CH, ß3hVal); 42.12 (CH2, ß3hLeu); 42.30 (CH2, ß3hLeu); 48.71 (CH,

ß3hAla); 49.08 (CH, ß3hAla); 50.72 (CH, ß3hLeu); 51.06 (CH, ß3hLcu); 58.12 (CH,

ß3hVal); 60.75 (CH, ß3hVal); 69.55 (CH, ß3hVal); 72.46 (CH, ß3hVal); 73.78 (CH,

ß3hLeu); 73.95 (CH, ß3hLeu); 74.23 (CH, ß3hAla); 74.85 (CH, ß3hAla).

153

7.4 Synthesis of ot/ß-, oc/y- and ß/y-Mixed Peptides

H-Asp-t^-ß^Phe-Ser-^-ß^Lys-Phe-^-ß^Glu-Glu-^-ß^Ala-Lys-OHtlS).

Fmoc-Lys-OH (223 mg, 0.72 mmol) was loaded onto the Wang resin (200 mg, 0.90

mmol/g, 100-200 mesh) according to (GP 1). The loading was determined to be 0.64

mmol/g (72%), corresponding to 0.13 mmol ofFmoc-Lys-OH. After capping (GP 2), the

peptide synthesis was performed according to (GP 4a), Treatment of the peptide-resin

according to (GP 5a) afforded the crude peptide IS (100 mg). Purification of a part of the

crude peptide (20 mg) by RP-HPLC (5-15% B in 40 min) according to (GP 6a) yielded

10 mg of 15 (calc. overall yield 50%) as TFA salt. White solid. RP-HPLC (5-10% B in 5

min, 10-50% B in 40 min, 50-95% B in 50 min; tR 33.20 min): purity >95%. ^-NMR

(600 MHz, CD3OH): \.\2(d,J = 6.8, Me, ß3hAla); 1.50 (m, CH2, Lys); 1.55 (m, CH2,

ß3hLys); 1.62 (m, CH2, ß3hLys); 1.65 (m, CH2, Lys); 1.67 (m, CH2 ,ß3hLys); 1.71 (m,

CH2, Lys); 1.79 (m, CH2, ß3hGlu); 1.91 (m, CH2, Glu); 2.20 (m, CH2, ß3hGlu); 2.26 (m,

CH2, ß3hGlu); 2.27 (m, CH2, ß3hLys); 2.30 (m, CH2, Glu); 2.41 (m, CH2, ß3hAla); 2.52

(dd, J = 7.5, 14.8, 1H, CH2, ß3hPhe); 2.80 (m, CH2, Phe); 2.88 (m, 1H, CH2, ß^Phe);

2.90 (m, CH2, Lys); 2.92 (m, CH2, ß3hLys); 3.77 (d, J = 5.6, CH2, Ser); 4.05 (m, CH,

Asp); 4.06 (m, CH, ß3hGlu); 4.15 (m, CH, ß3hLys); 4.21 (m, CH, Glu); 4.23 (m, CH,

Ser); 4.24 (m, CH, ß3hAla); 4.43 (m, CH, ß3hPhe); 4.50 (m, CH, Phe); 7.15-7.26 (m, 3

arom. H, Phe); 7.16-7.25 (m, 3 arom. H, ß3hPhe); 7.23-7.30 (m, 2 arom. H, Phe); 7.26-

7.29 (m, 2 arom. H, ß3hPhe); 7.94 (d, J = 9.0, NH, ß3hLys); 8.05 (d, J = 8.1, NH, Ser);

8.10 (d, J = 8.5, NH, ß3hGlu); 8.12 (d, J = 7.1, NH, Phe); 8.13 (d, J = 7.3, NH, Glu);

8.24 (d, J = 8.0, NH, Lys); 8.32 (d, J = 8.6, NH, ß3hPhe); 8.58 (br, NH, ß3hAla). 13C-

NMR (150 MHz, CD3OH): 28.0 (CH2, ß3hLys); 28.1 (CH2, Lys); 28.4 (CH2, ß3hGlu);

28.5 (CH2, Glu); 28.8 (Me, ß3hAla); 31.3 (CH2, ß3hGlu); 31.6 (CH2, Glu); 32.1 (CH2,

Lys); 34.4 (CH2, ß3hLys); 36.4 (CH2, Asp); 38.7 (CH2, Phe); 40.8 (CH2, Lys); 40.9 (CH2,

ß3hLys); 41.0 (CH2, ß3hPhe); 41.4 (CH2, ß3hPhe); 41.6 (CH2, ß3hLys); 41.7 (CH2,

ß3hGlu); 43.1 (CH2, ß3hAla); 62.9 (CH2, Ser); 127.8 (arom. C); 129.6 (arom. C); 130.3

(arom. C); 138.4 (arom. C); 127.9 (arom. C); 129.6 (arom. C); 130.4 (arom. C); 139.2

(arom. C); 169.0 (CO); 172.3 (C=0); 173.0 (C=0); 173.4 (C=0); 173.6 (C=0); 175.4

(C=0); 176.7 (C=0); 177.2 (C=0). MALDI-HR-MS: 1156.5739 ([M+H]+; calc. For

[C54HSiNiiOi7]+: 1156.5812).

154

H-Val-^-ß^Ala-Leu-^-ß^Val-Ala-^-ß^Leu-Val-^-ß'hAla-Leu-OHtlö).

Fmoc-Leu-OH (318 mg, 0.90 mmol) was loaded onto the Wang resin (200 mg, 0.90

mmol/g, 100-200 mesh) according to (GP 1). The loading was determined to be 0.58

mmol/g (68%), corresponding to 0.12 mmol of Fmoc-Leu-OH. After capping (GP 2), the

peptide synthesis was performed according to (GP 4a), Treatment of the peptide-rcsin

according to (GP 5a) afforded the crude peptide 16 (100 mg). Purification of a part ofthe

crude peptide (20 mg) by RP-HPLC (50% B in 10 min, 50-95% B in 45 min; tR 30.00

min) according to (GP 6b) yielded 10 mg of 16 (calc. overall yield 50%) as TFA salt.

White solid. RP-HPLC (5-10% B in 5 min, 10-90% B in 5 min, 90-99% B in 60 min; tR

34.52 min): purity >95%. MALDI-HR-MS: 924.6521 ([M+H]+; calc. for

[C46H85N9O10]+: 924.6419).

H-Ser-^-ß^Phe-Leu-t^-ß^Asp-Phe-^-ß^Val-Lys-^-ß'hSer-Ala-OHtn)

Fmoc-Ala- OH (311 mg, 1.00 mmol) was loaded onto the Wang resin (200 mg, 0.90

mmol/g, 100-200 mesh) according to (GP 1). The loading was determined to be 0.63

mmol/g (70%), corresponding to 0.12 mmol of Fmoc-Ala-OH. After capping (GP 2), the

peptide synthesis was performed according to (GP 4a). Treatment of the peptide-resin

according to (GP 5a) afforded the crude peptide 17 (100 mg). Purification of a part of the

crude peptide (20 mg) by RP-HPLC (5-15% B in 40 min) according to GP 6a yielded 10

mg of 17 (calc. overall yield 50%) as TFA salt. White solid. RP-HPLC (5-99% B in 30

min; 5% B in 20 min, 30% B in 30 min, 99% B in 30 min; tR 28.23 min): purity >95%.

^-NMR (600 MHz, CD3OH): 0.85 (m, 2 Me, ß3hVal); 0.88 (d, J - 6.5, Me, Leu); 0.92

(d, J = 6.6, Me, Leu); 1.37 (d, J = 7.4, Me, Ala); 1.41 (m, CH2, Lys); 1.50 (m, CH2, Leu);

1.62 (m, CH, Leu); 1.65 (m, CH2, Lys); 1.75 (m, CH2, Lys); 1.76 (m, CH, ß3hVal); 2.25

(m, 1H, CH2, ß3hVal); 2.29 (m, CH2, ß3hAsp); 2.43 (dd, J = 8.4, 14.5, 1H, CH2, ß3hPhe);

2.51 (m, CH2 , ß3hSer); 2.52 (m, 1H, CH2, ß3hVal); 2.54 (m, 1H, CH2, ß3hPhe); 2.64 (dd,

J = 1.1, 14.3, 1H, CH2, ß3hAsp); 2.69 (d: J = 8.8, 16.0, 1H, CH2, ß3hAsp); 2.84 (dd, J -

7.0, 14.0, 1H, CH2, ß3hPhe); 2.93 (dd, J = 6.5, 15.2, 1H, CH2, ß3hPhc); 2.94 (dd, J = 6.0,

14.3, 1H, CH2, Phe); 2.96 (m, CH2, Lys); 3.15 (dd, J = 8.0, 16.0, 1H, CH2, Phe); 3.52

(m, CH2, ß3hSer); 3.82 (m, CH, Ser); 3.86 (m, CH2, Ser); 4.06 (m, CH, ß3hVal); 4.23 (m,

CH, ß3hSer); 4.24 (m, CH, ß3hAsp); 4.28 (m, CH, Lys); 4.32 (m, CH, Ala); 4.46 (m, CH,

Phe); 4.50 (m, CH, ß3hPhe); 4.62 (m, CH, Leu); 7.16-7.23 (m, 3 arom. H, Phe); 7.18-

7.23 (m, 3 arom. H, ß3hPhe); 7.24-7.29 (m, 2 arom. H, Phe); 7.26-7'.28 (m, 2 arom. H,

155

ß3hPhe); 7.95 (d, J = 8.2, NH, p3hSer); 8.02 (d, J = 7.8, NH, Phe); 8.03 (d, J = 8.1, NH,

p3hVal); 8.10 (d, J = 7.6, NH, Lys); 8.13 (d, J = 6.7, NH, Leu); 8.23 (d, J = 7.7, NH,

Ala); 8.25 (d, J = 8.8, NH, ß3hPhe); 8.49 (d, J = 7.7, NH, ß3hAsp). 13C-NMR (150 MHz,

CD3OH): 17.9 (Me, Ala); 22.3 (Me, ß3hVal); 23.2 (Me, Leu); 23.6 (CH2, Lys); 26.1 (CH,

Leu); 28.1 (CH2, Lys); 32.5 (CH2, Lys); 37.9 (CH2, ß3hSer); 39.2 (CH2, ß3hVal); 39.3

(CH2, Phc); 40.6 (CH2, ß3hAsp); 40.7 (CH2, ß3hAsp); 41.1 (CH2, Lys); 41.2 (CH2,

ß3hPhe); 41.4 (CH2, ß3hPhe); 41.7 (CH2, Leu); 47.1 (CH, ß3hAsp); 49.8 (CH, Ala); 50.5

(CH, ß3hSer); 51.3 (CH, ß3hPhe); 54.2 (CH, ß3hVal); 54.3 (CH, Leu); 54.6 (CH, Lys);

56.5 (CH, Ser); 58.1 (CH, Phe); 62.0 (CH2, Ser); 64.5 (CH2, ß3hSer); 127.8 (arom. C,

ß3hPhe); 127.9 (arom. C, Phe); 129.6 (arom. C, ß3hPhe); 129.7 (arom. C, Phe); 130.3

(arom. C, ß3hPhe); 130.4 (arom. C, Phe); 138.7 (arom. C, ß3hPhe); 139.3 (arom. C, Phe);

168.1 (C=0); 173.0 (C=0); 173.3 (C=0); 173.8 (C=0); 174.1 (C=0); 174.9 (C=0);

175.0 (C=0); 176.5 (C=0). MALDI-HR-MS: 1069.5928 ([M+HJ+; calc. for

[C52H81N,oOi4]+: 1069.5856).

H-(i?)-Ser-(1S)-ß3hPhe-(/?)-Leu-(i?)-ß3hAsp-(i?)-Phe-(7?)-ß3hVal-(^)-Lys-(Ä)-ß3hSer-(Ä)-AIa-OH (18).

Fmoc-(R)-Ala-OH (311 mg, 1.00 mmol) was loaded onto the Wang resin (200 mg, 1.00

mmol/g, 100-200 mesh) according to (GP 1). The loading was determined to be 0.62

mmol/g (62%), corresponding to 0.12 mmol of Fmoc-(7?)-Ala-OH. After capping (GP 2),

the peptide synthesis was performed according to GP 4a. Treatment of the peptide-resin

according to (GP 5a) afforded the crude peptide 18 (100 mg). Purification of a part of the

crude peptide (20 mg) by RP-HPLC (5-15% B in 40 min) according to GP 6a yielded 10

mg of 18 (calc. overall yield 50%) as TFA salt. White solid. RP-HPLC (5-99% B in 30

min; 5% B in 5 min, 30% B in 20 min, 99% B in 30 min; tR 23.51 min): purity >95%. !H-

NMR (600 MHz, CD3OH): 0.65 (d, J = 6.6, Me, ß3hVal); 0.83 (d, J = 6.9, Me, ß3hVal);

0.93 (d, J = 6.5, Me, Leu); 0.96 (d, J = 6.6, Me, Leu); 1.34 (d, J = 7.3, Me, Ala); 1.51

(m, CH2, Lys); 1.55 (m, CH2, Leu); 1.62 (m, CH, ß3hVaI); 1.65 (m, CH, Leu); 1.67 (m,

CH2, Lys); 1.82 (m, CH2, Lys); 2.34 (dd, J = 7.4, 14.5, 1H, CH2, ß3hVal); 2.46 (m, CH2,

ß3hAsp); 2.50 (m, CH2, ß3hSer); 2.55 (m, CH2, ß3hAsp); 2.58 (m, CH2, ß3hPhe); 2.72

(ß3hPhe); 3.40 (dd, J = 7.8, 12.0, 1H, CH2, Ser); 2.48 (m, 1H, CH2, ß3hVal); 2.92 (m,

CH2, Lys); 3.04 (m, CH2, Phe); 3.61 (m, CH2, ß3hSer); 3.63 (w, 1H, CH2, Ser); 3.80 (m,

CH, Ser); 3.88 (m, CH, ß3hVal); 4.21 (m, CH, Lys); 4.25 (m, CH, ß3hSer); 4.30 (m, CH,

156

Leu); 4.36 (m, CH, Ala); 4.50 (m, CH, ß3hAsp); 4.53 (m, CH, ß3hPhe); 4.55 (m, CH,

Phe); 7.15-7.24 (m, 3 arom. H, ß3hPhe); 7.26-7.28 (m, 2 arom. H, ß3hPhe); 7.16-7.23

(m, 3 arom. H, Phe); 7.24-7.28 (m, 2 arom. H, Phe); 7.97 (d, J = 8.1, NH, ß3hSer); 8.02

(d, J = 8.9, NH, ß3hVal); 8.03 (d, J = 7.0, NH, Lys); 8.15 (d, J = 6.9, NH, Leu); 8.24 (d,

J = 7.2, NH, Ala); 8.27 (d, J = 8.8, NH, ß3hAsp); 8.29 (d, J = 9.1, NH, ß3hPhe); 8.31 (d,

J = 7.0, NH, Phe). 13C-NMR (150 MHz, CD3OH): 17.7 (Me, Ala); 19.3 (Me, ß3hVal);

23.3 (Me, Leu); 23.8 (CH2, Lys); 26.1 (CH, Leu); 28.1 (CH2, Lys); 32.2 (CH2, Lys); 39.1

(CH2, Phe); 40.9 (CH2, Lys); 41.2 (CH2, ß3hPhe); 41.5 (CH2, ß3hPhe); 41.7 (CH2, Leu);

45.6 (CH, ß3hAsp); 45.7 (CH2, ß3hAsp); 49.9 (CH, Ala); 50.1 (CH, ß3hPhe); 50.7 (CH,

ß3hSer); 54.1 (CH, Leu); 54.4 (CH, ß3hVal); 55.2 (CH, Lys); 56.8 (CH, Ser); 57.2 (CH,

Phe); 61.9 (CH2, Ser); 64.2 (CH2, ß3hSer); 127.7 (arom. C, ß3hPhe); 127.9 (arom. C,

Phe); 129.5 (arom. C, ß3hPhe); 129.6 (arom. C, Phe); 130.4 (arom. C, ß3hPhe); 130.5

(arom. C, Phe); 138.2 (arom. C, ß3hPhe); 139.3 (arom. C, Phe); 167.9 (C=0); 172.7

(C=0); 173.0 (C=0); 173.1 (C=0); 173.5 (CO); 173.7 (C=0); 174.4 (C=0); 174.5

(C=0); 174.9 (C=0); 176.9 (C=0). MALDI-HR-MS: 1069.5928 ([M+H]+; calc. for

[C52H8iNioOi4]+: 1069.5856).

H-Aib-(7?)-ß3hSer-Aib-(5)-ß3hPhe-Aib-(1S)-ß3hLys-Aib-(y?)-ß3hVal-OH(21).

Fmoc-(Ä)-ß3hVal-OH (317 mg, 0.90 mmol) was loaded onto the Wang resin (200 mg,

0.90 mmol/g, 100-200 mesh) according to (GP 1). The loading was determined to be

0.63 mmol/g (70%), corresponding to 0.13 mmol of Fmoc-(7?)-ß3hVal-OH. After capping

(GP 2), the peptide synthesis was performed according to (GP 4a). Treatment of the

peptide-resin according to (GP Sa) afforded the crude peptide 21 (100 mg). Purification

of a part of the crude peptide (20 mg) by RP-HPLC (5-15% B in 40 min) according to

(GP 6a) yielded 10 mg of 21 (calc. overall yield 50%) as TFA salt. White solid. RP-

HPLC (10% B in 5 min, 10-60% B in 45 min, 60-95% B in 55 min; tR 33.12 min): purity

>95%. !H-NMR (600 MHz, CD3H): 0.90 (d, J = 6.8, Me, ß3hVal); 0.92 (d, J = 7.0, Me,

ß3hVal); 1.15 (s, 2 Me, Aib); 1.17 (s, Me, Aib); 1.33 (s, Me, Aib); 1.42 (m, CH2,

ß3hLys); 1.43 (s, 2 Me, Aib); 1.46 (s, 2 Me, Aib); 1.57 (m, 1H, CH2, ß3hLys); 1.61 (m,

CH2, ß3hLys); 1.66 (m, 1H, CH2, ß3hLys); 1.85 (m, CH, ß3hVal); 2.27 (dd, J = 4.7, 14.6,

1H, CH2, ß3hLys); 2.37 (dd, J = 4.0, 14.0, 1H, CH2, ß3hPhe); 2.47 (m, 1H, CH2, ß3hLys);

2.48 (dd, J = 8.0, 15.8, 1H, CH2, ß3hSer); 2.52 (dd, J - 6.4, 14.4, 1H, CH2, ß3hSer); 2.53

157

(m, CH2, ß3hPhe); 3.61 (m, CH2, ß3hSer); 4.05 (m, CH, ß3hVal); 4.18 (m, CH, ß3hLys);

4.30 (m, CH, ß3hSer); 4.51 (m, CH, ß3hPhe); 7.16-7.23 (m, 3 arom. H, ß3hPhe); 7.24-

7.28 (m, 2 arom. H, ß3hPhe); 7.58 (d, J = 9.2, NH, ß3hPhe); 7.68 (d, J = 8.9, NH,

ß3hVal); 7.82 (d,J = 9.1,NH, ß3hLys); 7.99 (rf, J- 8.1,NH, ß3hSer); 8.03 (s,NH, Aib);

8.06 (s, NH, Aib); 8.16 (s, NH, Aib). 13C-NMR (150 MHz, CD3OH): 19.6 (Me, ß3hVal);

23.8 (CH2, ß3hLys); 24.7 (Me, Aib); 24.8 (Me, Aib); 24.9 (Me, Aib); 25.4 (Me, Aib);

26.3 (Me, Aib); 26.6 (Me, Aib); 27.0 (Me, Aib); 27.8 (CH2, ß3hLys); 34.9 (CH2, ß3hLys);

37.8 (CH2, ß3hVal); 40.1 (CH2, ß3hPhe); 43.1 (CH2, ß3hPhe); 43.6 (CH2, ß3hLys); 47.9

(ß3hLys); 49.9 (CH, ß3hPhe); 50.1 (CH2, ß3hScr); 50.9 (CH, ß3hSer); 53.5 (CH, ß3hVal);

60.4 (CH2, ß3hSer); 127.6 (arom. C); 129.4 (arom. C); 130.4 (arom. C); 139.5 (arom. C);

172.9 (C=0); 173.1 (C=0); 173.2 (C=0); 175.7 (C=0); 176.4 (C=0); 176.6 (C=0);

177.8 (C=0). MALDI-HR-MS: 875.5539 ([M+H]+; cale, for [C43H73N9Oiol+: 876.5553).

H-Aib-(5)-ß3hTyr-Aib-(*)-ß3hLys-Aib-(1S)-ß3hAsp-Aib-(i?)-ß3hVal-OH(22).

Fmoc-(#)-ß3hVal-OH (318 mg, 0.90 mmol) was loaded onto the Wang resin (200 mg,

0.90 mmol/g, 100-200 mesh) according to (GP 1). The loading was determined to be

0.61 mmol/g (68%), corresponding to 0.13 mmol of Fmoc-(#)-ß3hVal-OH. After capping

(GP 2), the peptide synthesis was performed according to (GP 4a). Treatment of the

peptide-resin according to (GP 5a) afforded the crude peptide 22 (100 mg). Purification

of a part of the crude peptide (20 mg) by RP-HPLC (5-15% B in 40 min) according to

(GP 6a) yielded 10 mg of 22 (calc. overall yield 50%) as TFA salt. White solid. RP-

HPLC (10% B in 5 min, 10-60% B in 45 min, 60-95% B in 50 min; tR 33.12 min): purity

>95%. }H-NMR (600 MHz, CD3OH): 0.91 (m, Me, ß3hVal); 1.30 (s, Me, Aib); 1.41 (s,

Me, Aib); 1.44 (s, Me, Aib); 1.48 (s, Me, Aib); 1.50 (s, Me, Aib); 1.58 (m, CH2, ß3hLys);

1.61 (m, CH2, ß3hLys); 1.68 (m, CH2, ß3hLys); 1.80 (m, CH, ß3hVal); 2.35 (m, CH2,

ß3hLys); 2.40 (dd, J = 4.8, 14.4, 1H, CH2, ß3hAsp); 2.42 (m, CH2, ß3hTyr); 2.53 (m, 1H,

CH2, ß3hAsp); 2.60 (m, 1H, CH2, ß3hTyr); 2.62 (m, CH2, ß3hAsp); 2.80 (dd, J = 5.0,

14.0, 1H, CH2, ß3hTyr); 2.91 (m, CH2, ß3hLys); 4.02 (m, CH, ß3hVal); 4.10 (m, CH,

ß3hLys); 4.40 (m, CH, ß3hTyr); 4.51 (m, CH, ß3hAsp); 6.68 (m, 2 arom. H, ß3hTyr); 7.12

(m, 2 arom. H, ß3hTyr); 7.59 (d, J = 9.0, NH, ß3hVal); 7.6 (br, NH, ß3hLys); 7.81 (d, J =

8.8, NH, ß3hTyr); 7.85 (d, J = 8.6, NH, ß3hAsp); 8.02 (s, NH, Aib); 8.12 (s, NH, Aib);

8.14 (s, NH, Aib). 13C-NMR (150 MHz, CD3OH): 19.5 (Me, ß3hVal); 24.0 (Me, Aib);

158

24.6 (Me, Aib); 25.1 (Me, Aib); 25.2 (Me, Aib); 25.4 (Me, Aib); 26.1 (Me, Aib); 26.3

(Me, Aib); 26.5 (Me, Aib); 27.8 (CH2, ß3hLys); 34.8 (CH2, ß3hLys); 37.6 (CH2, ß3hVal);

40.5 (CH2, ß3hTyr); 40.9 (CH2, ß3hLys); 42.3 (CH2, ß3hTyr); 42.8 (CH2, ß3hLys); 45.6

(CH, ß3hAsp); 48.1 (CH, ß3hLys); 50.6 (CH, ß3hTyr); 53.5 (CH, ß3hVal); 116.6 (arom.

C); 130.2 (arom. C); 131.4 (arom. C); 157.4 (arom. C); 172.6 (C=0); 172.7 (C=0); 172.7

(C=0); 173.2 (C=0); 174.8 (C=0); 175.6 (CO); 176.4 (C=0); 176.5 (C=0); 176.6

(C=0). MALDI-HR-MS: 920.5469 ([M+H]+; calc. for [C44H73N9Oi2]+: 920.5452).

H-(,S)-ß3hLeu-Ile-(1S)-ß3hLys-Aib-(if)-ß3hVal-Glu-(Ä)-ß3hPhe-OH(23).

Fmoc-(5)-ß3hPhe-OH (823 mg, 2.05 mmol) was loaded onto the Wang resin (500 mg,

0.82 mmol/g 100-200 mesh) according to (GP 1). The loading was determined to be 0.58

mmol/g (71%), corresponding to 0.29 mmol of Fmoc-(.S)-ß3hPhe-OH. After capping (GP

2), a portion of this Fmoc-amino acid-resin (200 mg) was used for synthesis (GP 4b).

After cleavage from the resin (GP 5a, 4 h), the thus obtained crude peptide was purified

by prep. RP-HPLC (5-50% B in 40 min) according to (GP 6c) to yield 85 mg (56%) of

23 as a TFA salt. White solid. RP-HPLC (5-50% B in 40 min; tR 26.99 min): purity

>97%. *H-NMR (600 MHz, CD3OH): 0.92 (d, J = 6.5, Me, ß3hVal); 0.94 (m, Me,

ß3hVal); 0.95 (d, J = 7.5, Me, Ile); 0.96-0.98 (m, 2 Me, ß3hLeu); 1.22 (m, CH2, Ile); 1.42

(m, CH2, ß3hLys); 1.43 (s, Me, Aib); 1.45 (s, Me, Aib); 1.61 (m, CH2, ß3hLys); 1.63 (m,

CH2, ß3hLys); 1.70 (m, CH, ß3hLeu); 1.72 (m, CH2, ß3hLeu); 1.8 (m, CH, Ile); 1.82 (m,

CH, ß3hVal); 1.87 (m, 1H, CH2, Glu); 1.94 (m, 1H, CH2, Glu); 2.29 (m, CH2, Glu); 2.34

(dd, J = 6.0, 15.5, 1H, CH2, ß3hLys); 2.43 (m, CH2, ß3hVal); 2.44 (dd, J = 4.5, 16.5, 1H,

CH2, ß3hLys); 2.49 (m, CH2, ß3hPhe); 2.58 (dd, J = 5.0, 16.0, 1H, CH2, ß3hLeu); 2.69

(dd, J - 7.6, 15.6, 1H, CH2, ß3hLeu); 2.89 (m, CH2, ß3hPhe); 2.92 (m, CH2, ß3hLys);

3.58 (m, CH, ß3hLeu); 4.15 (t, J = 7.4, CH, Ile); 4.22 (m, CH, Glu); 4.40 (m, CH,

ß3hPhe); 7.18-7.23 (m, 3 arom. H); 7.26-7.28 (m, 2 arom. H, ß3hPhe); 7.50 (d, J = 9.5,

NH, ß3hVal); 7.99 (d, J = 8.5, NH, ß3hPhe); 8.02 (d, J = 8.3, NH, ß3hLys); 8.03 (d, J =

7.0, NH, Glu); 8.07 (s, NH, Aib); 8.24 (m, J = 7.5, NH, Ile). J3C-NMR (150 MHz,

CD3OH): 16.2 (Me, Ile); 19.3 (Me, ß3hLeu); 25.4 (CH, ß3hLeu); 25.5 (Me, Aib); 26.10

(CH2, Ile); 26.15 (CH2, ß3hLeu); 26.20 (Me, Aib); 26.23 (Me, ß3hVal); 28.6 (CH2, Glu);

31.4 (CH2, Glu); 33.4 (CH, ß3hVal); 37.9 (CH, Ile); 38.3 (CH2, ß3hLeu); 39.2 (CH2,

ß3hLeu); 39.4 (CH2, ß3hVal); 41.2 (CH2, ß3hLeu); 49.0 (CH, ß3hLeu); 49.4 (CH,

159

ß3hPhe); 54.1 (CH, ß3hVal); 55.1 (CH, Glu); 60.1 (CH, Ile); 127.6 (arom. C); 129.4

(arom. C); 130.5 (arom. C); 139.3 (arom. C); 172.5 (C=0); 172.7 (OO); 173.39 (0=0);

173.43 (C=0); 174.1 (C=0); 175.0 (C=0); 176.5 (OO); 176.8 (C=0).

MALDI-HR-MS: 889.5748/911.5617 (100/58, [M+H]+/[M+Na]+; calc. for

[C45H77N8Oio]7[C45H76Ng010Na]+: 889.5757/911.5577).

H-^-ß'hLeu-Val-^ySJ-ß^^lataMeJ-Lys-^ySJ-ß^^lataMei-Val-^-ß^Phe-OH(24).

Fmoc-(S)-ß3hPhe-OH (933 mg, 2.33 mmol) was loaded onto the Wang resin (500 mg,

0.93 mmol/g 100-200 mesh) according to (GP 1). The loading was determined to be 0.72

mmol/g (77%), corresponding to 0.36 mmol of Fmoc-(S)-ß3hPhe-OH. After capping (GP

2), a portion of this Fmoc-amino acid-resin (200 mg) was used for synthesis (GP 4b).

After cleavage from the resin (GP 5a, 4 h), the crude peptide was purified by prep. RP-

HPLC (5-50%) B in 40 min) according to GP 6c to yield 59 mg (39%) of 24 as a TFA

salt. White solid. RP-HPLC (5-50% B in 40 min; tR 28.59 min): purity >95%. 'H-NMR

(600 MHz, CD3OH): 0.87 (d, J = 6.7, Me, Val); 0.89 (d, J = 6.6, Me, Val); 0.94 (m, 2

Me, ß2hLeu); 0.96 (m, 2 Me, Val); 1.02 (d, J = 6.8, Me, ß2'3hAla(aMe)); 1.12 (d, J = 7.2,

Me, ß2'3hAla(aMe)); 1.13 (d, J - 7.2, Me, ß2'3hAla(aMe)); 1.14 (d, J = 6.8, Me,

ß2'3hAla(aMc)); 1.29 (m, CH2, ß2hLeu); 1.42 (m, CH2, Lys); 1.62 (m, CH, ß2hLeu); 1.63

(m, 1H, CH2, Lys); 1.65 (m, CH2, Lys); 1.84 (m, 1H, CH2, Lys); 1.99 (m, CH, Val); 2.15

(m, CH, Val); 2.40 (ddJ = 7.8, 15.8, 1H, CH2, ß3hPhe); 2.47 (dd, J = 5.2, 16.2, 1H, CH2,

ß3hPhe); 2.56 (m, CH, ß2'3hAla(aMe)); 2.63 (m, CH, ß2,3hAla(aMe)); 2.80 (dd, J = 1A,

14.0, 1H, CH2, ß3hPhe); 2.87 (dd, J = 6.7, 12.4, 1H, CH2, ß3hPhe); 2.89 (m, 1H, CH2,

ß2hLeu); 2.90 (m, CH, ß2hLeu); 2.92 (m, CH2, Lys); 3.15 (m, 1H, CH2, ß2hLeu); 4.01 (m,

CH, ß2'3hAla(aMc)); 4.10 (m, CH, ß2'3hAla(aMe)); 4.12 (m, CH, Val); 4.15 (m, CH,

Val); 4.31 (m, CH, Lys); 4.46 (m, CH, ß3hPhe); 7.16-7.23 (m, 3 arow. H, ß3hPhe); 7.25-

7.28 (m, 2 arom. H, ß3hPhe); 7.98 (d, J = 8.8, NH, ß2'3hAla(aMe)); 7.94 (d, J = 8.5, NH,

ß2'3hAla(aMe)); 8.0 (d, J = 8.6, NH, Val); 8.02 (d, J = 8.8, NH, ß3hPhe); 8.19 (d, J =

7.6, NH, Lys); 8.23 (d, J = 7.5, NH, Val). 13C-NMR (150 MHz, CD3OH): 18.9

(ß2'3hAla(aMe)); 19.1 (Me, ß2hLeu); 19.4 (Me, ß2,3hAla(aMe)); 20.1 (Me, Val); 23.3

(Me, Val); 23.8 (CH2, Lys); 27.1 (CH2, ß2hLeu); 28.3 (CH2, Lys); 31.4 (CH, Val); 32.5

(CH2, Lys); 38.8 (CH2, ß3hPhe); 40.7 (CH, ß2hLeu); 40.8 (CH2, Lys); 41.3 (CH2,

ß3hPhc); 43.2 (CH2, ß2hLcu); 45.0 (CH, ß2'3hAla(aMe)); 46.2 (CH, ß2'3hAla(ccMe)); 49.3

160

(CH, ß2'3hAla(aMc)); 49.4 (CH, ß2'3hAla(aMe)); 49.5 (CH, ß3hPhe); 54.4 (CH, Lys);

60.4 (CH, Val); 61.3 (CH, Val); 127.7 (arom. C); 129.6 (arom. C); 130.4 (arom. C);

139.3 (arom. C); 173.1 (C=0); 173.3 (C=0); 173.5 (C=0); 174.8 (C=0); 175.9 (C=0);

177.2 0>O); 177.4 (CO). ESI-HR-MS: 831.5711/853.5393 (100/6, [M+H]7[M+Na]+;

calc. for [C43H75N808]+/[C43H74N808Na]+: 831.5702/853.5522).

H-^-ß'hPhe-His-^-ß^Leu-^-ß^Val-Ile-t^-ß^Tyr-OHtlS).

Fmoc-(5)-ß3hTyr(0'Bu)-OH (705 mg, 1.49 mmol) was loaded onto the Wang resin (298

mg, 1.00 mmol/g 100-200 mesh) according to (GP 1). The loading was determined to be

0.76 mmol/g (76%), corresponding to 0.23 mmol of Fmoc-(5)-ß3hTyr(0'Bu)-OH. After

capping (GP 2), the peptide was synthesized according to (GP 4a). After cleavage from

the resin (GP 5c), the crude peptide was purified by prep. RP-HPLC according to (GP

6d) (5 -10% B in 5 min, 10-50% B in 35 min, 50-99% B in 10 min) to yield 94 mg

(54%) of 25 as a TFA salt. White solid. RP-HPLC (5-10% B in 5 min, 10 -50% B in 35

min, 50-99% B in 10 min; tR 34.24 min): purity >95%. *H-NMR (600 MHz, CD3OH);

0.89 (d, J - 7.2, Me, He); 0.90 (d, J = 6.4, 2 Me, ß3hLeu); 0.93 (m, Me, ß3hVal); 1.12 (m,

CH2, Ile); 1.23-1.55 (m, CH2, ß2hLeu); 1.57 (m, CH, ß2hLeu); 1.70 (m, CH, Ile); 1.81 (m,

CH, ß3hVal); 2.39 (dd, J = 8.0, 16.0, 1H, CH2, ß3hTyr); 2.47 (m, 1H, CH2, ß3hTyr);

2.48-2.56 (m, CH2, ß3hVal); 2.57 (m, CH, ß2hLeu); 2.68-2.74 (m, CH2, His); 2.77 (m,

CH2, ß2hPhe); 2.90-3.01 (m, CH2, His); 3.01 (m, 1H, CH2, ß2hPhe); 3.20 (m, 1H, CH2,

ß2hPhe); 3.23-3.43 (m, CH2, ß2hLeu); 4.10 (t, J = 7.6, CH, Ile); 4.39 (m, CH, ß3hTyr);

4.65 (m, CH, His); 7.10 (m, 2H, ß3hTyr); 7.17 (m, 2H, ß3hTyr); 7.21-7.24 (m, 2H,

ß2hPhe); 7.25-7.29 (m, 3H); (ß2hPhe); 7.84 (d, J = 8.7, NH, ß3hVal); 7.90 (d, J = 8.5,

NH, ß3hTyr); 8.01 (d, J = 8.3, NH, Ile); 8.10 (t, J = 6.0, NH, ß2hLeu); 8.5 (d, J = 7.6,

NH, His). 13C-NMR (150 MHz, CD3OH): 15.9 (Me, Ile); 18.6 (Me, ß3hVal); 26.04 (CH2,

Ile); 26.04 (CH, Ile); 27.1 (Me, ß2hLcu); 27.2 (CH, ß2hLeu); 28.1 (CH2, His); 33.4 (CH,

ß3hVal); 37.5 (CH2, ß2hPhe); 37.9 (CH2, ß3hVal); 38.9 (CH2, ß3hTyr); 39.9 (CH2,

ß2hLeu); 40.5 (CH2, ß3hTyr); 43.0 (CH2, ß2hLeu); 43.3 (CH2, ß2hPhe); 49.8 (CH,

ß3hTyr); 53.5 (CH, ß3hVal); 53.9 (CH, His); 60.1 (CH, Ile); 127.5 (arom. C); 129.7

(arom. C); 130.2 (arom. C); 130.3 (arom. C); 131.5 (arom. C); 140.5 (arom. C); 172.4

(C=0); 173.2 (CO); 174.2 (C-O); 174.4 (C=0); 175.2 (C=0); 176.7 (C=0).

ESI-HR-MS: 847.5061/869.4842 (100/12, [M+H]7[M+Na]+; calc. for

[C45H67N808]+/[C45H66N808Na]+: 847.5076/869.4896).

161

a/y- and ß/y-Peptides

H-Ala-(y?)-y4hLeu-Glu-(iî)-y4hVal-Lys-(iî)-y4hLeu-Ser-(/?)-Y4hVal-Phe-OH(26).Fmoc-Phe-OH (775 mg, 2 mmol) was loaded onto the Wang resin (200 mg, 1.00 mmol/g,

100-200 mesh) according to (GP 1). The loading was determined to be 0.70 mmol/g

(70%), corresponding to 0.10 mmol of Fmoc-Phe-OH. After capping (GP 2), the peptide

synthesis was performed according to (GP 4a). Treatment of the peptide-resin according

to (GP 5a) afforded the crude peptide 4 (100 mg). Purification of a part of the crude

peptide (15 mg) by RP-HPLC (5-15% B in 40 min) according to (GP 6a) yielded 10 mg

of 26 (calc. overall yield 50%) as TFA salt. White solid. RP-HPLC (5-10% B in 5 min,

10-50% B in 40 min, 50-95% B in 50 min; tR 24.90 min): purity >95%. *H-NMR (600

MHz, CD3OH): 0.80 (m, 2Me, y4hLeu); 0.89 (d, J = 6.5 Hz, 2Mc, y4hLeu ); 0.90 (m,

2Me, y4hVal ); 0.91(4 J = 7.0, 2Me, y4hVal); 1.20 (m, CH2, y4hLeu); 1.30 (m, CH2,

y4hLeu ); 1.55 (d, J = 7.0, Me, Ala); 1.61(m, 1H, CH2, y4hVal); 1.67 (m, CH2, Lys); 1.71

(m, 1H, CH2, Lys); 1.72 (m, CH, y4hVal ); 1.78 (m, CH, y4hVal); 1.85 (m, 1H, CH2,

y4hVal); 1.83 (m, 1H, CH2, Lys); 1.91 (m, CH, y4hLeu); 1.92 (m, CH2, y4hLeu ); 1.98 (m,

CH2, Glu); 2.09 (m, CH2, y4hVal ); 2.17 (m, CH2, y4hVal); 2.19 (772, 1H, CH2, y4hVal );

2.24 (w, 1H, CH2, y4hLeu); 2.29 (m, CH2, y4Leu ); 2.30 (m, 1H, CH2, y4hVal ); 2.37 (m,

1H, CH2, y4hLeu); 2.45 (m, CH2, Glu); 2.91 (m, CH2, Lys); 3.18 (m, 1H, CH2, Phe); 3.50

(m, CH, y4hVal); 3.65(m, CH, y4hVal); 3.82 (m, CH2, Ser); 3.90 (m, CH, Ala); 3.91 (m,

CH, y4hLeu ); 3.92 (m, CH, y4hLeu); 3.96 (m, 1H, CH2, Phe); 4.12 (m, CH, Lys ); 4.23

(m, CH, Glu ); 4.36 (m, CH, Ser); 4.57(w, CH, Phe); 7.15-7.23 (m, 3 arom. H, Phe);

7.25-7.28 (m, 2 arom. H, Phe); 7.49 (d, J = 9.8, NH, y4hVal ); 7.90 (d, J = 9.10, NH,

y4hLeu ); 7.91 (d, J = 9.10, NH, y4hVal); 8.10 (d, J - 6.10, NH, Lys ); 8.04 (d, J = 7.19,

NH, Ser); 8.09 (d, J = 9.1, NH, y4hLeu ); 8.15 (d, J - 7.70, NH, Phe); 8.26 (d, J = 6.38,

NH, Glu). 13C-NMR (150 MHz, CD3OH): 18.2 (Me, Ala); 20.13 (CH2, Glu); 22.37 (Me,

y4hLeu); 23.36 (Me, y4hVal); 23.70 (Me, y4hLeu); 24.25 (CH2, Lys); 28.34 (CH2, Lys);

28.90 (CH2, y4hVal); 29.15 (CH2, y4hVal); 31.78 (CH2, Glu); 32.50 (CH2, Lys); 32.52

(CH, y4hLeu); 32.65 (CH, y4hLeu); 32.91 (CH2, y4hLeu); 33.0 (CH2, y4hVal); 33.08 (CH2,

y4hLeu); 33.31 (CH, y4hVal); 33.60 (CH, y4hVal); 33.65 (CH2, y4hVal); 38.50 (CH2,

Phe); 40.87 (CH2, Lys); 45.01 (CH2, y4hLeu); 45.21 (CH2, y4hLeu); 47.90 (CH, y4hLeu);

48.34 (CH, y4hLeu); 50.56 (CH, Ala); 55.20 (CH, y4hVal); 55.75 (CH, Phe); 55.83 (CH,

162

Glu), 56.24 (CH, y4hVal); 56.50 (CH, Lys); 58.01 (CH, Scr); 63.55 (CH2, Ser); 127.83

(arom. C); 129.52 (arom. C); 130.34 (arom. C); 138.61 (arom. C); 171.01 (C=0); 172.88

(C=0); 173.96 (C=0); 174.07 (CO); 175.98 (C=0); 176.04 (CO); 176.10 (C=0);

176.18 (CO); 176.75 (C=0). MALDI-HR-MS: ([M+H]+; 1117.7231 cale. For

[C56H96Nio013]+: 1116.7158).

H-Ala-(Ä)-y4hVal-Ser-(Ä)-y4hLeu-Glu-(i?)-y4hVal-Lys-(Ä)-y4hVal-Phe-OH(27).

Fmoc-Phe-OH (775 mg, 2 mmol) was loaded onto the Wang resin (200 mg, 1.00 mmol/g,

100-200 mesh) according to (GP 1). The loading was determined to be 0.70 mmol/g

(70%), corresponding to 0.10 mmol of Fmoc-Phe-OH. After capping (GP 2), the peptide

synthesis was performed according to (GP 4a). Treatment of the peptide-resin according

to (GP 5a) afforded the crude peptide 27 (100 mg). Purification of a part of the crude

peptide (15 mg) by RP-HPLC (5-15% B in 40 min) according to (GP 6a) yielded 10 mg

of 27 (calc. overall yield 50%) as TFA salt. White solid. RP-HPLC (5-10% B in 5 min,

10-50% B in 40 min, 50-95% B in 50 min; tR 29.94 min): purity >95%. 'H-NMR (600

MHz, CD3OH): 0.87 (d, J = 7.12, 2Me, y4hVal); 0.89 (d, J - 6.8, 2Me, y4hLeu); 0.92 (d,

J = 6.8, 2Me); 0.94 (d, J = 6.5, 2Me, y4hVal); 1.25 (m, IH, CH2, y4hLeu); 1.73 (m, IH,

CH2, Lys); 1.48 (m, IH, CH2, y4hLeu); 1.49 (m, IH, CH2, Lys); 1.54 (m, IH, CH2, Lys);

1.57 (d, J = 7.0, Me, Ala); 1.58 (m, IH, CH2, y4hVal); 1.62 (m, IH, CH2, y4hVal); 1.75

(m, IH, CH2, y4hVal); 1.68 (m, CH, y4hLeu); 1.74 (m, CH, y4hVal); 1.76 (m, CH2, Lys);

1.82 (m, CH, y4hVal); 1.87 (m, IH, CH2, Lys); 1.92 (m, IH, CH2, y4hVal); 1.96 (m, IH,

CH2, Glu); 2.00 (m, IH, CH2, y4hVal); 2.01 (m, CH2, y4hLeu); 2.03 (m, CH2, y4hVal);

2.10 (m, IH, CH2, Glu); 2.14 (m, IH, CH2, y4hVal); 2.18 (m, CH2, y4hVal); 2.23 (m, IH,

CH2, y4hLeu); 2.25 (m, IH, CH2, y4hVal); 2.30 (m, IH, CH2, y4hVal); 2.32 (m, IH, CH2,

y4hVal); 2.40 (m, IH, CH2, y4hLeu); 2.45 (m, CH2, Glu); 2.92 (m, CH2, Lys); 2.96 (m,

IH, CH2, Phe); 3.18 (m, IH, CH2, Phe); 3.53(/m, CH, y4hVal); 3.63 (m, CH, y4hVal); 3.67

(m, CH, y4hVal); 3.85 (m, CH2, Ser); 3.93 (m, CH, y4hLeu); 3.94 (m, CH, Ala); 4.14 (m,

CH, Glu ); 4.16 (m, CH, Lys ); 4.27 (m, CH, Ser); 4.58 (m, CH, Phe); 7.70 (d, J = 9.76,

NH, y4hLeu ); 7.89 (d, J = 9.76, NH, y4hVal); 7.97 (d, J = 6.38, NH, Ser); 8.04 (d, J =

9.31, NH, y4hVal ); 8.05 (d, J = 9.30, NH, y4hVal); 8.16 (d, J = 7.70, NH, Phe); 8.20 (d,

J = 6.02, NH, Glu ); 8.39 (d, J = 6.24, NH, Lys ); 7.16-7.23 (m, 3 arom. H, Phe); 7.24-

7.28 (m, 2 arom. H, Phe). 13C-NMR (150 MHz, CD3OH): 18.29 (Me, Ala); 18.92 (Me,

y4hVal); 19.01 (Me, y4hVal); 19.84 (Me, y4hLeu); 23.70 (Me, y4hVal); 26.10 (CH,

163

y4hLeu); 28.14 (CH2, Lys); 28.53 (CH2, Glu); 29.12 (CH2, y4hVal); 31.52 (CH2, Glu);

32.31 (CH2, y4hLeu); 32.81 (CH2, y4hLeu); 32.51 (CH2, Lys); 33.18 (CH2, y4hVal); 33.22

(CH, y4hVal); 33.47 (CH, y4hVal); 33.48 (CH2, y4hVal); 33.62 (CH, y4hVal); 33.77 (CH2,

y4hVal); 38.48 (CH2, Phe); 45.63 (CH2, y4hLeu); 47.44 (CH, y4hLeu); 47.78 (CH2, Lys);

50.58 (CH, Ala); 54.99 (CH, y4hVal); 55.73 (CH, Phe); 55.98 (CH, y4hVal); 56.25 (CH,

y4hVal); 56.38 (CH, Glu); 56.50 (CH, Lys); 58.20 (CH, Ser); 62.94 (CH2, Ser); 127.80

(arom. C); 129.48 (arom. C); 130.33 (arom. C); 138.62 (arom. C); 171.39 (C=0); 172.17

(C=0); 174.17 (C=0); 174.84 (C=0); 174.91 (C=0); 175.87 (C=0); 176.23 (C=0);

176.41 (CO); 176.45 (C=0). MALDI-HR-MS: ([M+H]'; 1103.7018 calc. For

[C55H94N1o013J+: 1102.7002).

H-ß3hSer-(/f)-y4hVal-ß3hLys-(i?)-Y4hLeu-ß3hTyr-(if)-y4hVal-ß3hAla-OH(28).

Fmoc-ß3hAla-OH (651 mg, 2 mmol) was loaded onto the Wang resin (200 mg, 1.00

mmol/g, 100-200 mesh) according to (GP J). The loading was determined to be 0.75

mmol/g (75%), corresponding to 0.15 mmol of Fmoc-ß3hAla-OH. After capping (GP 2),

the peptide synthesis was performed according to (GP 4a). Treatment of the peptide-resin

according to (GP 5a) afforded the crude peptide 28 (100 mg). Purification of a part of the

crude peptide (15 mg) by RP-HPLC (5-15% B in 40 min) according to (GP 6a) yielded

10 mg of 28 (calc. overall yield 50%) as TFA salt. White solid. RP-HPLC (5-10% B in 5

min, 10-50% B in 40 min, 50-95% B in 50 min; tR 17.79 min): purity >95%. 'H-NMR

(600 MHz, CD3OH): 0.87 (m, 2Me); 0.88 (d, Me, y4hVal); 0.90 (d, J = 6.8, 2 Me); 1.17

(d, J = 6.8, ß3hAla); 1.40 (m, 1H, CH2, ß3hLys); 1.44 (m, 1H, CH2, ß3hLys); 1.52 (m,

CH2, ß3hLys); 1.53 (m, 1H, CH2, y4hVal); 1.57 (m, 1H, CH2, y4hVal); 1.60 (m, 1H, CH2,

y4hVal); 1.66 (m, CH2, ß3hLys); 1.68 (m, 1H, CH2, y4hVal); 1.73 (m, 1H, CH2, y4hVal);

1.70 (m, 1H, CH2, y4hVal); 1.74 (m, CH, y4hVal); 1.84 (m, CH, y4hVal); 1.90 (m, CH,

y4hVal); 2.12 (m, CH2, y4hVal); 2.16 (m, 1H, CH2, y4hVal); 2.23 (m, 1H, CH2, y4hVal);

2.39 (m, CH2, ß3hAla); 2.41 (m, CH2, ß3hTyr); 2.43 (m, CH2, ß3hLys); 2.63 (m, CH2,

ß3hSer); 2.65 (m, 1H, CH2, ß3hTyr); 2.77 (m, 1H, CH2, ß3hTyr); 3.08 (m, CH2, y4hVal);

3.60 (m, CH, ß3hSer); 3.61 (m, CH, y4hVal); 3.65 (m, CH, y4hVal); 3.68 (m, CH, y4hVal);

3.77 (m, CH2, ß3hSer); 4.17 (m, CH, ß3hAla); 4.20 (m, CH, ß3hLys); 4.38 (m, CH,

ß3hTyr); 6.69 (m, 2 aryl-H) (ß3hTyr); 7.02 (m, 2 aryl-H) (ß3hTyr); 7.73 (d, J - 9.39,

NH, y4hVal); 7.79 (d, J = 8.48, NH, ß3hTyr); 7.66 (d, J = 9.24, NH, y4hVal); 7.85 (d, J =

164

8.66, NH, ß3hLys); 7.96 (d, J = 9.24, y4hVal); 7.90 (d, J = 7.92, NH, ß3hAla). 13C-NMR

(150 MHz, CD3OH): 18.69 (Me, y4hVal); 19.70 (Me, y4hVal); 20.57 (Me, ß3hAla); 23.89

(CH2, p3hLys); 28.95 (CH, y4hVal); 29.20 (CH, y4hVal); 29.24 (CH, y4hVal); 29.28 (CH2,

y4hVal); 34.02 (CH2, y4hVal); 34.22 (CH2, y^Val); 34.31 (CH2, y4hVal); 35.02 (CH2,

ß3hLys); 35.30 (CH2, ß3hSer); 40.72 (CH2, ß3hTyr); 41.64 (CH2, ß3hAla); 41.79 (CH2,

ß3hTyr); 42.37 (CH2, ß3hLys); 43.84 (CH, ß3hAla); 47.94 (CH, ß3hLys); 50.16 (CH,

p3hTyr); 52.12 (CH, ß3hScr); 55.46 (CH, y4hVal); 55.51 (CH, y4hVal); 55.65 (CH,

y4hVal); 62.58 (CH2, p3hSer); 116.30 (arom. C); 130.17 (arom. C); 131.52 (arom. C);

157.24 (arom. C); 171.84 (C=0); 173.31 (C=0); 1743.47 (C=0); 174.93 (C=0); 175.15

(C=0); 175.34 (C=0). MALDI-HR-MS: ([M+Hf; 905.6214 calc. For [C46H80N8Oio]+:

904.5997).

7.5 NMR Measurements and Structure Calculation

GeneralRemarks

NMR spectra of peptides were acquired at 600 MHz (lH) / 150.9 MHz (13C) with

presaturation of the solvent signal; 90k data points, 128 scans, 5.6s acquisition time.

{'HJ-BB-decouplcd 13C-NMR: 80-K data points, 20-K scans, 1.3-s acquisition time, Is

relaxation delay, 45° excitation pulse. Processed with 1.0-Hz exponential line

broadening. The following spectra were used for the resonance assignments: [ H, H]

DQF-COSY, [^HJ-TOCSY, [^C/HJ-HSQC, ["C^HJ-HMBC, [^HJ-ROESY. 2D-

NMR: solvent suppression with presat. DQF-COSY (500 MHz, CD3OH) with pulsed

field gradients (PFG) for coherence pathway selection: acquisition: 2K(t2) x 512(ti) data

points. 10 scans per ti increament, 0.17-s acquisition time in t2; relaxation delay 2.0 s.

TPPI quadrature detection in ti. Processing: zero filling and FT to IK x IK real/real data

points after multiplication with a sin2 filter shifted by n/2 in tl. HSQC with PFG (500,

125 MHz, CD3OH): acquisition: 2K(t2) x 512 (ti) data points, 48 scans per ti-increment.

13C-GARP decoupling during t2. 0.17-s acq. time in t2. Processing: zero filling and FT to

IK x IK real/real data points after multiplication with a sin2 filter shifted by n/2 in ti and

sin filter shifted by n/2 in t2. HMBC with PFG (500, 125 MHz, CD3OH): acquisition:

delay for evolution of long-range antiphase magn.: 50 ms. No l3C decoupling, otherwise

identical to parameters for HSQC. Processing: zero filling and FT to IK x IK after

multiplication with a sin2 filter in t2 and a Gaussian filter in ti; power spectrum in both

165

dimensions. ROESY (500 MHz, CD3OH). Acquisition: a ROESY spectrum with a

mixing time of 300 ms was acquired. CW-spin lock (2.7 kHz) between trim pulses,

2K(t2) x 512(ti) data points, 64 scans per ti increment. 0.17s acq. time in t2, other

parameters identical to DQF-COSY. Processing: zero filling and FT to IK x 512K

real/real data points after multiplication by a cos2 filter in t2 and ti. Baseline correction

with 3rd degree polynomial in both dimensions.

Assignments and volume integration of ROESY cross peaks were performed with the

aid of SPARKY [3]. Distance constraints and error limits were generated from cross peak

volumes by calibration with known distances (two-spin approximation, ± 20% error

limits) through a python extension within SPARKY. The volumes of cross peaks

involving methyl groups or other groups of isochronous protons were corrected by

division through the number of protons.

Generation of distance restraints

ROESY spectra in CD3OH with mixing time 300 ms were used. Possible contributions

from spin diffusion were excluded from the generation of distance restraints. Also cross

peaks with signal to noise ratio <1 at mixing time 300 ms and peaks overlapping more

than 30% at their basis were exluded. A distance of 3.0 Â between backbone NH and Hß,

and of 1.9 Â between Hœc-aC and Heç-aC were used as reference values to generate

distance restraints with upper and lower bounds as ± 20% of the calculated value.

Simulated annealing (SA) structure calculations

Program XPLOR-NIH v2.9.7 [4J. The standard parameter and topology files of XPLOR-

NIH (parallhdg.pro; topallhdg.pro) were modified to accommodate ß2-, ß3-aminoacid

residues. The SA calculation protocol (adopted from the torsional angle dynamics

protocol of Stein et al. [5] included 4000 steps (0.015 ps each) of high temperature

torsional angle dynamics at 20000K, followed 4000 (0.015 ps) steps of slow cooling to

1000K with torsion angle dynamics, 4000 steps (0.003 ps) of slow cooling with cartesian

dynamics to 300K and a final conjugate gradient minimization. The only non-bonded

interactions used were van der Waals repulsion functions. For each compound, at least 30

structures were calculated.

166

g-F/OH-ß-PeptidesTable 30. NOEs observed in the 300 ms ROESY spectrum of pepüde 3 in MeOH

Residue H-Atom Residue H-Atom dNOE[A] Residue H-Atom Residue H-Atom dNOE [A]

1 ß 1 Y 2.6 6 HN 6 Y 3.2

2 P 2 NU 2.9 7 HN 7 P 3.0

2 HN 2 Y 3.2 7 HN 7 Y 3.3

3 HN 3 a 2.8 7 HN 7 5 3.6

3 HN 3 ß 3.0 1 a 2 HN 2.5

3 HN 3 6 3.7 1 P 2 P 4.9

3 HN 3 «Si 3.3 1 8 2 HN 5.2

3 HN 3 ccrs 5.7 2 Otfe 3 HN 2.6

4 Y 4 &Rs 3.3 3 a 4 HN 2.5

4 HN 4 «te 2.8 3 P 4 HN 3.1

4 HN 4 ß 3.1 4 <*R£ 5 HN 3.2

4 HN 4 Y 2.9 4 HN 5 HN 5.3

5 HN 5 as. 3.9 4 P 5 HN 3.2

5 HN 5 CRs 3.1 4 Y 5 HN 4.3

5 HN 5 p 2.9 3 HN 4 HN 4.3

5 HN 5 Y 2.9 6 HN 5 aRe 3.4

6 HN 6 ß 3.0

Table 31. NOEs observed in the 300 ms ROESY spectrum of peptide 4 in MeOH

Residue H-Atom Residue H-Atom dNOE [A] Residue H-Atom Residue H-Atom dNOE[A]

1 P 1 8 2.8 6 HN 6 P 2.9

1 P 1 Y 2.3 6 HN 6 Y 3.5

2 P 2 a& 2.3 7 HN 7 Y 3.6

2 HN 2 Y 2.8 P 2 P 4.5

3 5 3 as. 2.7 P 2 HN 4.4

3 Y 3 a 2.9 P 2 Y 5.2

3 HN 3 Ö 3.5 OtSi 2 HN 3.3

4 a 4 Y 3.0 OtRe 2 HN 2.4

4 HN 4 a 3.5 2 P 3 HN 4.1

4 HN 4 Y 3.8 2 HN 3 HN 3.5

4 HN 4 Y 2.9 4 a 5 8 4.2

5 P 5 Ö 2.8 5 P 6 HN 4.2

5 HN 5 Y 3.2 5 5 4 ß 4.4

5 HN 5 P 29 5 P 6 HN 4.2

6 asi 6 Y 3.0 5 Y 4 a 4.9

6 HN 6 CCrj 2.8 5 HN 4 P 3.3

6 P 6 Y 2.6

167

Table 32. NOEs observed in the 300 ms

Residue H-Atom Residue H-Atom dN0E[A]

2 NH 1 a 4.3

2 a 3 NH 2.3

2 NH 3 P 4.0

2 ß 3 NH 4.2

2 NH 3 NH 4.7

3 a 4 NH 3.6

3 NH 4 NH 3.5

4 NH 5 NH 3.5

5 aRe 6 NH 3.2

5 P 6 Y 4.9

5 P 6 NH 4.0

5 a& 6 NH 5.1

6 «Si 7 NH 2.9

6 aRe 7 NH 2.6

1 «Re 4 P 2.7

Table 33. NOEs observed in the 300 ms

Residue 11-Atom Residue H-Atom dN0E[A]

2 «a 2 P 2.4

2 P 2 HN 2.9

2 y 2 HN 3.6

3 P 3 a 2.4

3 P 3 HN 2.9

3 8 3 HN 3.4

3 Y 3 un 3.4

4 P 4 HN 3.0

4 Y 4 HN 3.5

4 a 4 HN 2.9

5 aRe 5 HN 3.2

5 as. 5 HN 2.4

5 P 5 HN 3.0

5 Y 5 HN 3.2

6 P 6 HN 2.9

ROESY spectrum of peptide 5 in MeOH

Residue 11-Atom Residue H-Atom dN0E|A]

2 aRe 5 P 2.4

2 Y 5 P 3.7

2 NH 5 P 3.3

3 a 6 P 2.3

3 NH 5 P 3.6

3 Nil 6 P 3.2

3 NH 6 Y 5.1

4 HB 1 8 4.3

4 HN 6 P 3.7

4 HN 7 P 3.2

2 P 5 S 4.4

2 Y 5 HN 5.1

5 HN 7 P 3.8

7 P 4 Y 3.7

ROESY spectrum of peptide 6 in MeOH

Residue H-Atom Residue H-Atom dN0ElAj

2 P 3 HN 3.6

2 HN 1 aSl 2.6

2 HN 1 aRe 23

2 HN 3 p 4.6

4 a 5 HN 2.6

4 P 5 HN 3.2

5 as. 6 HN 3.1

5 a Re6 UN 2.3

3 p 2 UN 4.6

7 HN 6 a 2.8

1 as. 4 ß 5.8

1 (XRc 4 a 3.3

2 aRe 4 a 3.7

2 »Re 5 P 2.6

2 HN 4 a 4.3

168

6 a 6 ß 2.4 2 HN 4 ß 3.7

7 a si7 HN 2.6 2 HN 5 ß 3.8

7 cxr* 7 HN 2.8 3 HN 5 ß 3.5

7 ß 7 HN 2.9 4 Y 2 HN 4.8

1 ß 2 HN 4.4 4 ß 6 a 3.9

2 a Re3 HN 2.5

Table 34. NOEs observed in the 300 ms ROESY spectrum of peptide 7 in MeOH

Residue H-Atom Residue H-Atom dNOE [Â] Residue H-Atom Residue H-Atom dN0E[A]

1 ctsi 2 HN 3.0 5 ß 6 HN 4.2

1 WRe 2 HN 2.6 6 «Si 7 HN 2.9

1 ß 2 HN 4.1 6 aRe 7 HN 2.7

2 «Si 3 HN 3.8 «Si 4 a 5.1

2 OlRc 3 HN 2.5 a& 4 ß 3.1

1 aSi 4 ß 3.1 «Re 4 HN 2.4

1 «Re 4 ß 2.5 ß 4 ß 5.1

1 ß 2 ß 4.4 8 4 a 4.6

1 Y 4 ß 3.9 2 aR= 5 ß 2.4

2 ß 3 HN 4.2 2 OCR* 6 Y 4.6

2 HN 3 ß 5.1 2 Y 5 ß 3.6

2 HN 3 HN 4.3 2 HN 4 a 4.8

3 «Si 4 HN 3.8 2 HN 5 ß 3.4

3 CtRe 4 HN 2.5 2 HN 5 Y 5.3

3 8 4 ß 4.4 3 «Rc 6 ß 2.3

3 HN 4 HN 4.0 3 OCRs 6 Y 5.6

4 a 5 Y 4.4 3 HN 5 ß 3.7

4 a 5 HN 3.2 3 HN 6 ß 3.0

4 ß 5 HN 4.2 3 HN 6 Y 5.6

3 ß 4 HN 4.2 4 HN 6 ß 3.1

4 HN 5 ß 5.0 4 HN 6 Y 4.9

4 HN 5 HN 4.0 5 HN 7 ß

4 a 5 aa 5.2

169

Table 35. NOEs observed in the 300 ms

Residue II-Atom Residue H-Atom dN0E[A]

1 al 1 a2 1.9

1 ß 1 a* 3.0

2 P 2 y* 3.2

2 P 2 HN 2.9

2 5 2 HN 3.4

2 yl 2 Y2 2.0

3 a 3 HN 3.2

3 Y* 3 P 3.2

3 ß 3 HN 3.1

4 P 4 HN 2.9

4 ß 4 Y 2.5

4 Y 4 HN 3.2

5 Y 5 HN 2.8

5 P 5 HN 2.9

5 a* 5 HN 3.2

6 a* 6 HN 3.2

6 P 6 HN 3.0

7 a 7 HN 2.7

7 Y* 7 a 3.0

7 Y* 7 ß 3.0

7 Y* 7 HN 3.4

8 a* 8 HN 3.2

8 a* 8 P 2.8

8 Y* 8 P 3.0

8 P 8 HN 3.0

8 yl 8 Y2 2.0

9 a* 9 HN 3.2

9 y* 9 P 3.2

9 ß 9 HN 3.0

10 Y 10 ß 2.5

10 ß 10 HN 2.9

10 Y 10 HN 3.2

11 a* 11 HN 3.2

11 Y* 11 ß 3.0

11 P 11 HN 3.0

ROESY spectrum ofpeptide 8 in MeOH

Residue H-Atom Residue H-Atom dN0E[A]

1 a* 2 HN 2.9

2 HN 3 HN 4.3

4 HN 5 HN 4.3

5 HN 6 HN 4.2

6 a* 7 HN 2.9

6 HN 6 HN 4.4

7 a 8 HN 2.5

8 HN 9 HN 4.2

9 HN 10 HN 4.3

10 HN 11 HN 4.4

11 a* 12 HN 3.0

12 HN 13 P 5.0

13 HN 12 a* 2.9

1 a* 4 P 2.8

1 Y* 4 P 3.9

2 y* 5 P 3.3

3 HN 5 HN 4.9

3 HN 6 P 3.5

3 Y* 6 P 3.9

4 HN 6 HN 5.0

4 a* 7 a 3.0

4 Y 7 ex 3.8

4 Y 7 P 3.7

4 HN 7 a 3.8

4 Y 7 Y* 4.5

5 HN 8 ß 3.2

6 HN 8 P 3.4

6 a* 9 P 3.1

7 a 10 P 2.8

7 Y* 10 ß 3.5

7 Y* 10 Y 4.5

7 HN 10 P 3.0

8 HN 10 P 3.6

8 HN 11 ß 3.0

9 P 6 HN 3.8

170

11 Y* il HN 3.5

12 ß 12 HN 3.0

12 a* 12 HN 3.1

12 y* 12 HN 3.5

13 a* 13 HN 3.1

13 Y* 13 ß 3.0

13 P 13 HN 3.0

13 Y* 13 HN 3.6

1 P 2 HN 4.3

Note: * = pseudoatom used for calculations

Mixed Peptides

Residue H-Atom Residue II-Atom dNOE[A]

1 a 1 P* 3.5

2 a 2 P 2.6

2 a 2 Y 3.5

2 a 2 HN 3.0

2 P 2 HN 2.8

2 y* 2 HN 3.3

3 a 2 P* 2.9

3 a 3 Y* 3.1

3 a 3 HN 3.3

3 ß* 3 HN 3.2

3 Y* 3 HN 3.7

4 a 4 P* 3.1

4 a 4 UN 3.2

5 a 5 P* 2.9

5 a 5 HN 3.2

5 Y 5 P* 3.3

5 HN 5 P* 3.4

5 HN 5 Y 3.2

Note: *= pseudoatom used for calculations

9 P 7 HN 3.9

9 HN 11 P 3.5

10 ß 6 HN 4.6

10 HN 7 HN 4.7

10 HN 8 HN 4.6

10 HN 13 P 3.1

12 P 9 HN 2.8

12 P 10 HN 3.8

13 Y* 10 Y 5.0

Residue H-Atom Residue H-Atom dNOE[A]

6 a 6 ß* 3.5

6 a 6 HN 3.2

6 P* 6 HN 3.3

7 P* 7 a 3.4

7 P* 7 HN 3.5

2 a 3 a 4.9

2 a 3 HN 2.4

2 P 3 HN 3.2

2 Y* 3 HN 4.5

3 a 4 HN 2.4

3 P* 4 HN 4.1

3 HN 4 HN 3.8

5 a 6 HN 2.4

6 P* 7 HN 4.2

2 a 5 a 4.2

2 a 6 P* 4.9

3 HN 5 a 5.0

4 HN 2 a 3.6

Table 36. NOEs observed in the 300 ms ROESY spectrum of pepüde 14 in MeOH

171

Table 37. NOEs observed in the 300 ms ROESY spectrum of peptide 19 in MeOH

Residue H-Atom Residue H-Atom dNOE[A] Residue H-Atom Residue H-Atom dNOE[A]

1 a* 1 P 3.2 5 a* 6 HN 3.2

1 5 1 P 3.3 5 P 6 HN 3.5

2 HN 2 Pa* 3.3 5 HN 4 Ps,* 3.2

3 Y 3 P 2.8 5 HN 4 HN 3.3

3 Y 3 HN 3.0 6 HN 5 HN 4.0

3 a* 3 HN 3.3 6 HN 5 Y* 4.8

3 ß 3 a* 3.2 7 HN 6 Pa* 3.2

4 Ps,* 4 HN 3.2 7 HN 6 HN 3.4

5 P 5 a* 3.1 a* 3 HN 3.4

5 P 5 p* 3.0 P 3 HN 5.0

5 P 5 HN 3.0 P 4 Psi* 4.4

5 P* 5 HN 3.5 P 4 HN 5.2

5 a* 5 HN 3.2 P 5 P 5.0

6 pSl* 6 HN 3.2 8 5 P 5.2

7 P 7 P* 3.2 2 pR=* 5 OlR* 4.0

7 HN 7 P* 3.3 7 HN 6 Pa* 3.2

7 P 7 ß* 3.0 7 HN 6 HN 3.4

7 P 7 HN 3.0 1 a* 3 HN 3.4

7 HN 7 p* 3.3 3 P 6 Psi* 4.0

1 P* 2 HN 2.8 3 P 6 HN 4.1

1 P 2 HN 4.2 3 P 7 HN 5.0

2 HN 1 5 5.2 4 Pr.* 7 HN 4.7

2 HN 3 P* 3.4 5 P 7 a* 5.0

3 HN 2 HN 3.3 4 Pr/ 7 a* 4.6

3 P 4 HN 3.2 5 P 7 HN 4.0

4 HN 3 Y 4.9 6 Psi* 3 Y 5.1

4 HN 3 P* 2.8 7 HN 5 a* 4.9

Note: * = pseudoatom used for calculations

172

Table 38. NOEs observed in the 300 ms ROESY spectrum of pepüde 22 in MeOH

Residue H-Atom Residue H-Atom ^NOE [A] Residue H-Atom Residue H-Atom 4lOE [Â]

2 HN 2 a2 3.2 2 ccRe 2 HN 2.7

2 P 2 Y* 3.1 2 ß 3 UN 3.2

2 ß 2 «si 3.0 2 HN 3 HN 4.0

2 P 2 HN 3.0 4 P 5 HN 3.5

2 Y* 2 HN 3.5 4 HN 3 HN 3.3

3 HN 3 ßSl* 3.2 4 HN 3 ßSl* 3.0

4 P 4 a* 3.1 5 HN 4 HN 4.0

4 P 4 HN 2.9 5 HN 4 a* 2.9

4 UN 4 a* 3.4 5 ßSl* 6 HN 3.0

4 HN 4 Y* 3.5 6 HN 5 HN 3.3

6 HN 6 a-te 3.0 6 «Re 7 HN 2.7

6 HN 6 as. 3.9 6 P 7 HN 3.2

6 HN 6 ß 2.9 7 HN 6 HN 3.7

6 ß 6 asi 3.0 7 un 8 HN 3.2

6 P 6 g* 30 2 P 4 HN 4.0

6 HN 6 Y* 3.4 2 P 5 PSl* 3.8

8 ß 8 a* 3.1 3 P*,* 6 «Ke 3.4

8 HN 8 a* 3.5 4 ß 6 HN 4.0

8 P 8 Y 2.7 4 P 7 HN 4.3

8 HN 8 Y 3.0 4 P 7 P&* 3.3

8 HN 8 ß 2.9 5 HN 2 P 3.0

1 pSl* 2 HN 3.1 5 ßfe* 8 a* 3.7

Note: * = pseudoatom used for calculations

173

Table 39. NOEs observed in the 300 ms ROESY spectrum of peptide 25 in MeOH

Residue H-Atom Residue Atom dNOE [A] Residue H-Atom Residue H-Atom dNoE [A]

2 a 2 ß* 3.0 1 p. 2 HN 3.6

2 a 2 HN 3.2 1 P* 2 HN 4.0

3 a 3 r 3.1 2 a 3 HN 2.6

3 a 3 HN 3.3 2 HN 3 HN 3.7

3 a 3 Ps, 2.7 3 Ps, 4 HN 3.7

3 a 3 Pr* 3.1 3 Prc 4 HN 2.9

3 UN 3 ßsi 3.0 3 HN 4 p* 3.9

3 HN 3 Pro 2.8 3 HN 4 P 4.4

3 HN 3 Y* 4.2 4 asi 5 HN 2.8

4 P 4 Y 2.6 4 P 5 HN 3.3

4 P 4 HN 3.0 4 HN 3 HN 4.0

4 Y 4 HN 2.9 5 a 6 HN 2.4

5 a 5 P 2.7 5 P 6 HN 3.0

5 y* 5 HN 3.1 6 P 5 P 4.9

5 a 5 UN 3.1 6 HN 5 Y* 4.2

5 P 5 HN 2.9 2 a 5 a 3.4

6 P 6 HN 3.0 2 a 5 P 5.0

6 a* 6 HN 2.9 2 a 6 HN 4.1

6 a* 6 P 2.8 3 HN 5 a 4.0

Note: * = pseudoatom used for calculations

ßMcosapeptide__^

Table 40. NOEs observed in the 400 ms NOESY spectrum of peptide 29 in MeOH

Residue H-Atom Residue H-Atom dNOE [A] Residue HAtom Residue H-Atom dN0E [A]

1 OtRe 1 P 2.9 HN 3 2 HN 3.4

2 P 2 Y 2.6 3 «Si 4 HN 2.6

2 P 2 HN 2.9 4 «Re 5 HN 2.5

2 Y 2 HN 3.0 4 HN 5 HN 3.3

3 P 3 aRe 2.6 5 «Si 6 HN 2.6

3 P 3 Y 2.5 6 «Si 7 HN 2.7

3 P 3 HN 3.0 6 HN 5 HN 3.2

3 Y 3 HN 3.1 6 HN 7 HN 3.4

3 CCr* 3 HN 2.8 7 asi 8 HN 2.7

s

§s§Jg

|!§

§*b

§«^§

ö§

a»Ig

^§

§#§

^§

§§

J§

§^^§

I^§

t-c«c»osc>222:=;222222222^22^^"2^^r'>m*:î-''~>,''iïo^f--000000

t^C>OOO^OO^C>Ot^^OO^C»C«[^O^OtX)f^C>00\OCVOOVOt^TrC\OOOt^OOOOsOU-i^^

^Tj-'a-w-iuimu-bu-ivoy3vor^r-~ooooooCTio\222m!il!^"^^l

——"22222222

Tj-Tj--^-inu-%iriw-1u-ivooy3t^t-~ooooc<;o\a^022~"Z;

——fStN2rn22222222

175

15 as; 15 5 2.8 9 P 6 aa 2.8

15 ß 15 y2.6 9 P 6 HN 3.1

15 P 15 HN 3.0 9 P 7 HN 3..3

15 «Si 15 y2.8 9 HN 11 P 3.2

15 6 15 HN 3.0 9 g* 12 8* 4.5

15 y 15 HN 3.1 10 ß 7 aRe 2.5

16 C-Re 16 HN 2.8 10 P 7 UN 2.9

16 »Si 16 P 2.6 10 ß 8 HN 3.0

16 HN 16 y 3.0 10 HN 12 P 3.0

16 P 16 y 2.7 11 ß 8 y 3.4

17 HN 17 «Si 3.0 11 HN 13 P 3.1

17 «Re 17 Y 2.7 11 HN 14 P 2.8

17 ß 17 y 2.5 11 y 14 y 5.0

17 p 17 HN 3.0 12 a* 15 P 2.5

17 Y 17 HN 2.9 12 P 9 MRg 2.3

18 Ctfe 18 P 2.6 12 HN 15 P 2.7

18 OtRe 18 HN 2.8 12 S* 15 5 4.5

18 P 18 HN 3.0 14 P 11 a* 2.5

19 «Re 19 HN 2.8 14 Y 17 Y3.5

19 o-rc 19 P 2.6 15 P 13 HN 3.3

19 HN 19 P 3.0 16 P 13 HN 3.0

19 HN 19 Y 3.3 16 Y19 ß 3.4

20 «Si 20 ß 2.9 16 HN 19 P 3.2

20 ß 20 5 2.9 17 HN 15 HN 3.5

20 P 20 Y2.6 17 HN 19 P 3.3

2 «Si 3 HN 2.7 18 ß 15 5 3.0

2 HN 1 «Re 2.5 19 P 16 aRe 2.6

Note: *= pscudoatom used for calculations

176

a-Heptapeptide wit a central Aib„^___

Table 41. NOEs observed in the 300 ms ROESY spectrum of peptide 30 in MeOH

Residue H-Atom Residue H-Atom dNOE [A] Residue H-Atom Residue H-Atom dNOE[A]

1 ß 1 a 2.5 2 a 3 Y4.1

1 a 1 Y3.0 2 a 3 HN 2.1

2 a 2 ß 2.5 3 a 2 P 4.0

2 a 2 HN 2.8 3 HN 2 P 3.3

2 ß 2 HN 3.1 3 HN 2 HN 4.2

3 P 3 a 2.3 3 a 4 HN 2.1

3 6 3 a 3.5 3 P 4 HN 3.0

3 Y3 a 2.7 4 P 5 HN 2.5

3 HN 3 a 2.5 4 HN 5 HN 2.9

3 IIN 3 Y 2.8 5 a 6 HN 2.5

3 P 3 HN 2.6 5 HN 6 HN 2.8

4 HN 4 P 3.1 5 HN 4 P 3.0

5 P 5 a 2.4 6 a 7 HN 2.6

5 ii* 5 a 3.6 6 ß 7 HN 3.6

5 y2* 5 HN 3.6 6 HN 7 HN 3.5

5 HN 5 a 2.8 1 Y 3 HN 5.0

5 HN 5 P 2.6 1 Y3 a 4.2

5 P 5 yl* 2.4 2 ß 5 P 3.7

5 HN 5 yl* 3.0 3 a 6 Y 3.6

6 ß 6 a 2.5 3 a 6 e 4.4

6 Y 6 a 2.9 3 P 5 HN 3.7

6 HN 6 a 2.6 3 P 6 Y 3.8

6 HN 6 Y 3.1 3 P 6 HN 4.0

6 HN 6 P 2.7 4 P 6 HN 3.9

7 a 7 P 3.7 4 P 7 P 4.2

7 HN 7 P 3.2 4 ß 7 HN 4.7

ß 2 a 3.6 5 P 2 HN 4.3

a 2 P 4.6 5 y2* 7 HN 4.9

a 2 HN 2.2 5 HN 2 P 4.3

P 2 HN 3.0 5 HN 3 a 3.4

Y* 2 HN 3.8 6 HN 2 P 4.8

y* 2 a 4.2 6 HN 3 a 3.2

a 2 a 3.9 6 HN 4 a 3.9

2 a 3 P 4.1 6 8 3 a 4.4

Note: * = pseudoatom used for calculations

177

Zn cordinating ß-peptides

Table 42. NOEs observed in the 300 ms ROESY spectrum ofpeptide 31 in H20

Residue H-Atom Residue H-Atom dNOE [A] Residue H-Atom Residue H-Atom dNOE[A]

1 a* 2 P 2.6 8 ß 8 HN 3.0

2 y 2 HN 3.1 8 a 8 HN 3.2

4 a 4 Pa 2.6 1 a* 2 HN 2.7

4 a. 4 Prs 2.9 1 a* 2 Y* 4.1

4 a 4 8 3.1 2 a 3 HN 2.6

4 a 4 Y 3.0 3 a 4 HN 2.4

4 a 4 HN 3.1 4 Y 5 HN 3.4

4 y 4 Pa 2.9 5 a* 6 HN 2.6

4 HN 4 Psi 3.0 5 P 6 HN 3.0

4 HN 4 PR* 2.8 7 ß 8 HN 3.2

4 HN 4 y 3.6 7 a 8 HN 2.4

5 a* 5 P 2.6 4 HN 6 P 3.4

5 P 5 E* 3.2 3 a 6 P 2.5

5 ß 5 HN 3.0 2 a 7 P 2.6

8 ß 8 Y 3.0

Note: * = pseudoatom used for calculations

Table 43. NOEs observed in the 300 ms ROESY spectrum ofpeptide 31 in H20 with ZnCl2

Residue H-Atom Residue H-Atom dNOE [A] Residue H-Atom Residue H-Atom dNOE[A]

1 a* 1 P 2.7 7 P 7 y* 2.8

2 P 2 HN 3.0 7 Y* 7 HN 3.0

3 P 3 HN 3.0 8 a* 8 HN 2.8

4 a 4 Psi 2.5 8 P 8 HN 3.0

4 a 4 pRe 2.7 8 P 8 HN 3.2

4 a 4 Ö* 2.9 1 a* 2 HN 2.7

4 a 4 Y 3.1 4 a 5 HN 2.2

4 a 4 HN 3.0 4 Y 5 HN 3.6

4 Y 4 Psi 2.7 4 HN 3 a 2.2

4 HN 4 Psi 2.8 5 P 6 Y 3.8

4 HN 4 ßsi 2.8 6 HN 5 ß 3.0

4 HN 4 Y 3.9 7 Y* 8 ß 4.0

4 a 5 ß 2.7 8 HN 7 Y* 3.5

5 P 5 Ô* 3.6 4 HN 6 ß 3.6

178

5 s* 5 HN 4.3

6 ß 6 HN 3.0

Note: * = pseudoatom used for calculations

2.7

7öWe 44. NOEs observed in the 300 ms ROESY spectrum of peptide 32 in MeOH

Residue H-Atom Residue H-Atom <1noe[A] Residue 11-Atom Residue H-Atom aW [A]

1 P 1 a* 2.6 8 P 8 HN 3.0

2 P 2 Y* 2.9 8 y* 8 HN 3.0

2 P 2 HN 2.9 1 a* 2 HN 2.5

2 r 2 HN 3.0 5 HN 4 a 2.2

3 P 3 Ï 2.6 5 a* 6 HN 3.2

3 P 3 HN 2.9 6 a* 7 HN 2.5

3 a* 3 HN 2.3 7 a* 8 HN 2.8

3 Y 3 HN 3.1 2 Y* 5 P 2.9

4 a 4 Y2.8 2 HN 5 P 3.0

4 P* 4 HN 2.7 5 UN 6 P 3.0

5 a* 5 HN 2.9 4 a 7 P 2.2

5 P 5 HN 2.9 4 P* 1 a* 2.9

5 P 5 y* 2.4 5 P 7 P 3.1

6 P 6 Y* 2.6 4 HN 7 P 2.9

7 P 7 a* 2.6 5 P 2 a* 2.4

7 P 7 Y* 2.8 5 5* 8 P 3.8

7 P 7 HN 3.0 5 HN 8 P 3.3

7 Y* 7 HN 3.4 6 P 3 a* 2.9

7 a* 7 HN 2.9 8 P 5 a* 3.2

8 p 8 Y 2.7

Note; * = pseudoatom used for calculations

179

Table 45. NOEsobserved in the 300 ms ROESY spectrum of peptide 32 in H20 + ZnCl2

Residue H-Atom Residue H-Atom dNoE[A]

1 a* 1 P 2.5

1 ß 1 y* 2.7

1 P 1 e* 4.6

2 P 2 Y* 2.6

2 P 2 a* 2.8

2 P 2 HN 3.0

3 P 3 a* 2.8

3 P 3 y2.6

3 a* 3 HN 2.7

3 P 3 HN 3.0

4 a 4 P* 2.6

4 a. 4 y2.8

4 a 4 HN 3.0

4 P* 4 y 2.7

4 P* 4 HN 3.0

4 P* 4 8* 3.5

4 y 4 HN 3.8

5 a* 5 HN 3.1

5 P 5 HN 3.0

5 y* 5 HN 2.6

5 P 5 a* 2.7

5 P 5 HN 3.0

6 P 6 y* 2.8

6 Ö 6 HN 3.2

6 P 6 a* 3.2

6 a* 6 HN 2.6

6 P 6 y* 2.5

6 P 6 HN 3.0

Note: *= pseudoatom used for calculations

Residue H-Atom Residue H-Atom dNoE[A]

6 y* 6 HN 3.2

7 P 7 HN 3.0

7 a* 7 HN 2.6

7 a* 7 P 3.1

7 Y* 7 HN 2.9

7 Y* 7 P 2.6

8 P 8 a* 2.7

8 a* 8 HN 2.6

8 P 8 HN 3.0

8 P 8 a* 2.5

8 y* 8 HN 2.8

1 a* 2 HN 2.6

2 y* 3 HN 4.6

2 p 3 HN 4.2

2 a* 3 HN 2.6

3 P 4 HN 4.2

3 a* 4 HN 2.6

4 a 5 HN 2.5

4 y 5 HN 3.4

5 HN 4 p* 4.4

5 p 6 HN 3.8

5 a* 6 HN 2.7

6 a* 7 HN 2.8

7 HN 6 P 4.2

3 HN 7 P 3.2

2 HN 8 P 3.9

2 HN 8 y* 3.8

1 y* 8 P 4.1

I

180

Table 46. NOEs observed in the 300 ms

Residue H-Atom Residue H-Atom dN0E [A]

1 a* 1 Y* 3.2

1 P 1 a* 2.3

1 P 1 Y* 2.3

2 f 2 Y* 2.4

2 a 2 HN 2.6

2 P 2 y* 2.2

2 P 2 HN 2.8

2 Y* 2 HN 2.7

3 a* 3 P 2.3

3 a* 3 Ö* 2.5

3 a* 3 Y* 2.9

3 ß 3 8* 2.2

3 P 3 HN 2.9

3 Ö* 3 HN 2.6

3 7 3 HN 2.8

4 a 4 ß 2.3

4 a 4 HN 2.8

4 Y 4 P 2.9

4 ß* 4 HN 3.0

4 ß* 4 Y 2.6

4 ß* 4 HN 2.6

4 Y 4 HN 3.6

5 Y* 5 a* 2.2

5 a* 5 HN 3.2

5 a* 5 P 2.6

5 a 5 HN 2.8

5 Y* 5 P 2.4

6 a* 6 P 2.2

6 a* 6 Y* 2.8

6 S 6 P 2.8

6 Y* 6 P 2.8

6 e* 6 P 2.3

7 Y* 7 P 2.2

7 HN 7 Y* 2.6

Note: *= pseudoatom used for calculations

ROESY spectrum of peptide 33 in MeOH

Residue H-Atom Residue H-Atom dN0E [A]

7 HN 7 P 2.7

7 P 7 Y* 2.3

8 P 8 Y* 2.3

8 HN 8 Y* 2.7

8 HN 8 P 2.7

8 a* 8 Y* 2.5

1 P 2 Y* 5.0

1 P 2 HN 3.4

2 Y* 3 P 4.7

2 Y* 3 HN 3.8

2 HN 3 8* 3.9

2 HN 3 P 5.5

3 P 4 ß* 4.8

3 P 4 HN 2.9

3 HN 4 HN 4.5

5 P 6 S 2.9

5 P 6 HN 3.0

6 5 7 HN 4.4

6 Y* 7 HN 4.6

5 Y* 6 HN 3.7

7 Y* 8 HN 3.9

1 Y* 3 P 4.3

2 Y* 4 a 3.7

2 Y* 4 P* 4.5

2 Y* 5 HN 5.0

1 P 3 a* 4.7

5 HN 3 P 3.2

1 P 3 8* 3.7

3 5* 7 Y* 3.4

3 Y 7 Y* 4.8

2 HN 4 a 5.0

4 a 7 P 3.0

4 a 7 Y* 5.5

181

Table 47. NOEs observed in the 300 ms ROESY spectrum ofpeptide 33 in H20-solution (KH2PO4

buffer + 10%DA pll 7.6) without Zn'

Residue H-Atom Residue Atom dNoE[A] Residue H-Atom Residue Atom dNOE[A]

1 a* 1 P 2.4 6 c* 6 Y* 2.5

1 a* 1 Y* 2.4 6 Y* 6 P 2.8

1 P 1 Y* 2.4 6 Y* 6 NH 2.8

2 a* 2 NH 3.1 6 Y* 6 8 4.3

2 P 2 Y* 2.3 6 NH 6 P 2.8

2 P 2 NH 3.7 7 a* 7 Y* 2.8

2 Y* 2 NH 3.4 7 P 7 a* 2.3

3 a* 3 P 2.4 7 P 7 Y* 2.3

3 a* 3 8* 2.5 7 Y* 7 NH 3.0

3 a* 3 Y 2.9 7 NH 7 P 3.0

3 a* 3 P 3.0 8 P 8 Y* 2.7

3 a* 3 8* 2.8 8 Y* 8 a* 3.4

3 a* 3 Y 3.2 8 Y* 8 NH 4.3

3 P 3 Y 2.4 8 Y* 8 NH 3.4

3 P 3 NH 2.8 1 a* 2 NH 3.7

3 8* 3 Y 2.1 2 P 1 a* 3.5

3 8* 3 NH 2.8 2 Y* 1 a* 4.0

3 Y 3 NH 3.0 2 Y* 3 P 4.6

4 a 4 Psi 2.4 2 Y* 3 NH 4.1

4 a 4 Plfc 3.0 3 P 4 a 4.8

4 a 4 8* 2.5 3 NH 2 P 3.4

4 a 4 Y 2.9 4 a 5 NH 2.1

4 a 4 NH 3.1 4 pRs 5 NH 4.2

4 Ps 4 pRe 1.7 4 s* 5 NH 5.6

4 Pa 4 5* 3.4 4 8* 5 P 3.2

4 Ps. 4 Y 3.1 4 Y 5 Nil 3.7

4 Ps 4 NH 3.1 a* 3 8* 3.0

4 Pr* 4 8* 3.3 a* 3 Y 4.1

4 Prs 4 Y 2.7 Y* 3 5* 3.8

4 PRe 4 NH 2.7 Y* 4 5* 4.7

4 8* 4 Y 2.4 Y* 5 Y* 3,3

4 8* 4 NH 4.9 2 Y* 4 Psi 4.4

4 8* 4 Pr* 3.4 2 Y* 4 Pr, 5.3

4 8* 4 Y 2.1 3 a* 1 a* 4.1

4 Y 4 NH 3.5 3 P 5 NH 4.0

182

5 a* 5 ß 2.3

5 a* 5 f 2.3

5 ß 5 5* 3.0

5 ß 5 r 2.3

5 ß 5 NH 2.6

5 NH 5 Y* 4.8

6 ß 6 Y* 2.6

6 8* 6 ß 2.9

6 5* 6 Y* 3.2

6 8* 6 NH 3.0

Note: * = pseudoatom used for calculations

Table 48. NOEs observed in the 300 ms ROESY spectrum ofpeptide 33 in H20-soln. (d"-Tris-buffer+ 10% DA pH 7.2), in the presence of 1 equiv. ZnCl2

Residue H-Atom Residue H-Alom dNoE[A] Residue H-Atom Residue H-Atom dNOE[Al

a* P 3.0 4 Y 4 NH 3.5

a* s* 3.8 6 P 6 Y* 2.4

a* Y* 2.7 6 Ö 6 NH 2.8

P Y* 3.2 6 6* 6 8 2.0

s* Y* 3.2 6 S* 6 Y* 2.3

2 a* 2 NH 2.8 6 Y* 6 NH 2.6

2 P 2 Y* 2.2 7 a* 7 P 2.4

2 ß 2 NH 2.9 7 a* 7 Y* 2.7

2 Y* 2 NH 2.9 7 ß 7 Y* 2.3

3 a* 3 P 2.3 7 P 7 NH 2.6

3 a* 3 Ô 2.4 7 Y* 7 NH 2.8

3 a* 3 Y 2.7 1 a 2 NH 2.4

3 a* 3 ß 2.6 2 P 3 NH 3,3

3 a* 3 8* 2.7 2 Y* 1 a* 4.0

3 a* 3 Y 3.2 2 Y* 1 E* 4.4

3 a* 3 Nil 2.4 2 Y* 3 P 4.3

3 P 3 8* 2.3 2 Y* 3 NH 4.2

3 ß 3 Y 2.3 2 NH 3 NH 5.0

3 P 3 NH 2.8 3 a* 4 NH 2.6

3 8* 3 Y 2.0 3 P 4 NH 3.2

3 8* 3 NH 2.8 3 8* 2 NH 3.7

3 Y 3 NH 3.1 4 a 5 NH 2.1

3 ß 6 a* 3.4

3 8* 1 P 3.4

3 5* 7 Y* 3.9

3 Y 7 Y* 4.7

4 ßfe 7 Y* 2.6

4 5* 1 Y* 3.4

4 8* 8 Y* 5.0

5 a* 3 P 3.3

7 ß 4 a 5.0

7 Y* 3 a* 4.9

183

4 a

4 a

4 a

4 a

4 a

4 ßs,

4 Pa

4 Ps

4 ßsi

4 Pu.

4 Pr,

4 5*

4 Ô*

Mntf; * = rise

4 ßs, 2.4 4 8*

4 Pifc 3.2 4 Y

4 5* 2.5 1 a'

4 Y 3.0 3 P

4 NH 2.9 3 5*

4 Pu. 1.8 3 8*

4 6* 2.4 3 8*

4 Y 3.0 3 Y

4 NH 2.9 3 Y

4 Y 2.6 3 Y

4 NH 2.8 3 Y

4 Pte 3.3 3 Y

4 Y 2.2 3 Y

5 NH 3.3

5 NH 3.5

3 8* 3.0

5 NH 3.9

1 e2 4.0

1 Y* 1.1

7 Y* 3.2

7 P 3.6

7 Y* 4.5

7 Y* 4.5

7 Y* 4.5

7 Y* 4.5

7 Y* 4.5

pseudoatom used for calculations

184

Cyclic ß-peptides (RGD)

Table 49. NOEs observed in the 300 ms ROESY spectrum ofpeptide 34 in MeOH

Residue H-Atom Residue H-Atom dM0E[A]

ß as. 2.7

Y* aSi 3.0

ß OCRs 3.2

Y* a-te 2.8

HN «Re 2.9

ß y* 2.6

ß HN 2.8

Y* HN 3.2

2 «Si 2 Cr* 1.7

2 as. 2 ß 2.5

2 P 2 «Re 3.0

2 «Si 2 Y* 2.7

2 ß 2 Y* 2.4

2 ß 2 HN 3.0

2 ß 2 HNx 4.1

2 8* 2 HN 2.8

2 Y* 2 HN 2.7

2 HN 2 HNx 5.4

3 HN 3 a 3.2

3 Psi 3 a 2.4

3 ßa 3 ßRe 1.9

3 ßsi 3 Yl 3.7

3 ßs, 3 Y2 2.5

3 ßs. 3 HN 2.8

3 ßn. 3 HN 2.5

3 vi 3 HN 4.0

4 ßs. 4 a 2.6

4 y* 4 a 2.5

4 HN 4 a 3.0

4 ßs. 4 ßR* 2.0

4 ßsi 4 Y* 3.0

Note: * = pseudoatom used for calculations

Residue H-Atom Residue H-Atom dNcœlA]

4 ßs, 4 HN 3.0

4 ßR* 4 Y* 3.4

4 ß 4 HN 2.8

4 Y* 4 HN 4.4

1 HN 2 aR= 2.8

1 ß 2 ß 4.9

1 ß 2 HN 3.4

1 ß 2 Y* 5.4

2 ctsi 3 HN 2.5

2 OlRe 3 Yl 4.5

2 OtRe 3 HN 3.0

2 ß 3 a 5.2

2 ß 3 Yl 5.5

2 ß 3 HN 2.5

2 HN 3 HN 3.9

3 a 4 HN 2.5

3 ßfe 4 HN 3.1

3 yl 4 HN 2.9

3 Y2 4 HN 3.5

3 HN 4 HN 3.7

«Re 3 a 4.5

«Re 4 HN 4.8

ß 4 ßRe 5.2

ß 4 HN 5.0

HN 3 a 4.2

HN 3 Yl 4.9

HN 4 ßsi 4.1

HN 4 Y* 4.2

2 ß 4 ßs, 5.2

2 ß 4 HN 3.5

4 HN 1 HN 3.4

185

Table 50. NOEs observed in the 300 ms ROESY spectrum of peptide 35 in MeOH

Residue H-Atom Residue H-Atom dNOE [A] Residue H-Atom Residue 11-Atom dNOE [A]

P «Si 2.6 4 ßs, 4 ßRe 1.9

a si T 2.7 4 Yi 4 HN 3.6

a Re Y 3.1 4 ßRs 4 HN 2.7

a ReHN 2.8 4 ßsi 4 HN 3.0

ß Y 2.5 2 HN 1 as. 2.5

Y HN 3.0 2 a& 3 HN 2.7

ß HN 3.0 1 ß 2 HN 3.3

2 as, 2 ß 2.6 2 ß 3 HN 2.5

2 «s, 2 y* 2.9 3 a 2 HN 4.7

2 as, 2 8* 3.6 3 HN 2 HN 5.2

2 ß 2 5* 2.9 3 a 2 ß 4.2

2 ß 2 8* 3.3 3 HN 2 aRe 3.2

2 «Si 2 Kr* 1.9 3 ßsi 2 aRe 4.5

2 ß 2 Y* 2.8 3 HN 4 a 2.5

2 un 2 5* 3.2 3 HN 4 HN 3.9

2 ß 2 HN 2.9 3 ßsi 4 HN 4.3

2 HNx 2 P 3.7 3 ßRS 4 HN 3.0

3 oc 3 ßs, 2.5 4 HN 3 Y* 3.9

3 HN 3 ßs, 3.0 HN 4 a 2.3

3 HN 3 ß*. 2.6 HN 4 Pr* 3.8

3 Y* 3 ßsi 3.1 aRe 4 HN 4.0

3 ß* 3 pR* 2.0 HN 4 HN 3.5

3 a 3 HN 3.0 HN 4 yl 3.0

4 a 4 ßsi 2.5 2 ß 4 HN 3.1

4 a 4 HN 3.0

pseudoatom used for calculations

186

Table 51. NOEs observed in the 300 ms ROESY spectrum of peptide 36 in MeOH

Residue H-Atom Residue H-Atom dNOE[Â] Residue H-Atom Residue H-Atom dNOE [Â]

ß 1 HN 3.0 1 ß 2 ß 4.5

ß 1 &Re 3.2 2 HN 1 ß 2.9

«si 1 aRe 1.9 2 ß 3 P 4.0

UN 1 Y* 3.6 2 P 3 HN 2.6

2 «S! 2 P 3.1 2 e* 3 ß 4.3

2 CtRs 2 ß 3.0 2 HN 1 a* 33

2 yl 2 ß 2.9 1 HN 2 HN 4.7

2 y2 2 ß 2.8 3 ß 4 HN 3.6

2 HN 2 ß 3.0 3 HN 2 F.* 5.0

2 HX 2 ß 3.6 3 HN 4 ß 4.7

3 ß 3 y2 2.6 3 HN 4 HN 3.8

3 P 3 yl 2.4 1 HN 4 a* 3.1

3 P 3 HN 2.6 1 HN 4 ß 3.6

3 yl 3 y2 1.9 2 ß 4 ß 3.9

4 HN 4 «Re 2.8 2 P 4 HN 4.8

4 ß 4 yl 2.5 2 e* 4 ß 3.6

4 P 4 y2 2.4 2 HN 4 HN 4.9

4 P 4 HN 2.9 2 HNx 4 ß 3.6

Note: *= pseudoatom used for calculations

Table 52. NOEs observed in the 300 ms ROESY spectrum of peptide 37 in MeOH

Residue H-Atom Residue H-Atom dNOE [A] Residue H-Atom Residue H-Atom dN0E[A]

ß ocsi 2.6 4 a 4 Y* 2.9

P «Re 3.1 4 a 4 HN 2.9

ß Y* 2.9 4 Prc 4 HN 3.0

P HN 2.9 4 P* 4 HN 2.9

HN Y* 3.1 4 HN 4 y* 4.7

«Ru eta 1.8 1 P 2 HN 3.4

2 ß 2 «Re 2.9 1 P 2 5* 4.8

2 ß 2 «Si 2.8 2 a& 3 HN 2.7

2 5* 2 ß 3.7 2 O-Re 3 HN 3.2

2 Y* 2 HN 2.9 2 P 3 HN 2.8

187

2 ß 2 HN 3.0

2 «Re 2 as 1.9

3 a 3 HNx 4.5

3 OCre 3 «Si 1.9

3 a 3 HN 3.4

3 a 3 Psi 2.6

3 1IN 3 Ï* 4.5

3 a 3 r 3.2

3 ßsi 3 UN 3.0

3 ßRe 3 HN 2.7

4 a 4 ßsi 2.7

Note: *= pseudoatom used for calculations

Cyclic ß-peptides (Somatostatin)

Table 53. NOEs observed in the 300 ms ROESY spectrum ofpeptide 38 in MeOH

Residue H-Atom Residue H-Atom dN0E[A] Residue H-Atom Residue H-Atom dN0E[A]

1 HN 1 Y* 3.5 4 HN 4 ßl 2.7

1 HN 1 ßi 3.2 4 HN 4 ß2 2.6

1 HN 1 ß2 3.0 2 HN 1 ßl 5.0

2 P 2 »Si 2.6 2 HN 1 ß2 4.7

2 ß 2 «Re 2.8 2 ß 1 s 4.9

2 ß 2 Y* 3.4 2 HN 1 HN 5.0

2 ß 2 HN 2.9 3 HN 2 «Si 3.4

2 HN 2 y* 3.3 3 HN 2 «Re 3.2

3 HN 3 ßl 3.1 3 HN 2 ß 3.7

3 HN 3 ß2 3.0 3 HN 2 HN 4.5

Note: * = pseudoatom used for calculations

a 4 HN 2.3

HN 3 HN 4.3

y2 4 HN 3.7

HN 4 a 2.4

ß 4 Y* 4.8

HN 4 Psi 3.8

HN 4 Ï* 4.0

HN 4 HN 3.9

ß 4 HN 3.7

ß 4 HN 4.5

188

Table 54. NOEs observed in the 300 ms ROESY spectrum of peptide 39 in MeOH

Residue H-Atom Residue H-Atom dNQE[A] Residue H-Atom Residue H-Atom dNoE[A]

ßl 1 ß2 2.0 4 ß* 4 HN 3.3

HN 1 ß* 3.2 HN 2 HN 4.1

HN 1 y* 3.9 ß* 2 HN 3.8

HN 1 ß* 3.2 e 2 ß 4.7

2 a* 2 ß 3.1 e 2 HN 4.0

2 a* 2 HN 3.5 Y* 2 HN 3.3

2 y* 2 ß 3.1 2 a* 3 HN 3.2

2 HN 2 ß 3.0 2 ß 3 ß* 4.7

2 HN 2 5* 3.6 2 ß 3 HN 3.4

2 HN 2 Y* 3.4 2 HN 3 HN 4.8

3 a* 3 HN 3.6 2 ß 4 HN 4.1

3 ß* 3 HN 3.1

Note: * = pseudoatom used for calculations

Table 55. NOEs observed in the 300 ms ROESY spectrum of peptide 41 in MeOH

Residue H-Atom Residue H-Atom dNOE [A] Residue H-Atom Residue H-Atom dN0E[A]

1 ß 1 «Si 2.6 3 y2 3 HN 3.5

1 ß 1 «Ri 3.0 4 ß 4 HN 2.8

1 ß 1 Y* 3.2 4 yl 4 HN 3.1

1 P 1 HN 3.0 4 y2 4 HN 3.2

2 ßl 2 ß2 1.9 3 a 2 HN 2.7

2 a 2 Y* 2.9 2 HN 3 UN 2.9

3 a 3 Yl 3.1 3 HN 4 HN 3.9

3 a 3 72 32 1 ß 4 HN 2.7

3 a 3 HN 2,9 2 y# 4 ß 5.3

3 ßl 3 ß2 2.0 2 Y* 4 HN 5.4

3 Yl 3 HN 3.5 2 HN 4 ß 4.0

Note: * = pseudoatom used for calculations

189

7.6 References

[1] W. C. Chan., P. D. White, 'Fmoc Solid Phase Peptide Synthesis: A Practical

Approach', Vol. 222, Oxford University Press, Oxford, 2000.

L2J W. S. Hancock, J. E. Battersby, Anal. Biochem. 1976, 71, 260.

[3] T. D. Goddard, D. G. Kneller, Version 3.110, San Francisco, 2004.

[4] C. D. Schwieters, J, J. Kuszewski, N. Tjandra, G. M. Clore,./. Magn. Reson.

2003,760,66.

[5] E. G. Stein, L. M. Rice, A. T. Brunger, J. Mag. Reson. 1997,124, 154.

190

CURRICULUM-VITAE

Name

Date of Birth

Raveendra I. Mathad

ah16in June 1976

Nationality Indian

Education

1994-1997

1997-1999

1999-2000

2000-2001

2001-2003

2003-2007

B.Sc.

K. L.E. College, Bangalore University, India

M. Sc. ChemistryCentral College, Bangalore University, India

Research Project on Peptide Synthesisat Bangalore University

NMR Project Assistant at SIF, IISc, Bangalore, India

Research Associate at General Electric R&D Centre

Bangalore, India

Ph. D. Thesis, ETH Zürich, Switzerland

"Synthesis of Mixed a/ß-Peptides and NMR-Solution

Structure of F/OH-Substituted and Other ß-Peptides"

Zürich, May 2007