Robust transition metal markers for labelling of … transition metal markers for labelling of peptides via solid phase synthesis methods Dissertation for the degree of Doktor der

Robust transition metal markers for labelling of

peptides via solid phase synthesis methods

Dissertation for the degree of

Doktor der Naturwissenschaften

in the Fakultät für Chemie

at the Ruhr-Universität Bochum

presented by

Dave Richard van Staveren

Bochum, June 2001

This work was carried out between July 1999 and May 2001 at the

Max-Planck-Institut für Strahlenchemie

Mülheim an der Ruhr, Germany

Submitted on: May 16th 2001

Examination: June 27th 2001

Referent: Prof. Dr. K. Wieghardt

Korreferent: Prof. Dr. W. S. Sheldrick

Prüfer: Prof. Dr. W. Sander

Acknowledgements

I would like to acknowledge everybody who showed interest in my work and supported me

during this Ph. D. period. I am especially indebted to:

Prof. Dr. Karl Wieghardt, for the opportunity to work in his group and the access to chemicals

and equipment. I really appreciate that I was allowed to complete my work after my advisor

joined another university.

Prof. Dr. Nils Metzler-Nolte, for the freedom he gave me during my Ph. D. research, his

constant encouragement, many helpful comments and the skills in NMR spectroscopy he

taught me.

Dr. Thomas Weyhermüller and Heike Schucht for the numerous high-quality X-ray crystal

structure determinations. I appreciate that I was given the opportunity to have a close look at

the process of X-ray data collection and the subsequent elucidation and refinement of the

structure.

Dr. Eberhard Bothe, Petra Höfer and Helmut Schmidt for their technical assistance with the

electrochemical and spectro-electrochemical measurements. I am grateful to Dr. E. Bothe for

his assistance with the low temperature electrochemical measurements and his helpful

discussions.

Dr. Michael Bühl, for performing the Density Functional Theory calculations and the pleasant

cooperation. The accuracy of the results impressed me to a large extent.

Dr. Eckard Bill, Frank Reikowski and Bernd Mienert for their help with the EPR and

Mössbauer spectroscopic data acquisition and interpretation.

Jörg Bitter and Kerstin Sand for running many NMR spectra, and their patience and

willingness to cooperate.

Manuela Trinoga for the numerous skillfully performed HPLC purifications and the nice

conversations.

Herr Selbach for his assistance with the solid phase peptide syntheses. I also would like to

thank Thomas Happ and Ulrich Hoffmanns for their preliminary investigations concerning the

solid phase peptide synthesis.

Andy Göbels for his assistance with the circular dichroism spectroscopic measurements.

Prof. Dr. Phalguni Chaudhuri for his interest in the molybdenum chemistry and his helpful

discussions.

I also would like to thank Dr. Craig Grapperhaus and Dr. Bas de Bruin for the nice

conversations and the many helpful suggestions and comments along the way. I also would

like to acknowledge Tapan Paine for his helpful comments concerning the organic syntheses.

Furthermore, I also would like to thank Udo Beckmann, Ricardo Garcia, Dr. Diran Herebian,

Dr. Shuji Kimura for the conversations and helpful comments.

Weiterhin möchte ich mich ganz herzlich bei Silke Klein bedanken. Obwohl Du keinen

wissenschaftlichen Beitrag leistetest, war Deine Unterstützung in jeglicher Hinsicht doch von

unschätzbarem Wert. Ohne Dich, liebe Silke, hätte diese Doktorarbeit auf keinen Fall diesen

Umfang und diese Qualität erreicht.

Eveneens wil mijn moeder bedanken. Zonder jouw steun al die jaren tijdens mijn schooltijd,

studie en mijn promotieonderzoek had dit proefschrift nooit tot stand kunnen komen. Echt

heel erg bedankt voor alles, en gelukkig kunnen we ons op een aanzienlijk stressvrijere

toekomst verheugen.

voor Oene

i

Table of contents

Chapter 1 Introduction 1

1.1 General introduction 1

1.2 Metal ions in biological systems 2

1.3 Bio-organometallic chemistry 5

1.4 Application of organometallic complexes in immuno-assays 9

1.5 Objectives and outline of this thesis 12

Chapter 2 Peptides 15

2.1 General aspects 15

2.2 Peptide synthesis in solution 16

2.3 Solid phase peptide synthesis 18

2.4 Biological properties of enkephalin 20

Chapter 3 The marker Mo(ηηηη-Cp-COOH)(ηηηη-allyl)(CO)2 23

3.1 Reasons for selecting molybdenum carbonyl complexes as markers 23

3.2 Properties of Mo(η-Cp)(η-allyl)(CO)2 and synthesis of the marker 24

3.3 Coupling of the marker with amino acids and peptides 27

3.3 X-ray crystal structure of the phenylalanine derivative 28

3.4 1H NMR spectroscopic investigations 31

3.5 Infrared spectroscopy 34

3.6 Electrochemistry and air-sensitivity 35

Table of contents

ii

Chapter 4 Fluxional processes in complexes of the type

Mo(His)(ηηηη-2-R-allyl)(CO)2

37

4.1 Introduction 37

4.2 Synthesis of the complexes 38

4.3 Solid state structures 39

4.4 Behaviour in solution 45

4.5 Results from Density Functional Theory calculations 52

4.6 Electrochemical investigations 57

4.7 Electronic and infrared spectroscopic investigations 63

4.8 EPR spectroscopic investigations 71

4.9 Density Functional Theory calculations on the oxidised complexes 75

4.10 Concluding remarks 78

Chapter 5 Markers based on the complex Mo(His)(ηηηη-allyl)(CO)2 80


5.2 Synthesis of the enantiomeric markers and diastereomeric

bioconjugates

81

5.3 X-ray crystallography 83

5.4 NMR spectra of the phenylalanine and dipeptide derivatives 86

5.5 Solid phase synthesis 87

5.6 NMR spectra of the [Leu]-enkephalin bioconjugates 89

5.7 Circular dichroism spectroscopy 93

5.8 Electrochemistry and infrared spectroscopy 95


Chapter 6 Spectroscopic properties and reactivity of the complexes

Mo(bpa)(CO)3 and Mo(benzyl-bpa)(CO)3

98


6.2 Synthesis 99


6.4 NMR and electronic spectroscopy 105

6.5 Electrochemistry 109

Table of contents

iii

6.6 Spectro-electrochemistry 111

6.7 EPR spectroscopy 115


Chapter 7 The Mo(bpa)(CO)3 unit as a marker 117


7.2 Synthesis in solution 118

7.3 Solid phase synthesis 123

7.4 Infrared spectroscopic and electrochemical investigations 124

7.5 Concluding remarks on the labelling of [Leu]-enkephalin and future

outlook

125

Chapter 8 Ferrocene and cobaltocenium conjugates of amino acids and

dipeptides, a hydrogen bonding study

127

8.1 Introduction 127

8.2 Synthesis 129


8.4 Investigation of the conformation in solution 139

8.5 Mössbauer spectroscopy and electrochemistry 143


Chapter 9 Summary 148

Experimental Section 155

Crystallographic data 192

References 200

Curriculum Vitae 212

iv

Abbreviations

Å angström

Ar aryl group

B magnetic field

br broad

CD circular dichroism

cm centimeter

CT charge transfer

CV cyclic voltammetry

d doublet

dd double doublet

δ chemical shift, isomer shift

∆EQ quadrupole splitting

E potential

Ep,a anodic potential

Ep,c cathodic potential

EI electron impact

EPR electron paramagnetic resonance

esd estimated standard deviation

ESI electro-spray interface

Et ethyl

ε molar extinction coefficient

FAB fast atom bombardment

G gauss

g g-value

hr hour

HOMO highest occupied molecular orbital

Abbreviations

v

HPLC high performance liquid chromatography

Hz hertz

I nuclear spin

IR infrared

J coupling constant

K kelvin

LUMO lowest unoccupied molecular orbital

λ wavelength

m multiplet, meter, milli-, medium (intensity)

M molar, mega-

Me methyl

min minute

MO molecular orbital

m/z mass per charge product

NMR nuclear magnetic resonance

ν stretching vibration

OTTLE-cell optically transparent thin layer electrochemical cell

ppb parts per billion

ppm parts per million

q quartet

RT room temperature

S spin

s singlet, second, strong (intensity)

SOMO single occupied molecular orbital

σ standard deviation

T tesla, temperature

t triplet

tert tertiary

UV ultra violet

Vis Visible

vs. versus

w weak (intensity)

Abbreviations

vi

Abbreviations for chemicals and solvents

Ala alanine

b-bpa N-benzyl-N,N-di(2-picolyl)amine

benzyl-bpa N-benzyl-N,N-di(2-picolyl)amine

bpa di(2-picolyl)amine

Boc tert-butyloxycarbonyl

2-ClTrt 2-chloro-trityl

Cys cysteine

DCM dichloromethane

dipea di-isopropyl-ethylamine

DMF dimethylformamide

DMSO dimethylsulfoxide

EtOAc ethylacetate

EtOH ethanol

Fmoc Fluorenyl-9-methoxycarbonyl

Gly Glycine

HBTU O-(benzotriazole-1-yl)-N,N,N’,N’-tetramethylurionium hexafluorophosphate

His histidine

HOBt 1-hydroxy-1H-benzotriazole

Leu leucine

Lys lysine

MeCN acetonitrile

MeOH methanol

MSNT 1-(mesitylene-2-sulphonyl)-3-nitro-1H-1,2,4-triazole

PEA S-1-phenyl-ethylamine

Phe phenylalanine

tacn 1,4,7-triazacyclononane

TBTU O-(benzotriazole-1-yl)-N,N,N’,N’-tetramethylurionium tetrafluoroborate

TFA trifluoro-acetic acid

THF tetrahydrofuran

TIS tri-isopropylsilane

Tyr tyrosine

1

1Introduction

1.1 General introduction

During the second half of the 20th century the well-defined borders between the classical

disciplines of chemistry, which are organic chemistry, inorganic chemistry and biochemistry,

have started to disappear. Each of these fields has profited to a large extent from achievements

made in the other areas of chemistry. Nowadays a fairly good share of reactions in organic

chemistry employ (transition) metals, either as elements or in the form of compounds [1].

Also the use of biochemical reagents in the form of immobilised enzymes has become of great

importance for organic chemistry, in particular for stereospecific reactions [2]. The

biochemists on their turn have benefited immensely from discoveries made by their organic

colleagues. For example, Merrifield’s invention of solid phase peptide synthesis [3] provided

(bio)chemists a method by which small peptides (up to 40 amino acids) with essentially any

desired primary structure can be prepared [4-6]. In the last few decades the areas of

biochemistry and inorganic chemistry have become much more related as well, because it has

been discovered that a variety of metal ions is indispensable for living cells.

Chapter 1

2

1.2 Metal ions in biological systems

Until about 60 years ago, only some (earth-)alkaline metal ions, such as Na+, K+, Ca2+ and

Mg2+, were considered essential for life. The earth-alkaline metal ion Ca2+ plays a number of

important roles in organisms [7-9]. For example bones consist of an organic phase, mainly

collagen, and an inorganic phase, of which hydroxyapatite with chemical formula

Ca5(PO4)3OH is the main constituent. Furthermore, Ca2+ ions act as messengers in eukaryotic

cells in a manner similar to the action of cAMP. Transient increases of the Ca2+ concentration

in the cytosol trigger various cellular responses, like muscle contraction, release of

neurotransmitters and the breakdown of glycogen to glucose-6-phosphate. Moreover, Ca2+

serves as a regulator of the citric acid cycle, by activating the enzymes pyruvate

dehydrogenase, isocitrate dehydrogenase and α-ketoglutarate dehydrogenase.

The alkaline metal ions Na+ and K+ are also important for cells [7-9]. The plasma-

membrane enzyme [Na+-K+]-ATPase pumps two K+ ions out of and three Na+ ions into the

cell with the concomitant hydrolysis of intracellular ATP. This enzyme therefore regulates the

concentration of electrolytes in the cell, which enables animal cells, all of which lack a cell

wall, to control their water content osmotically. Furthermore, the electrochemical gradient that

is generated by the enzyme is important for the signal transmission by nerve cells.

The earth-alkaline metal ion Mg2+, the fourth most abundant cation in the human body

after Na+, K+ and Ca2+, is indispensable for life as well [7-9]. For example, stabilisation of the

helical structure of DNA is accomplished by Mg2+ ions in cooperation with Na+ ions, by

neutralising the negatively charged phosphate backbone. In fact, enzymes that mediate

reactions with nucleic acids or nucleotides, e.g. ATP, usually require Mg2+ for activity [7-9].

At the present time, it is more or less general knowledge that apart from these four

(earth-)alkaline metal ions, also numerous transition metal ions are necessary for living

organisms [9-11]. The roles of the transition metal ions are diverse and include involvement

in catalytic reactions, transporting tasks as well as electron transfer processes. Moreover, the

metal ion can also have a structural role, for example in the case of a zinc-finger in proteins.

This structural motif consists of a Zn2+ ion, which is in an [S4], [NS3] or [N2S2] coordination

environment, contributed by cysteinato sulfur atoms and histidine Nε atoms. These zinc

Introduction

3

fingers constitute important secondary structural elements that are found predominantly in

DNA-binding proteins [12, 13]. As a result of this zinc-finger, the protein possesses a loop-

like structure, which enables it to bind efficiently and selectively to the DNA’s major groove

at specific nucleobase sequences. In this way, the protein regulates vital processes, such as the

expression or inhibition of genes.

When the active site of an enzyme contains a metal ion that is coordinated by functional

groups from the protein, it is called a metallo-enzyme. These kinds of enzymes utilise the

versatile chemistry of the transition metal ions. A metallo-enzyme with a transporting function

is for example hemoglobin, which delivers dioxygen from the lungs to the tissues. In its deoxy

form, the active site of hemoglobin consists of a low-spin Fe2+ ion that is coordinated by a

planar tetradentate heme ligand and by a histidine Nε atom, with the sixth coordination site

available for dioxygen. The binding of dioxygen is accompanied by the transfer of an electron

from the Fe-ion to the dioxygen molecule, resulting in a high-spin Fe3+ ion and a superoxide

ligand. Upon release of dioxygen the reverse process takes place, yielding again deoxy

hemoglobin.

Catalytic reactions of metallo-enzymes can be divided in two classes, i.e. reactions that

occur without a change of the metal ion’s valency and those that involve a change of the

oxidation state of the metal ion. An example of the former is the Zn2+ containing enzyme

carbonic anhydrase, which catalyses the reaction of H2O and CO2 to HCO3- and a proton. The

Zn2+ ion in human carbonic anhydrase I and II is coordinated by two histidine Nε atoms, one

histidine Nδ atom and a water molecule or hydroxide ion, as revealed by protein X-ray

crystallography [14, 15]. The role of the metal ion is to serve as a Lewis acid and, thus, to

decrease the pKa of the coordinated water molecule, in this way activating a nucleophilic

attack of the Zn-OH moiety on the CO2 carbon atom.

A catalytic reaction involving a change of the metal ion’s oxidation state takes for example

place in Cu-Zn Superoxide Dismutase (Cu-Zn SOD). The active site of this enzyme consists

of a Zn2+ and a Cu2+ ion, which are bridged by a deprotonated imidazole from a histidine, with

the Nδ coordinated to Zn2+ and the Nε to the Cu2+ ion. Furthermore, the Cu2+ ion is ligated by

two additional histidine- Nε atoms and one histidine Nδ atom, whereas the Zn2+ is further

coordinated by two histidine Nδ atoms and a carboxylate oxygen atom from an aspartate. This

Chapter 1

4

enzyme catalyses the reaction of two superoxide-ions with two protons to yield one molecule

of dioxygen and one molecule of hydrogen peroxide. The produced hydrogen peroxide, which

is also a harmful substance for living cells, is subsequently transformed by the enzyme

catalase. It has been revealed that the catalytic mechanism of Cu-Zn SOD consists of two

steps: the so-called ping-pong mechanism. The Cu(II) state of the enzyme undergoes a redox-

reaction with an O2- molecule, yielding a Cu(I) state and a molecule of dioxygen. In the

second step, this Cu(I) state reacts with a second O2- molecule and two protons, resulting in

formation of a H2O2 molecule and regeneration of the Cu(II) state.

The metallo-enzymes mentioned thus far have transition metal ions that are merely

coordinated by classical coordination chemical donor atoms, such as N, O and S atoms.

However, also metallo-enzymes with organometallic ligands are known to be present in

nature, one example thereof being the [Fe-Ni]-hydrogenase, which contains a Fe(CN)2(CO)

fragment. The active site of the reduced (active) and the two-electron oxidised (inactive)

forms of this enzyme, isolated from the bacterium D. Gigas, is depicted in Scheme 1.1, as

revealed by protein X-ray crystallography [16, 17]. In the reduced form, the nickel ion is

coordinated by two terminal cysteinato sulfur atoms and by two cysteinato sulfur atoms that

bridge to the Fe(CN)2(CO) moiety. In the two-electron oxidised form, an extra bridge in the

form of an hydroxo or oxo ligand is present.

Scheme 1.1 Representation of the active site of [Ni-Fe]-hydrogenase from D.Gigas as

revealed by protein X-ray crystallography [16, 17]: left: active (reduced) form;

right: inactive (oxidised) form.

NiS Fe

SS

S

Cys

Cys

CysCys

COCN

CNNi

S Fe

SS

S

Cys

Cys

CysCys

COCN

CNO(H)

Introduction

5

Apart from the active reduced form (the Ni-B form) and the inactive two-electron oxidised

form (the Ni-A form), also two one-electron oxidised forms are known to exist, as concluded

from EPR and infrared spectroscopic investigations [18-24].

This enzyme is capable to transform dihydrogen efficiently in two protons and two

electrons. In the case of D. Gigas these electrons are used to fulfil the function of the enzyme,

namely the reduction of sulfate. Recently a review on the bio-organometallic chemistry of the

metal containing hydrogenases ([Fe only], [NiFe] and [NiFeSe]) appeared in the literature

[25].

1.3 Bio-organometallic chemistry

The era of organometallic chemistry started with the discovery of ferrocene by Kealy and

Pauson in 1951 [26]. Since then, research in this area has developed rapidly, which has lead to

characterisation of a large variety of structurally exotic and novel compounds. In the early

years of organometallic chemistry, there was a common belief among scientists that these

compounds had no biological relevance. It was therefore a huge surprise that metal carbon

bonds were actually found to be present in nature as well, the first example being the X-ray

crystal structure of vitamin B12 [27], which possesses a Co-CH3 bond.

On the border between organometallic chemistry and biochemistry a new field emerged in the

past few decades, which is called bio-organometallic chemistry. To most chemists, the name

bio-organometallic chemistry still implies a contradiction, because organometallic complexes

are thought of to be very reactive compounds, only being stable in the absence of water and

dioxygen. However, this generalisation is not justified, since many organometallic complexes

display high stability in aerobic aqueous media, e.g. ferrocene or the cobaltocenium cation.

The field of bio-organometallic chemistry can be divided in five classes of

compounds, based upon structural features and properties. The first class consists of

organometallic compounds that are encountered in nature, like vitamin B12 and the enzyme

[Fe-Ni]-hydrogenase. To the second class belong complexes that contain in addition to a

classical organometallic ligand, such as a Cp-ring or a carbonyl ligand, also a biologically

relevant molecule, like an amino acid, peptide or nucleic acid, coordinated to the metal ion.

Chapter 1

6

The third class is comprised of compounds that resemble structural features of biomolecules,

e.g. amino acids, but at the same time also possessing an organometallic fragment. Among the

fourth class of compounds, organometallic complexes are classified that do not contain a

single biomolecule, but exhibit biological activity, for example against tumors. The fifth class

of compounds constitutes organometallic complexes that are bound in a covalent way to

biomolecules.

Because the structure of the active site of the [Fe-Ni]-hydrogenase, which is a class one

compound, has already been depicted and discussed above, only examples of each of the other

four remaining classes will be presented at this stage. The synthetic bio-organometallic

chemistry started more or less with the use of amino acids as ligands. Each of the naturally

occurring amino acids contains two potential donor atoms, namely the carboxylate oxygen

atom and the nitrogen atom of the primary amine or in the case of proline the secondary

amine. Several of the amino acids contain, in addition to these two potential ligating atoms,

side-chains with another potential donor atom. The nature of these additional donor atoms

varies largely and constitutes carboxylates (Asp, Glu), amino groups (Lys), aromatic N-donors

(His), thiolate and thioether sulfur atoms (Cys and Met, respectively) as well as aliphatic and

aromatic alkoxides (Ser and Tyr, respectively). In the past two decades a large number of class

two compounds has been reported by the groups of Sheldrick and Beck, and recently an

overview of these appeared in the literature [28].

Two class two compounds showing different ways via which amino acids can bind to metal

ions are depicted in Scheme 1.2 [29]. The Ru2+ ion in the complex on the left of Scheme 1.2 is

coordinated by a η5-pentamethyl cyclopentadienyl ligand and by the carboxylate oxygen atom,

amine nitrogen atom and thioether sulfur atom of a deprotonated methionine. The Ru2+ ion in

the sandwich compound displayed on the right of Scheme 1.2 on the other hand is ligated in a

η5-fashion by a pentamethyl cyclopentadienyl ligand and in a η6-manner by the phenyl ring of

a phenylalanine.

Introduction

7

Scheme 1.2 Molecular structure of two Ru(II) complexes illustrating different binding

possibilities of amino acids (from ref. [29])

RuS

H2NO

O

RuCO2H

H2N

+

In Scheme 1.3 another example of a class two compound is shown, with a modified

nucleobase serving as a didentate ligand. In this complex, the Mo4+ ion is coordinated by two

η5-cyclopentadienyl ligands and the N-3 and exocyclic N-atom of a N-1 methylated thymine

[30]. In addition to the presence of a methyl group on N-1, the coordinated thymine differs

from the free nucleobase in another way, i.e. the exocyclic NH2 group is now deprotonated,

making it an amido ligand.

Scheme 1.3 Molecular structure of a Mo(IV) complex with a modified nucleobase as a

chelating ligand (from ref. [30])

Mo

NN

HN

O

+

The class three compounds consist of substances that structurally resemble biomolecules, for

example amino acids, while at the same time containing an organometallic fragment. To this

class belong the unnatural amino acids ferrocenylalanine and cymantrenylalanine, the

molecular structures of which are depicted in Scheme 1.4. The difference between these two

complexes and the Ru-phenylalanine sandwich compound displayed in Scheme 1.2 is evident.

The ruthenium compound constitutes an unnatural amino acid as well, but it contains the

amino acid phenylalanine. Ferrocenylalanine and cymantrenylalanine on the other hand cannot

be dissected into a naturally occurring amino acid and a metal fragment.

Chapter 1

8

Both ferrocenylalanine and cymantrenylalanine are quite stable compounds, and can be

introduced into a peptide, even by solid phase peptide synthesis methods [31-34]. However,

these unnatural amino acids have a drawback, because their synthesis yields a racemic mixture

of the D and L forms that can be separated into both enantiomers, but this is a very time-

consuming process [35-37]. Cymantrenylalanine can be modified before or after the peptide

synthesis by photochemical substitution of a carbonyl ligand by a trisubstituted phosphane

ligand [38, 39]. Because tri-alkyl and tri-aryl phosphanes with various substituents can be

introduced, the bulkiness of this unnatural amino acid and the electron-availability of the Mn+

ion can be fine-tuned this way.

Scheme 1.4 Molecular structures of cymantrenylalanine (left) and ferrocenylalanine (right)

MnCO

COOCNH2

CO2H

Fe NH2

CO2H

To the class four compounds belong for example several metallocenes, like titanocene

dichloride, vanadocene dichloride and a cationic molybdenocene dichloride derivative, whose

molecular structures are depicted in Scheme 1.5. Although these compounds contain no

biomolecule at all, they are classified under the field of bio-organometallic chemistry because

of their anti-tumor activity against a variety of tumors, as discovered by Köpf and Köpf-Maier

[40-42]. Of these three anti-tumor agents, titanocene dichloride displays the highest activity,

albeit less than several bis-β-diketonato titanium compounds, some of which already entered

clinical trials [43, 44].

Scheme 1.5 Molecular structures of three anti-tumor active organometallic complexes

TiClCl

VClCl

MoClCl

2+

Introduction

9

The fifth class consists of organometallic complexes that are bound to a biomolecule or

biologically relevant molecule in a covalent way and these compounds outnumber the

combined number of class 1-4 complexes. In the last decades several strategies have been

developed to tether organometallic complexes to amino acids or peptides and two examples

are shown in Scheme 1.6. On the left the complex Co(Cp)(CO)2 is covalently bound to β-

alanine via an amide bond [45]. The formation of an amide linkage from an organometallic

acid, in this case Co(Cp-COOH)(CO)2, and an amine, in this case β-alanine, constitutes a very

attractive method because standard peptide synthesis methods (see Chapter 2) can be applied.

On the left of Scheme 1.6, the complex Cr(η-benzene)(CO)3 is covalently bound to a Boc-

protected leucine derivative via an alkyne linkage [46]. This alkyne bond is formed from

Cr(η-C6ClH5)(CO)3 and the alkyne modified Boc-leucine via a Pd-catalysed Sonagoshira

reaction, employing CuI and Pd(PPh3)2Cl2 under exclusion of dioxygen.

Scheme 1.6 Two examples of class five bio-organometallic compounds (from ref. [45, 46])

Co

COOC

HN

O

COOHCr

COCO

OC

NH

OHN O

O

1.4 Application of organometallic complexes in immuno-assays

Quantitative analytical methods with antigens and antibodies are nowadays part of many

clinical, pharmaceutical and basic scientific investigations [47-51]. In these so-called

immuno-assays, either the antigen (also called hapten) or antibody is labelled with a substance

that displays spectroscopic properties that are characteristic for this compound. In this way the

concentration of the analyte, for example peptides or hormones, can be determined accurately

and quickly, also at very low concentrations. Even when the analyte could be determined

feasibly by other methods, like chromatography, immuno-assays are often preferred due to

their speed and simplicity.

Chapter 1

10

In the past, labelling of the antigens or antibodies was mainly performed by introduction of

radioactive isotopes, e.g. 3H and 125I [52-56]. However, these radiological immuno-assays

have a number of disadvantages, such as the consequent need for protection against radiation

and the generation of radioactive waste. Another drawback is the short lifetime of some

labelling reagents, which requires that the markers need to be synthesised constantly. These

shortcomings of radioactive labels stimulated the investigation and development of non-

radioactive markers, or “cold” markers. In the late 1970’s, Cais and co-workers reported the

first examples of immuno-assays with haptens bound to organometallic complexes [57, 58],

with detection and quantification of the analyte taking place by Atomic Absorbtion

Spectroscopy (AAS). A few years later it was demonstrated that this method is as accurate as

radio-immunological assays for several clinical determinations [59].

Apart from AAS, also other quantitative analytic methods applicable to metal complexes have

been used for detection, which include luminescence and electrochemistry. For the former

detection method, several complexes with the lanthanide ions Sm3+, Eu3+and Tb3+ bound to

the haptens have been used [60-65]. The quantification of the analyte occurred in these cases

by time-resolved luminescence measurements.

For electrochemical detection, ferrocene labelled compounds have been widely

studied. Morphine covalently attached to the ferrocene moiety was the first biologically

relevant molecule subjected to electrochemical investigations [66, 67]. It was demonstrated

that binding of the conjugate to the corresponding receptor resulted in a diminished

accessibility of the Fc/Fc+ transition due to the fact that the protein matrix of the receptor

surrounds the ferrocene. Unfortunately, this effect alone did not lend itself to quantification of

the analyte at concentrations that are easily and accurately determined by AAS [68]. However,

this problem can be overcome if unlabelled hapten is added, together with the use of oxidases

as signal amplifiers. In that case the unlabelled hapten will compete with the labelled hapten

for binding to receptor sites because its affinity is at least as high as that of the labelled hapten.

Subsequently, the liberated labelled hapten will serve as a substrate for the oxidases, and

therefore the measured current depends on the amount of liberated labelled hapten. This

somewhat cumbersome method has been applied successfully for the quantification of

thyroxine [69, 70], lidocaine [71] and human choriongonadotropine [72, 73], which

demonstrates its potential usability.

Introduction

11

After these first reports of quantification by electrochemical methods, amino acids and

proteins have been labelled with the ferrocene moiety, with detection taking place by HPLC-

ECD (High Performance Liquid Chromatography-ElectroChemical Detection) [74-76]. The

limit of detection was in these cases in the picomolar range.

The ferrocene derivative shown in Scheme 1.7 has been used for determination of

alkaline phosphatase [77]. Hydrolysis of the phosphate moiety lowers the redox potential by

210 mV, a difference that allows quick and reliable determination of the enzyme by

electrochemical methods.

Scheme 1.7 Application of a ferrocene derivative for the determination of alkaline

phosphatase (from ref. [77])

HN

O

O POH

OHO

alkalinephosphatase

HN

O

+

E1/2 = +390 mV vs Fc/Fc+

OH

H2O

+

E1/2 = +180 mV vs Fc/Fc+

H3PO4

Fe

Fe

In the last 15 years, Jaouen and co-workers developed a new immuno-assay method based

upon metal carbonyl complexes as markers, which is called Carbonyl Metal Immuno-Assay

(CMIA) [78-103]. After the pioneering work by Cotton and Kraihanzel in the early sixties

[104], it has been well known that metal bound carbonyls display intense and characteristic

bands in the 1800-2200 cm-1 region of the infrared spectrum. Biomolecules show almost no

absorption in this range, in this way not obscuring the detection of the CO stretching

vibrations of the metal-bound carbonyls. Jaouen and co-workers have shown that this method

is applicable to the quantification of various hormones, drugs and proteins, and recently they

published examples of simultaneous determination and quantification of several steroids [105,

106].

Chapter 1

12

One example from the extensive work of Jaouen is shown in Scheme 1.8: a β-estradiol

derivative tethered to a Re(Cp)(CO)3 moiety [99]. The identity of the molecule was

established by spectroscopy as well as by X-ray crystallographic methods. Interestingly, this

compound was shown to exhibit an affinity similar to β-estradiol itself for the corresponding

receptor. In the form depicted in Scheme 1.8, this molecule can be used in CMIAs, but the

corresponding iso-electronic and iso-structural Tc-analogue can also be used as a selective

radiopharmaceutical, as demonstrated by Wenzel and Klinge [107].

Scheme 1.8 Re(Cp)(CO)3 bound covalently to a β-estradiol derivative via an alkyne bond

(from ref. [99])

ClH2C

HO

OH

C C

ReCO

COOC

1.5 Objectives and outline of this thesis

As can be concluded from this introductory chapter thus far, labelling of biomolecules with

specific organometallic fragments permits easy and reliable detection of the bioconjugates by

spectroscopic methods. Although a large number of labelling reagents are known, many

compounds have a drawback because they are still quite reactive towards dioxygen, resulting

in rapid decomposition of the marker in air. Furthermore, the markers reported thus far are

either based on electrochemical detection, such as ferrocene, or on detection by infrared

spectroscopy. In addition, none of the markers reported thus far has been shown to be

compatible with solid phase peptide synthesis methods. Especially the development of

methods to introduce markers onto a peptide via solid phase peptide synthesis methods will be

the key-aim of this Thesis.

Introduction

13

These points mentioned above show that the design and development of new markers is still

of practical interest. The first objective of my thesis is the synthesis and characterisation of

functionalised metal carbonyl complexes that are quite resistant to dioxygen and that allow

detection by infrared spectroscopy as well as by electrochemical methods. First, these

functionalised markers will be coupled to amino acids and dipeptides and subsequently the

properties of these derivatives will be investigated. These conjugates are still low molecular

weight compounds, which makes their purification and characterisation not tedious. If these

amino acid and dipeptide derivatives exhibit low reactivity towards dioxygen, the introduction

of the marker onto larger peptides by solid phase peptide synthesis methods will be

investigated. The development of methods to introduce these markers onto peptides by solid

phase peptide synthesis methods will be the main objective of this Thesis. Jaouen and co-

workers have shown the feasibility of the CMIA method by labelling steroids and proteins

isolated from living organisms, such as bovine serum albumin. If spectroscopic labels can be

introduced onto a peptide by solid phase peptide synthesis methods, this will open new

interesting (bio)chemical and pharmaceutical research possibilities, because various small

peptides have a neurological action in organisms.

Furthermore, it will be interesting to investigate the influence of the biomolecule on the

spectroscopic properties of the metal complex. In addition, also the effect of the marker on the

properties of the biomolecule will be investigated, for example on the hydrogen bonding

properties.

These goals mentioned above are the main objectives of this dissertation, but another project

is the investigation of ferrocene and cobaltocenium complexes bearing amino acid and

dipeptide substituents as structural mimics for specific secondary structural elements of

proteins. Although various ferrocene derivatives of this kind have been reported recently (see

Chapter 8), it will be particularly interesting to investigate the influence of the positive charge

of the cobaltocenium complexes in relation to their neutral ferrocene analogues.

This thesis is organised in the following way: in Chapter 2 a short overview of coupling

reactions in peptide chemistry is given. Also the preparation of the peptides, both by solution

and solid phase synthesis methods, used in the following chapters of this Thesis is discussed.

In Chapter 3, the suitability of the complex Mo(η-Cp-COOH)(η-allyl)(CO)2 to serve as a

labelling reagent is investigated.

Chapter 1

14

In Chapter 4, the influence of replacement of the Cp-ring of the complex Mo(η-Cp)(η-

allyl)(CO)2 by its iso-electronic analogue L-histidinate is investigated. This substitution was

found to result in enhanced air-stability and it was investigated which type of allyl ligand

derivative (η-allyl or η-2-Me-allyl) yields a complex with more favourable marker properties.

During the course of these studies it was found that the complexes Mo(L-His-NεR)(η-

allyl)(CO)2 and Mo(L-His-NεR)(η-2-Me-allyl)(CO)2 (His = N, Nδ, O-histidinate; R = H or

C2H4C(O)OCH3) are fluxional in solution, both in their neutral and oxidised forms. The origin

of the fluxionality is elucidated by combining results from spectroscopic investigations and

from Density Functional Theory calculations.

In Chapter 5, the enantiomeric complexes Mo(L-His-NεR)(η-allyl)(CO)2 and Mo(D-His-

NεR)(η-allyl)(CO)2 (R = C2H4COOH) are coupled to biomolecules with varying complexity.

In addition to the common spectroscopic techniques, such as infrared and NMR spectroscopy,

these bioconjugates lend themselves also for investigation by circular dichroism spectroscopy

because of their chirality.

In Chapter 6 the electrochemical and spectroscopic properties of the complexes

Mo(bpa)(CO)3 and Mo(benzyl-bpa)(CO)3 (bpa = di(2-picolyl)amine) are investigated and

compared to those of their 1,4,7-triazacyclononane and hydrido-tris-pyrazolylborate

analogues. In Chapter 7, a convenient synthetic method for the introduction of the Mo(benzyl-

bpa)(CO)3 moiety on biomolecules of varying complexity is presented.

In Chapter 8, the spectroscopic and hydrogen bonding properties of several ferrocene

compounds bearing amino acid and dipeptide substituents are investigated, both in the solid

state and in solution. Furthermore, analogous cobaltocenium derivatives are prepared, and the

effect of the overall positive charge of these compared to the ferrocene analogues is

investigated.

15

2Peptides

2.1 General aspects

Peptides are comprised of amino acids that are linked together via amide (peptide) bonds. In

organisms, the synthesis of proteins is an efficient process, mainly taking place on the

ribosomes. However, the synthesis of many polypeptide antibiotics, such as Gramicidin S, is

mediated by enzymes rather than ribosomes, as discovered by Lipmann [108]. The ribosomal

as well as the non-ribosomal polypeptide synthesis proceeds from the N-terminus to the C-

terminus, which therefore appears to be general feature of biological peptide synthesis.

In the laboratory, the synthesis of peptides is unfortunately not so efficient as mother nature’s

process, but an extra degree of freedom is available because the direction of the synthesis

(from N to C-terminus or vice versa) is not restricted. The preparative synthesis of peptides

can be divided in two classes, i.e. peptide synthesis in solution and solid phase peptide

synthesis (SPPS). The peptides in this dissertation have been synthesised by both methods,

each of which has their advantages and drawbacks. Both methods are discussed in the next

section and are compared mutually.

Chapter 2

16

2.2 Peptide synthesis in solution

All of the dipeptides and their metal complex conjugates in this thesis have been synthesised

via reactions in solution. A dipeptide is formed by nucleophilic attack of the amine nitrogen

of one amino acid on the carboxylate carbon atom of another amino acid, under elimination of

a water molecule. To make this reaction possible under mild conditions, it is necessary to

transform the carboxylic acid (or carboxylate) into a better leaving group. In that way the

electron density on the carbonyl carbon atom decreases, making a nucleophilic attack of an

amine nitrogen atom more favourable. The reagents that transform the carboxylate moiety

into a better leaving group are called coupling reagents. Apart from the necessity of coupling

agents another prerequisite must be met to avoid side-reactions, namely protection of one of

the amino groups in order to prevent formation of homo-dipeptides. Solution peptide coupling

reactions have employed several types of reagents, such as N,N’-dicyclohexylcarbodiimid,

isobutyl chloroformiate (IBCF) and phosphonium reagents [4-6, 109]. The carbodiimid

method, first reported by Hess and Sheehan [110], has been used extensively in the past, but

the major disadvantage of this method is that N,N’-dicyclohexylurea forms, which is difficult

to separate from the peptide. Furthermore, under basic conditions a side reaction is frequently

observed, yielding a different urea-peptide derivative [111] that is even more difficult to

remove from the peptide.

The most attractive coupling method for the synthesis of dipeptides in solution appeared to be

the mixed-anhydride method, the reaction scheme of which is shown in Scheme 2.1. [112].

First the carboxylic acid of one of the amino acids is deprotonated by N-methyl morpholine

and subsequently the resulting carboxylate group is transformed into a mixed anhydride

derivative for example by reaction with IBCF, yielding a reactive intermediate. Upon

nucleophilic attack of the amine nitrogen atom of the other amino acid, the dipeptide forms

under elimination of iso-butanol and CO2. Theoretically, unwanted reactions can occur via

attack of the amine nitrogen atom on the other carbonyl carbon atom, but the corresponding

side-products were never found to be present in detectable amounts in the isolated peptides.

To prevent formation of homo-dipeptides, the NH2 group of one of the amino acids was

protected with a Boc (tert-butyloxycarbonyl) group and the carboxylate function of the other

amino acid was protected as a methyl ester. Virtually all amino acids are commercially

Peptides

17

available in their Boc and methyl ester protected forms. These two protecting groups are

orthogonal, which means that one of them can be selectively cleaved while leaving the other

unaffected. Boc groups can be removed by the reaction with CF3COOH in CH2Cl2, yielding

the trifluoro acetate ammonium salt and the volatile isobutylene and CO2. Methyl esters can

be efficiently hydrolysed under basic conditions in water or in a water / organic solvent (e.g.

MeOH or 1,4-dioxane) mixture.

Scheme 2.1 Reaction scheme for the formation of a dipeptide via the IBCF method,

according to ref. [112]

R

NH

O

PG

O

OCl

R

NH

O

PGOOH

O

O

R

NH

O

PGHN

R'O

PG'HO H2N

a

R'+

O

PG'CO2 + +

+

b

PG = N-terminus protecting group, PG’ = C-terminus protecting group

Reagents and conditions: a) N-methylmorpholine in THF, 5 min stirring; b) THF, 1hr stirring

The dipeptides used in this thesis synthesised by the mixed-anhydride method are Boc-Phe-

Leu-OMe (1a) and Boc-Ala-Phe-OMe (1b).The yield of the isolated dipeptides varies per

batch and is between 80 and 90 %, with the purity being around or higher than 95 %, as

concluded from the 1H NMR spectra. Both dipeptides were characterised by infrared

spectroscopy, electron impact mass spectrometry, 1H and 13C NMR spectroscopy (see

experimental section).

The dipeptides can be stored for years in their N and C-termini protected forms, without any

noticeable decomposition taking place. For coupling with organometallic complexes or a

functionalised di(2-picolyl)amine ligand containing an acid group (see Chapters 3, 5, 7 and 8),

the dipeptides are Boc-deprotected by reacting them with CF3COOH in CH2Cl2 (1/1 v/v) for

one hour, followed by evaporation of the mixture in vacuo. Any residual trifluoro acetic acid

is removed by addition of diethylether and subsequent evaporation to dryness in vacuo. The

Chapter 2

18

trifluoro acetate salts of the dipeptides were obtained as white hygroscopic powders in

essentially quantitative yield. After transformation into the free amine by addition of NEt3 or

dipea, the dipeptide can be tethered to the functionalised ligand or to the complexes

containing an acid functionality by employment of HBTU or TBTU (see p.viii for the

molecular structure) as the coupling reagent. In contrast to the coupling of amino acids, the

use of HBTU and TBTU is necessary because IBCF does not show any reactivity in the case

of organometallic complexes [46].

The solution phase synthesis of peptides is very efficient in the case of small peptides.

However, each coupling step usually proceeds in a 80-90 % yield, with losses occurring

during extraction, isolation and purification of the peptides. Although the yields appear to be

very good on first sight, the losses accumulate rapidly during the synthesis of larger peptides.

If one is interested in larger peptides, i.e. more 5-6 amino acids, large quantities of the starting

dipeptide are required. Five coupling-steps each proceeding in 80% yield implies that the final

peptide is isolated in a 30% yield. In addition, small amounts of side-products arise during

each coupling step [112]. The longer the peptide, the larger the similarity in properties

between the desired product and the side products, which means that the purification of the

peptide becomes more difficult. The disadvantages of the synthesis of peptides in solution in

the case of oligopeptides can be overcome by solid phase synthesis methods, which are

discussed in the next section.

2.3 Solid phase peptide synthesis

As noted above, the synthesis of larger peptides via the solution synthesis method is difficult

to achieve, primarily because the purification becomes tedious. In addition, the overall yield is

not high and this method is very-time consuming because of the necessary work-up after each

coupling step.

These shortcomings of peptide synthesis in solution stimulated the development of peptide

synthesis on a solid support, a method which is called Solid Phase Peptide Synthesis (SPPS).

One year after Merrifield’s initial report of SPPS [4], he improved the method by synthesising

a highly pure nona-peptide in a 68% yield [113]. From that time on the era of SPPS had really

started and it attracted the attention of many (bio)chemists and pharmacists. The method

Peptides

19

consists of using a resin (a solid support) to which initially an N-protected amino acid is

coupled via its carboxylate group. After a deprotection step, the next N-protected protected

amino acid is coupled to it, followed by again a deprotection step. Usually, the resin is reacted

with a three-fold or more excess of the amino acid to ensure that the yield of the coupling

reactions approaches 100%. The peptide grows stepwise on the resin, which can be washed

after each deprotection and coupling step in order to remove the side-products, excess amino

acid and coupling reagents, and only when the desired primary structure is obtained, the

peptide is finally cleaved from the resin.

Merrifield’s original method employed Boc-protecting groups for the N-termini of the amino

acids and the side-chains were protected in form of benzyl and tert-butyl groups. Removal of

the Boc group requires a CF3COOH/CH2Cl2 mixture, but cleavage of the side-chain

protecting groups can only be affected by much harsher conditions, such as highly poisonous

and corrosive liquid HF. The necessity of CF3COOH and HF stimulated the development of

other SPPS-compatible protecting groups and the Fmoc (9-fluorenylmethoxycarbonyl; see p.

viii for the molecular structure) moiety was found to be a good substitute for protecting the N-

termini of amino acids. This group can be removed under basic conditions, for example with

25 % piperidine in DMF. Another advantage of the Fmoc group is its strong absorbance

around 280 nm, which allows the progress of the deprotection steps to be monitored by UV-

Vis spectroscopy. At the present time, SPPS has become a fully automated process,

employing a peptide-synthesiser.

Although the Merrifield method still has its advocates, many chemists switched to the Fmoc

method because of its milder reagents. The pentapeptide-derivatives synthesised in this thesis,

all of which are leucine-enkephalin metal conjugates, were synthesised via the Fmoc synthesis

strategy [114]. The reason for the choice of the pentapeptide leucine-enkephalin, with primary

structure H-Tyr-Gly-Gly-Phe-Leu-OH, is that it is a naturally occurring neuro-peptide with an

action similar to that of morphine (see next section).

Scheme 2.2. Molecular structure and schematic representation of the HMBA-AM resin

CH2 NH

C

O

CH2 OH CH2OH=

Chapter 2

20

The resin employed is a HMBA-AM (HydroxyMethylBenzoic Acid-AMide) type (see

Scheme 2.2). The first amino acid, L-leucine was connected to it via an ester linkage

according to a literature procedure [114]. This step as well as the preceding ones is depicted

schematically in Scheme 2.3. Four consecutive deprotection and coupling steps (see

experimental part for details) yielded the resin-bound leucine-enkephalin and from this stage

on all deprotection steps and coupling of the metal fragment or ligand (see Chapters 5 and 7)

were performed manually. First the 2-ClTrt group on the tyrosine was cleaved by 5% tri-

isopropyl silane (TIS) and 5% CF3COOH in CH2Cl2, followed by removal of the Fmoc group

by 25% piperidine in DMF. This yields the resin bound NH2 form of the peptide, which was

used to couple the metal fragments or ligand to by employing the coupling reagent TBTU.

Cleavage from the resin was affected by a saturated NH3 solution in MeOH, yielding the

leucine-enkephalin conjugate, with the C-terminus in the form of an acid amide group. These

compounds were purified by preparative HPLC, which is a common purification step in SPPS

after cleavage of the peptide from the resin [114].

Scheme 2.3 Synthesis of the leucine-enkephalin conjugates via SPPS

CH2OH CH2O Leu-Fmoc CH2O Leu-Phe-(Gly)2-(2-ClTrt)Tyr-Fmoc

CH2O Leu-Phe-Gly-Gly-Tyr-Fmoc CH2O Leu-Phe-Gly-Gly-Tyr-H

CH2O Leu-Phe-Gly-Gly-TyrR

OH2N Leu-Phe-Gly-Gly-Tyr

R

O

d

b

f

ec

a

Reagents and conditions: a) Fmoc-leucine, MSNT (see p. viii), N-methylimidazole, 1 hr stirring in

DCM; b) four consecutive automated deprotection and coupling steps; c) 5% TIS and 5 % TFA in

DCM, 3 × 2 minutes; e) R-COOH, TBTU, dipea, 16 hrs shaking; f) sat. NH3 / MeOH, 48 hrs

2.4 Biological properties of enkephalin

The endogenous morphine-like substance enkephalin, first isolated by Hughes and co-workers

from pig brain [116], is a mixture of two pentapeptides, with primary structures H-Tyr-Gly-

Gly-Phe-Met-OH ([Met]-enkephalin) and H-Tyr-Gly-Gly-Phe-Leu-OH ([Leu]-enkephalin).

The ratio between these two peptides is species-dependent, being 4 : 1 in pig brain and reverse

Peptides

21

in cattle brain [117]. It has been demonstrated that these pentapeptides are also present in

various human tissues, such as the human brain, the human spinal fluid and in blood plasma

[118]. Both of these substances can bind to the opiate receptor, in this way acting as natural

painkillers (or endorphins). To the same receptor, the compound morphine and its synthetic

analogue heroin can bind with high affinity. The molecular structure of these two substances

is depicted in Scheme 2.4. At first sight no similarity can be observed between the rigid

condensed ring system of these compounds and the flexible pentapeptides. However, two

different X-ray crystal structures of [Leu]-enkephalin [119-121] revealed a folded

conformation, with the space-filling of the backbone, tyrosine-phenol ring and the

phenylalanine benzyl group of the peptide displaying similarities with that of the solid state

structure of morphine [122]. The strong affinity of morphine and [Leu]-enkephalin for the

opiate receptor has been attributed to this similarity of the 3D-structures.

Scheme 2.4 Two opiate-receptor binding substances; left: the alkaloid morphine, right: its

synthetic derivative heroin

HO OHO

NCH3

O O

NCH3HH

H H

C O

CH3

O

C O

CH3

However, the situation is not so clear-cut as shown above, because in two other X-ray crystal

structures, the [Leu]-enkephalin displays an extended conformation, with each molecule only

being involved in intermolecular hydrogen bond interactions [123-125].1H NMR spectra were interpreted by two different groups as indicative for a solution

structure consisting of a β-turn [126, 127], similar to the folded conformations revealed by the

first two X-ray crystal structures mentioned above. A third NMR spectroscopic study [128],

however, found no basis for intramolecular hydrogen bonding. Also the results from

Chapter 2

22

theoretical calculations indicated the extended structure to be the more stable one [129].

These conflicting results were explained by Khaled and co-workers, who demonstrated by

combining the results from a variety of spectroscopic techniques, such as NMR, CD and UV

spectroscopy, that the two conformations of [Leu]-enkephalin exist in a concentration-

dependent equilibrium [130].

It has been suggested that the two different conformations of this molecule bind to two

different subtypes of opioid receptors: the β-bend folded conformation might bind to the µ-

receptor, and the extended conformation to the δ-receptor [131, 132]. However, this

suggestion has not been completely verified up to date. A way to provide convincing evidence

for the conformation that binds to the receptor, would be to obtain an X-ray crystal structure

of the receptor with [Leu]-enkephalin bound to it. Possibly, the use of enkephalin with 4-Br-

phenylalanine might be necessary to obtain a high-resolution structure of the receptor-

substrate complex on account of the introduced heavy atom.

23

3The marker Mo(ηηηη-Cp-COOH)(ηηηη-allyl)(CO)2

3.1 Reasons for selecting molybdenum carbonyl complexes as markers

As already mentioned in Chapter 1, one of the objectives of this dissertation is the design and

synthesis of novel markers for biomolecules based on organometallic complexes that contain

carbonyl ligands, in this way making infrared spectroscopic detection of the bioconjugates

possible. Although various metal carbonyl compounds have been used for these purposes,

many of them are quite reactive towards dioxygen, which results in rapid decomposition of

the marker in dioxygen containing atmospheres. Furthermore, it would be convenient to

develop markers that allow in addition to infrared detection, also detection by electrochemical

methods. The examples of ferrocene derivatives provided in the introductory chapter show

that electrochemical detection is possible at low concentrations. For these marker purposes

molybdenum carbonyl complexes appeared attractive for three reasons. First of all, a plethora

of molybdenum carbonyl complexes has been reported, which indicates that such compounds

are quite stable. Secondly, many of these low-valent organometallic molybdenum complexes

show reversible electrochemical transitions and, thirdly, the low toxicity of the element

molybdenum makes these compounds suitable for applications in biological assays.

Chapter 3

24

The ligand to which the anchoring group is attached should be bound very tightly to the metal

ion in order to avoid dissociation of this ligand. Therefore, it should be a five or six electron

donor ligand, which implies that ligands other than a η5-cyclopentadienyl or η6-benzene ring

should be tridentate. Literature studies revealed that in particular a wealth of compounds with

Mo(η-allyl)(CO)2 and Mo(CO)3 cores, which sum 13 and 12 electrons respectively, have been

reported. Complexes with such a moiety would fulfil the two requirements, because the 18-

electron rule would be obeyed with an additional five or six electron donor ligand and,

moreover, they contain carbonyl ligands, which allows the corresponding bioconjugates to be

detected by infrared spectroscopy.

3.2 Properties of Mo(ηηηη-Cp)(ηηηη-allyl)(CO)2 and synthesis of the marker

The first attractive candidate complex to serve as an infrared spectroscopic marker appeared

to be Mo(η-Cp)(η-allyl)(CO)2, a compound that was first synthesised by Green and co-

workers in 1963 [133]. Three years later King reported the solution infrared spectrum of this

complex to consist of four carbonyl bands instead of the expected two, which indicates that

this compound exists as mixture of two isomers [134]. By subjecting this complex to variable

temperature 1H NMR spectroscopic investigations, Davison and Rode demonstrated the

isomers to be in equilibrium in solution [135], and they suggested that these differ in the

orientation of the allyl ligand, as visualised in Scheme 3.1. The exo isomer has its central

hydrogen atom pointed towards the Cp-plane, whereas the allyl ligand in the other isomer,

called endo, has undergone a 180° rotation with respect to the Mo-allyl axis, thus having its

terminal carbon atoms directed towards the Cp-plane. By comparison of the 1H NMR spectra

of Mo(η-Cp)(η-allyl)(CO)2 with those of Mo(η-Cp)(η-2-Me-allyl)(CO)2 and Mo(η-

Indenyl)(allyl)(CO)2 (Indenyl = C9H7), Faller and co-workers confirmed the assignment of the

isomers on the basis of large magnetic anisotropy effects owing to the indenyl ring [136, 137].

Scheme 3.1 Conformation of the observed isomers for Mo(η-Cp)(η-allyl)(CO)2 in solution

MoCO

CO

MoCO

CO

endoexo

The marker Mo(η-Cp-COOH)(η-allyl)(CO)2

25

The complex Mo(η-Cp)(η-allyl)(CO)2 would fulfil all requirements to serve as an infrared-

spectroscopic handle of biomolecules if substitution of the Cp-ring with an acid group could

be achieved, in this way making coupling reactions via well-established peptide-synthesis

methods possible (see Chapter 2). First, however, the parent complex was synthesised in order

to investigate its reactivity towards water and dioxygen. Especially the stability in aqueous

media is important for the possible use in biological assays. The original synthesis of the

parent compound has been improved by Hayter [138] and involves a two-step reaction

starting from molybdenum hexacarbonyl, as shown in Scheme 3.2. In a one pot reaction

Mo(CO)6 is reacted with allyl bromide in an acetonitrile/toluene mixture, leading to formation

of Mo(η-allyl)Br(CO)2(MeCN)2 [139]. The easy work-up only consists of isolation of the

product by filtration and drying it in vacuo. In the second step, the compound Mo(η-

allyl)Br(CO)2(MeCN)2 is reacted with LiCp in dry THF, resulting in displacement of the

bromide and the two acetonitrile ligands by the cyclopentadienyl ligand. This reaction

requires in addition to very dry THF also exclusion of dioxygen, whereas for the work-up

only anaerobic conditions are necessary, which indicates that the complex Mo(η-Cp)(η-

allyl)(CO)2 is not susceptible to water. In addition, this compound was found to be stable

towards dioxygen in the solid state for more than one hour. These two positive findings were

encouraging to explore how the anchoring group could be introduced onto the complex Mo(η-

Cp)(η-allyl)(CO)2.

Scheme 3.2 Route of synthesis for the parent complex Mo(η-Cp)(η-allyl)(CO)2.

MoCO

CO

Mo(CO)6 Mo(η-allyl)Br(CO)2(NCMe)2a b

Reagents and conditions: a) allyl-bromide, 6 hrs reflux in toluene / acetonitrile; b) LiCp in THF added

at –78° C, followed by 45 mins stirring at RT.

In order to obtain the derivative of this compound with an acid group on the Cp-ring, two

strategies came into consideration. The first one constitutes substitution of the Cp-ring before

it is introduced to the Mo(η-allyl)(CO)2 group, whilst the second possibility is

functionalisation of a bound Cp-ring of the parent compound Mo(η-Cp)(η-allyl)(CO)2. The

former would have the disadvantage that the acid functionality requires protection, for

Chapter 3

26

example in the form of a methyl ester, which makes a deprotection step after complex

formation necessary in order to obtain the acid. The first thing tried was to synthesise Na(Cp-

COOMe) from NaCp and dimethyl carbonate according to a literature procedure [140]. The

product obtained this way, however, was invariably not pure and, hence, the idea of

introduction of the anchoring group onto the Cp-ring prior to complex formation was not

further pursued.

An attractive alternative for obtaining the desired complex Mo(η-Cp-COOH)(η-allyl)(CO)2

appeared to be lithiation of the parent compound by using n-butyllithium, followed by

carboxylation with solid CO2, as shown in Scheme 3.3. This type of reaction is a common

method for the introduction of a carboxylic acid substituent onto a Cp-ring [141, 142], and

even the Cp-ring of a similar compound, namely Mo(η-Cp)(CO)3Me, can be carboxylated this

way [143, 144]. However, this type of reaction has thus far not been reported for metal-Cp

complexes that contain allyl ligands. The synthesis was performed in dry THF under

anaerobic conditions and after an aqueous work-up including acidification, the complex

Mo(η-Cp-COOH)(η-allyl)(CO)2 (2) was isolated in 89% yield as a yellow powder. The

evidence for the presence of a carboxylic acid was first of all obtained from the infrared

spectrum. In comparison to the IR spectrum of the parent complex, two additional vibrations

are present at 1680 cm-1 and 3450 cm-1, which are assigned as the νC=O and νOH,

respectively. The EI mass spectrum clearly shows that Mo(η-Cp-COOH)(η-allyl)(CO)2 has

formed by the presence of the M+ peak at m/z = 304. Furthermore, fragments at m/z = 276 and

m/z = 248 are observed, which are due to loss of one or two carbonyl ligands, respectively. In

addition, the results from elemental analysis are also fully consistent with the proposed

constitution.

Scheme 3.3 Synthesis of the marker complex Mo(η-Cp-COOH)(η-allyl)(CO)2 (2)

MoCO

CO

MoCO

CO

OH

O

ca , b

Reagents and conditions: a) n-butyl lithium in THF at –78° C; 15 mins stirring; b) addition of CO2 (s)

at –78°C, followed by stirring at RT for 1 hr; c) H3O+


27

The resonances in the 1H NMR spectrum of 2 at room temperature are broadened, which is

indicative for occurring coalescence, whereas two sets of signals are observed in the 13C NMR

spectrum. This shows that 2, like the parent complex Mo(η-Cp)(η-allyl)(CO)2, exists as a

mixture of the exo and endo form, but the ratio between these could not be determined

reliably at room temperature because of the broadened resonances in the 1H NMR spectrum.

Since this compound merely serves as a starting material, it was not subjected to variable

temperature NMR spectroscopic investigations.

3.2 Coupling of the marker with amino acids and peptides

As outlined in Chapter 1, an objective of this graduate work is the synthesis of new markers

and subsequent coupling of these to small biomolecules, such as amino acids and dipeptides.

These amino acid and dipeptide conjugates are still low molecular weight compounds, which

facilitates their purification as well as the investigation of their (spectroscopic) properties. It is

of practical interest to tether the marker to oligopeptides by SPPS only when these small

conjugates already exhibit good stability towards dioxygen and water.

Scheme 3.4 Synthesis and constitution of the amino acid and dipeptide derivatives

MoCO

CO

OH

O

MoCO

CO

AA-R

O

H-AA-R+

H-AA-R

3a Phe-OMe3b Leu-NH23c Gly-OMe4 Phe-Leu-OMe

a

Reagents and conditions: a) HBTU in DMF / NEt3; 45 mins stirring at RT

The following amino acid derivatives were selected for coupling to complex 2: phenylalanine

methyl ester, glycine methyl ester, leucine amide and the dipeptide H-Phe-Leu-OMe. The

presence of the carboxylic acid on the organometallic complex makes coupling reactions via

standard peptide synthesis methods possible. The amino acid conjugates 3a-3c (see Scheme

3.4) were synthesised by reacting 2 for 45 minutes at ambient temperature under an

atmosphere of argon with stoichiometric amounts of the appropriate amino acid in a DMF /

NEt3 mixture, employing HBTU as the coupling reagent. Upon addition of an aqueous 2M

NaHCO3 solution, the compounds precipitated as yellow powders, which were collected by

Chapter 3

28

filtration and dried in vacuo, with yields varying between 67% (for 3c) and 91% (for 3a). The

amino acid derivatives 3a-3c differ in C-terminal protecting groups (NH2 for 3b and OCH3

for 3a and 3c) and complexity because 3c is achiral, whereas 3a and 3b are chiral. To

investigate whether coupling of 2 could also be expanded to larger biomolecules, this

complex was reacted with the dipeptide H-Phe-Leu-OMe in a similar manner. After an

identical work-up, the dipeptide derivative Mo(η-C5H4-C(O)-Phe-Leu-OMe)(η-allyl)(CO)2

(4) was obtained in a 62% yield.

Evidence for the assumed constitution of 3a-3c and 4 was first of all obtained from the results

from elemental analysis, all of which fit well. Furthermore, the electron impact mass spectra

of 3a-3c and 4 show a general feature: apart from the molecular peak with the appropriate

mass, fragments owing to the loss of carbonyl ligands are observed. Moreover, these

fragments as well as the M+ peak show a mass-distribution consistent with the isotopic

distribution of the element molybdenum.

The purity of the isolated compounds was around 90-95 %, as concluded from their 1H NMR

spectra, but analytically pure microcrystalline samples for 3b, 3c and 4 could be obtained by

slow pentane diffusion into a THF solution at +4° C. In the case of the phenylalanine

derivative 3a, this purification step resulted in formation of X-ray quality crystals. Below the

X-ray crystal structure of 3a is described and compared to related compounds in the literature.

Thereafter, the 1H NMR spectra are discussed followed by presentation of the infrared

spectroscopic data for 3a-3c and 4.

3.3 X-ray crystal structure of the phenylalanine derivative

In order to elucidate the solid state structure of 3a, a single crystal of this compound was

subjected to X-ray analysis. The unit cell of 3a was found to consist of two

crystallographically independent molecules, which are depicted in Figure 3.1. Geometrical

details for these two molecules, named 3aA and 3aB, are summarised in Table 3.1.

Both molecules contain an identical ligand-set, namely a trihapto allyl ligand, two carbonyl

ligands and a substituted pentahapto Cp-ring, but the geometry of these two molecules differs


29

in the arrangement of the allyl and carbonyl ligands and the folding of the Phe-OMe

substituent. In molecule 3aA, the allyl ligand is located towards the substituent on the Cp-

ring, while the carbonyl ligands point toward the unsubstituted part of the Cp-ring. In

molecule 3aB, the Phe substituent is folded upward and the Mo(η-allyl)(CO)2 unit has

undergone a 150° rotation about the Mo-Cp(centroid) axis with respect to 3aA. Thus, the

relative orientation of the allyl and carbonyl ligands for 3aB is the opposite from 3aA.

Figure 3.1 ORTEP plot for both independent molecules of 3a (left: 3aA, right: 3aB)

Table 3.1. Selected bond lengths (Å) for both independent molecules of 3a

Molecule 3aA Molecule 3aB

Mo(1)-C(1) 2.315(7) Mo(2)-C(31) 2.313(9)

Mo(1)-C(2) 2.327(10) Mo(2)-C(32) 2.382(8)

Mo(1)-C(3) 2.348(9) Mo(2)-C(33) 2.399(10)

Mo(1)-C(4) 2.363(8) Mo(2)-C(34) 2.361(9)

Mo(1)-C(5) 2.354(7) Mo(2)-C(35) 2.309(7)

Mo(1)-C(19) 2.250(11) Mo(2)-C(49) 2.402(11)

Mo(1)-C(20) 2.264(10) Mo(2)-C(50) 2.261(11)

Mo(1)-C(21) 2.52(2) Mo(2)-C(51) 2.366(12)

Mo(1)-C(22) 1.977(12) Mo(2)-C(52) 1.952(11)

Mo(1)-C(23) 1.968(12) Mo(2)-C(53) 1.952(11)

Chapter 3

30

Rotation of a Cp-ring about the metal-Cp axis is a process with a low activation barrier [145].

However, the concomitant presence of two different rotational isomers of this kind in a unit

cell has been rarely observed before. Two examples are the W(CO)3X (X = I, CH3) fragment

in the X-ray crystal structures of W(η-C5H4R)(CO)3X (R = succinamidyl ester) [146].

Another remarkable feature of this crystal structure is the endo conformation of the allyl

ligand. All solid-state structures reported thus far for Mo(η-C3H5)(CO)2 compounds with

various Cp-derivatives display an exo conformation of the allyl ligand [147-151]. For the

methyl-allyl analogue Mo(η-Cp)(η-2-Me-C3H4), an allyl-endo conformation was observed in

the solid state [152], but in this case the allyl-exo conformation is probably unfavourable due

to steric interactions of the methyl group with the Cp-ring.

The crystal structure of 3a clearly shows that the endo conformation is favoured, but large

anisotropic displacement parameters of the allyl-carbon atoms of both molecules indicate a

disorder of these moieties. The nature of this disorder can either be static or dynamic, but

unfortunately a split-atom model did not improve the quality of the structure. Therefore the C-

C distances of the allyl ligands were refined to be equal within certain errors, which makes

discussion of these bond lengths not possible. Due to the poor quality of the crystals (thin

needles), the esd’s of all bond lengths are relatively large. Hence, a detailed discussion of the

bond lengths of 3aA and 3aB as well as detailed comparison of these with analogous

compounds is not possible. As a matter of fact, all Mo-donor atom bond lengths for 3aA and

3aB are equal within 3σ. The Mo-C(carbonyl) distances of both molecules are similar to those

reported for Mo(η-Cp)(η-allyl)(CO)2 and its derivatives [147-151]. The average Mo-CCp

distances (3aA 2.34 Å, 3aB 2.35 Å) are in the range reported for Mo(η-Cp)(η-allyl)(CO)2

(average Mo-CCp distance = 2.33 Å) [147] and Mo(η-C5H4-C(O)CH3)(η-allyl)(CO)2 (average

Mo-CCp distance = 2.35 Å) [148].

It is observed that the amide bonds in both molecules have a trans-configuration. Usually all

amide bonds in peptides between non-proline amino acids are trans-configured due to

decreased steric interference for this conformation and, quantitatively, the trans- peptide bond

is about 33 kJ mol-1 higher than the cis-form [7]. In the case of proline residues, this energy

difference is considerably less, which accounts for the frequently occurring cis-peptide bond

involving this amino acid [7].


31

In the crystal lattice, 3a forms infinite chains with alternating molecules 3aA and 3aB. The

monoclinic space group C2 implies translational symmetry in such way that each of the

molecules 3aA and 3aB shows an identical orientation. The amide groups of neighbouring

molecules are involved in hydrogen bond interactions, with N⋅⋅⋅O contacts of 2.879 Å (for

N(7)⋅⋅⋅O(36)) and 2.919 Å (for N(37)⋅⋅⋅O(6)). These distances are typical for hydrogen bonds

between amide-moieties in peptides [7, 8].

3.4 1H NMR spectroscopic investigations

X-ray crystallography revealed the endo-conformation of the allyl ligand to be the major

component for 3a in the crystalline state. For the compounds 3a-3c and 4, the situation is

more complex in solution and both isomers are present, as concluded from the observation of

two sets of resonances for the allyl hydrogen atoms in the 1H NMR spectra. The ratio between

both isomers is about 4:1 for 3a-3c and 4 in acetone-d6 as well as in toluene-d8 and it is

apparently not affected by the solvent and the nature of the substituent. Exact ratios will be

presented later, but first it will be derived which isomer (exo or endo) is thermodynamically

favoured in solution.

The presence of chiral centres in the case of 3a, 3b and 4 has a significant impact on the

appearance of their 1H NMR spectra in comparison to the 1H NMR spectrum of 3c. First the1H NMR spectroscopic behaviour of 3c will be treated, followed by discussion of the 1H

NMR spectrum of 3a. These two compounds have the same C-terminus protecting group and

only differ in the side chain of the amino acid.

Selected regions from the 1H NMR spectra of 3c in toluene-d8 are displayed in Figure 3.2 at

two different temperatures. At 283 K, the resonances owing to the allyl hydrogen atoms for

both the exo and endo conformation appear as well-resolved signals. Due to rapid rotation of

the Cp-ring, both of the syn and anti hydrogen atoms of each conformer are equivalent.

Consequently, the resonances owing to the allyl hydrogen atoms appear as three signals in a 2

: 2 : 1 ratio for each isomer. At 353 K, rapid exchange between the minor and major isomer

on the NMR timescale occurs, which leads to observation of an averaged signal for the allyl

hydrogen atoms, also with intensity 2 : 2 : 1. Faller and coworkers reported the central

hydrogen atom of the exo-isomer of Mo(η-Cp)(η-allyl)(CO)2 to be deshielded with respect to

that of the endo-isomer, due to the small magnetic anisotropy effects of the Cp-ring [137].

Because the central allyl-hydrogen atom resonance of the major isomer of 3c is downfield

Chapter 3

32

compared to that of the minor isomer, it can be assumed that the major isomer in solution has

the allyl-exo conformation. Further evidence for this is obtained by comparing the 1H NMR

spectrum of 3c with that of the acetyl substituted compound Mo(η-Ac-Cp)(η-allyl)(CO)2

[148]. First of all, this acetyl compound was found to exist predominantly (>95%) in the exo

form in solution. Considering the similarity of the substituent of this complex compared to 3c

(planar acetyl vs planar amide), it is very likely that the major isomer in compound 3c also has

an allyl-endo conformation. Furthermore, the central hydrogen atom of the allyl ligand in this

compound resonates at 3.82 ppm [148], which is very similar to the chemical shift of the

major isomer in 3c (3.68 ppm).

Figure 3.2 Parts of the 1H NMR spectra of 3c (400 MHz; toluene-d8) at 283 and 353 K.Open and filled circles: allyl signals from the major and minor isomer, respectively.

Allyl H-atom signals in the 353 K spectrum are marked with a square. * = impurity

Selected parts from the 1H NMR spectra of 3a at two different temperatures in toluene-d8 are

depicted in Figure 3.3. Due to the presence of a chiral centre on the amino acid substituent,

both of the syn and anti hydrogen atoms are magnetically inequivalent. Therefore, at low

temperatures ten signals owing to the hydrogen atoms of the allyl ligand are observed, i.e. five

for each isomer. At higher temperatures, these turn into five signals of equal intensity due to

rapid exo-endo interconversion on the NMR timescale. The overlap of two resonances in both

the low and high temperature 1H NMR spectrum is coincidence. In the spectrum at 273 K, it is

observed that the central hydrogen atom of the major isomer resonates about 0.30 ppm

downfield compared to the minor isomer (3.62 vs 3.31 ppm). For similar reasons as discussed

above for 3c, it can be concluded that the exo form of 3a is thermodynamically favoured in

solution.


33

Figure 3.3 Parts of the 1H NMR spectra of 3a (400 MHz; toluene-d8) at 283 and 363 K.

Open and filled circles: allyl signals from the major and minor isomer, respectively.

Allyl H-atom signals in the 363 K spectrum are marked with a square. * = impurity

The appearance of the 1H NMR spectra in acetone-d6 at room temperature corresponds to that

of the high temperature 1H NMR spectra in toluene-d8. The spectroscopic data in acetone have

been reported for characterisation, due to better solubility of the compounds in this solvent.

All of the allyl hydrogen atom resonances for the exo and endo isomers of 3a-3c and 4 could

be assigned by 2D-NMR techniques, such as 1H-1H COSY, 1H-13C HMQC and 1H-1H

NOESY. Especially the NOESY spectra were very informative, because positive cross-peaks

in those spectra indicate chemical exchange between the same hydrogen atom in two different

chemical environments [153].

In aceton-d6 at 300 K, the ratios between the exo and endo isomer are 80 / 20 (for 3a) and 79 /

21 (for 3c), which are very similar to the ratios in toluene-d8 (78 / 22 for 3a and 79 / 21 for

3c). These values were determined by careful integration of suitable allyl hydrogen atom

resonances. A more pronounced solvent dependence of the exo / endo ratio has been reported

for Mo(η-Cp)(η-allyl)(CO)2 [137, 154]. The exo isomer of that compound has the higher

dipole moment and, thus, it is stabilised in polar solvents [137, 154]. Evidently, the overall

dipole moment of the amino acid derivatives described in this chapter is dominated by these

substituents and there is only a minor contribution, if any, from the metal-allyl moiety.

Chapter 3

34

A 79 / 21 ratio between the two species at 300 K implies a thermodynamic energy difference

of 3.3 kJ mol-1. Although the exo isomer for 3a-3c and 4 is thermodynamically favoured in

solution, the X-ray crystal structure of 3c revealed the endo isomer to be the major component

in the crystalline solid. This might be attributed to favourable packing effects for the endo

isomer in the solid state, but it must be noted that the crystals were grown by slow diffusion of

pentane into a THF solution and the isomer ratio in THF has not been investigated for any of

these compounds.

3.5 Infrared spectroscopy

Selected vibrational data for 3a-3c and 4 are presented in Table 3.2. The presence of hydrogen

bonds in the solid state of 3a, as revealed by its X-ray crystal structure, is also clearly

reflected in the infrared spectra of this compound. In KBr, the νNH stretching vibration is

broad and centred around 3307 cm-1, whereas it shifts to 3429 cm-1 and becomes much

sharper in a CH2Cl2 solution. Similar shifts are observed for 3b, 3c and 4, which indicates that

the amide groups of these compounds are also involved in hydrogen bonds in the solid state.

Table 3.2. Selected vibrational data for 3a-3c and 4

Compound νNH (KBr)[a] νNH (CH2Cl2)[a, b] νCO (KBr)[a] νCO (CH2Cl2)[a, c]

3a[d] 3307 3429 1966, 1945, 1867 1950, 1869

3b 3238 3515[e], 3402 1947, 1869 1953, 1870

3c 3279 3452 1952, 1869 1952, 1869

4 3296 3420 1951, 1868 1953, 1870

[a] in cm-1; [b] 1 × 10-2 M; [c] 1 × 10-3 M; [d] X-ray analysis revealed the presence of two

crystallographically independent molecules in the unit cell; [e] vibration owing to amide NH2


35

Not surprisingly, in both the KBr and solution infrared spectra, the νCO vibrations owing to

the metal bound carbonyls display the highest intensity. The solution infrared spectra provide

information on the carbonyl vibrations of isolated molecules. In KBr on the other hand, the

carbonyl vibrations can be influenced by lattice effects, in this way possibly resulting in more

than the expected number of vibrations. In fact, three νCO vibrations are observed in the KBr

spectrum of 3a, perhaps due to the existence of two independent molecules in the unit-cell,

each of which is expected to display different carbonyl vibrations.

In the solution infrared spectra of 3a-3c and 4 only two carbonyl vibrations can be observed,

although 1H NMR spectroscopic investigations revealed that both the exo and endo isomer are

present in an approximate 4:1 ratio. This is in contrast with the infrared spectrum of the parent

complex Mo(η-Cp)(η-allyl)(CO)2, which displays four bands, i.e. two bands for each isomer

[134]. The bands owing to the exo and endo isomer in the infrared spectrum of the parent

complex in cyclohexane are separated by approximately 10 cm-1. In principle, each isomer of

3a-3c and 4 should display slightly different vibrations, but these could occur at nearly

identical wavenumbers. Furthermore, it should be noted that low-polarity solvents, such as

alkanes, are mandatory for obtaining small line-widths [155]. Unfortunately, the spectra of

3a-3c and 4 needed to be recorded in dichloromethane due to insolubility of these compounds

in alkanes. Therefore, it might be that any small difference between the exo and endo isomers

of 3a-3c and 4 is obscured by the natural line-broadening in that solvent. In fact, for

application as biological markers the signal-to-noise ration would be enhanced as a

consequence of this overlap.

3.6 Electrochemistry and air-sensitivity

The results presented in the previous section show that Mo(η-Cp)(η-allyl)(CO)2 can serve as

an effective infrared spectroscopic marker. Because an aim of this thesis is the development

of markers that also permit detection by electrochemistry, first compound 2 was subjected to

electrochemical investigations. Complex 2 displayed a semi-reversible oxidation at +0.42 V

vs. Fc/Fc+ in acetonitrile (0.5 mM 2; 0.1 M NBu4PF6 as supporting electrolyte). At low scan-

rates this oxidation is irreversible, whereas it approaches a reversible appearance at scan rates

Chapter 3

36

higher than 2 V s-1. By analysis of the cyclic voltammograms at different scan rates, the life-

time of the oxidised species is estimated to be 0.5 second. Because complex 2 is a starting

material and contains a possibly destabilising carboxylic acid group, also the electrochemical

behaviour of a bioconjugate of this complex was investigated. Compound 3a displays a semi-

reversible oxidation at +0.30 V vs Fc/Fc+ in MeCN. Similar to the behaviour of 2, the

oxidation of 3a also approaches reversibility at higher scan rates and a life-time of

approximately 3 seconds was estimated.

These results show that this marker does not allow detection by electrochemistry. In the solid

state, the compounds 3a-3c and 4 are stable for several hours in air, but after that period the

surface becomes green and eventually brown. In solution, however, these compounds are

much more sensitive towards dioxygen, with decomposition taking place within 30 minutes in

air. This marker was therefore not tethered to larger biomolecules via solid phase peptide

synthesis methods. Instead, the spectroscopic and electrochemical properties of another class

of allyl dicarbonyl compounds possessing a tridentate histidinate ligand were investigated,

which is the subject of the next chapter.

37

4Fluxional processes in complexes of the type

Mo(His)(ηηηη-2-R-allyl)(CO)2

4.1 Introduction

As concluded from Chapter 3, the complex Mo(η-Cp-COOH)(η-allyl)(CO)2 can be

covalently linked to amino acids and peptides to serve as an infrared spectroscopic marker,

but detection by electrochemical methods is not possible, because the conjugates display

irreversible oxidations. Moreover, the stability of these compounds towards dioxygen is not

satisfying to such a degree to further explore the use of this marker for the labelling of larger

biomolecules. However, the fact that a large number of complexes with the Mo(η-allyl)(CO)2

moiety has been reported indicates that these compounds are not necessarily unstable and

reactive species. The easy accessibility of a suitable precursor for the Mo(η-allyl)(CO)2 group

in the form of Mo(η-allyl)(Br)(CO)2(NCMe)2 stimulated further investigations on markers

with this moiety. The idea was to replace the Cp-ring by a preferably bulkier non-

organometallic ligand, which possibly might result in diminished reactivity towards dioxygen.

The Cp-ring is a five-electron donor (in the covalent electron-count model), which implies

that the new ligand should be a tridentate L2X ligand, with two nitrogen donors and one

Chapter 4

38

alkoxo or carboxylato oxygen donor. On the basis of these rationalisations, a suitable

candidate appeared to be the deprotonated histidine ligand. Another advantage of the

histidinato ligand is the possibility to substitute the hydrogen atom on the uncoordinated

imidazole nitrogen atom by an alkyl-group, in this way introducing a functional group via

which this complex can be tethered to biomolecules. In fact, as shown in the next section,

introduction of a methyl propionate substituent on the Nε atom is easily accomplished.

In this chapter, the spectroscopic and electrochemical properties of the complex

Mo(L-His)(η-allyl)(CO)2 and its Nε -methyl propionate substituted derivative are investigated

in relation to analogous complexes with the 2-Me-allyl ligand. The objective behind this was

to determine which of the two types of complexes is more suitable for the labelling purposes

of biomolecules. During the course of these studies, it was found that these compounds are

fluxional in solution, both in their neutral and one-electron oxidised form. Herein,

spectroscopic investigations are combined with results from Density Functional Theory

calculations, leading to a detailed insight of the involved species and their interconversion

pathways.

4.2 Synthesis of the complexes

The route of synthesis via which the complexes for this study were synthesised is shown in

Scheme 4.1. The reaction of Mo(η-allyl)(Br)(CO)2(NCMe)2 [139] with equimolar amounts of

L-histidine and KOH in a MeOH / H2O mixture at room temperature yields Mo(L-His)(η-

allyl)(CO)2 (5) in excellent yield. This method is more convenient than the previously

reported route [156], which consists of reacting K[Mo(L-His)(CO)3] with allyl bromide,

because this tricarbonyl compound is highly sensitive towards dioxygen. The yield and purity

of isolated 5 synthesised by the new route were higher than via the previously reported one. In

an analogous manner to the synthesis of 5, the complex Mo(L-His)(η-2-Me-allyl)(CO)2 (6) is

obtained in good yield by reacting Mo(η-2-Me-allyl)(Cl)(CO)2(MeCN)2 [139] with

stoichiometric amounts of L-histidine and KOH in EtOH. In contrast to the synthesis of 5,

EtOH was used in this case rather than MeOH because 6 is much more soluble in MeOH than

5 and does not precipitate from that solvent. Also for this 2-Me-allyl derivative, the new

method is more efficient than the originally reported one by Beck [157].

Fluxional processes in complexes of the type Mo(His)(η-2-R-allyl)(CO)2

39

Scheme 4.1 Route of synthesis and constitution of the complexes 5-8

N

HN

H2N

O O

MoOC

OC

R

N

N

H2N

O OMo

OC

OC

Mo(η-2-R-allyl)X(CO)2(NCMe)2 a

R

5 R = H

6 R = CH3

H3COO

b7 R = H

8 R = CH3

R = H, X = BrR = CH3, X = Cl

Reagents and conditions: a) L-histidine and KOH in MeOH / H2O (for 5) or EtOH / H2O (for 6),

stirring at RT for 30 mins;b) 3-bromo-methylpropionate + Cs2CO3 in DMF, heating at 80°C for 2 hrs.

By reacting 5 and 6 for two hours at 80° C with 3-bromo-methyl propionate in DMF in the

presence of excess Cs2CO3, the Nε-alkylated compounds Mo(L-His-Nε-C2H4CO2Me)(η-

allyl)(CO)2 (7) and Mo(L-His-Nε-C2H4CO2Me)(η-2-Me-allyl)(CO)2 (8) were obtained in

good yield. The purity of these complexes is around 95% as concluded from analytical HPLC

investigations, but purification by preparative HPLC yields analytically pure (purity > 99.8

%) samples. By using a preparative column with a large diameter (see experimental part for

characteristics), a few grams of each compound can be purified per day.

4.3 Solid state structures

X-ray quality crystals could be obtained for all four compounds. The details concerning how

these crystals were grown can be found in the experimental section. ORTEP diagrams for the

structures 5⋅⋅⋅⋅MeOH and 6 are shown in Figure 4.1 and those for the structures 7⋅⋅⋅⋅2MeOH and 8

are depicted in Figure 4.2, with relevant bond distance information summarised in Table 4.1.

As far as possible, similar atom labelling schemes for those four X-ray crystal structures have

Chapter 4

40

been adapted, in this way facilitating the discussion. The unit cell of 6 contains two

crystallographically independent molecules, which display only slightly different bond

lengths and bond angles. Because the bond distances of these two molecules are all within 3σ,

only those for one of the molecules are presented and discussed here. However, for the

discussion of intermolecular hydrogen bond interactions both molecules will be treated (see

below) and atoms belonging to the second molecule will be denoted with a prime.

Figure 4.1 ORTEP plots for 5⋅MeOH (left) and for one of the crystallographically

independent molecules of 6 (right)

The MeOH molecule in 5⋅MeOH is disordered over two positions and is involved in a

hydrogen bond interaction with the carboxylate group (average O(50)···O(10) = 2.74 Å). In

7⋅⋅⋅⋅2MeOH on the other hand, the MeOH molecules do not show any sign of disorder, with the

hydroxyl group of one MeOH molecule being hydrogen bonded to O(9) (O···O contact =

2.750(10) Å) and the OH group of the other MeOH molecule forming a hydrogen bond to

O(10) (O···O contact = 2.800(10) Å). Because the Mo(1)-O(9) bond lengths in 5⋅MeOH and

7⋅⋅⋅⋅2MeOH are equal within 3σ, these hydrogen bonds in those two structures apparently do

not influence the Mo(1)-O(9) bond length significantly. Therefore, the disordered MeOH

molecule in 5⋅⋅⋅⋅MeOH and the two MeOH molecules in 7⋅⋅⋅⋅2MeOH have been omitted for

clarity in the ORTEP projections.


41

Figure 4.2 ORTEP projections for 7⋅2MeOH (left) and 8 (right)

Table 4.1 Selected bond lengths (Å) for 5⋅⋅⋅⋅MeOH, 6, 7⋅⋅⋅⋅2MeOH and 8

5⋅⋅⋅⋅MeOH 6[a] 7⋅⋅⋅⋅2MeOH 8

Mo(1) - N(1) 2.217(2) 2.254(5) 2.216(2) 2.2699(11)

Mo(1) - O(9) 2.222(2) 2.150(5) 2.214(2) 2.1517(11)

Mo(1) - N(11) 2.272(2) 2.276(5) 2.256(2) 2.2711(12)

Mo(1) - C(30) 1.945(2) 1.955(7) 1.944(3) 1.969(2)

Mo(1) - C(40) 1.948(2) 1.936(6) 1.945(3) 1.951(2)

Mo(1) - C(20) 2.326(2) 2.336(7) 2.328(3) 2.337(2)

Mo(1) - C(21) 2.212(3) 2.216(7) 2.209(2) 2.225(2)

Mo(1) - C(22) 2.320(2) 2.319(7) 2.328(3) 2.304(2)

[a] The unit cell contains two crystallographically independent molecules with similar bond lengths.

The coordination sphere around the Mo-atoms in 5⋅MeOH, 6, 7⋅2MeOH and 8 is comprised of

identical donor atoms, namely two carbon atoms from two carbonyl ligands, three carbon

atoms from an allyl (for 5⋅MeOH and 7⋅2MeOH) or 2-Me-allyl (for 6 and 8) ligand and the

carboxylato oxygen atom, amine and Nδ nitrogen atoms of either a histidine (for 5⋅MeOH and

6) or Nε-substituted histidine (for 7 and 8). The Mo(η-allyl)(CO)2 (for 5⋅MeOH and

Chapter 4

42

7⋅2MeOH) and Mo(η-2-Me-allyl)(CO)2 (for 6 and 8) moieties are in a facial arrangement

with the terminal CH2-carbon atoms of the allyl or 2-Me-allyl ligand oriented towards the

carbonyl ligands. Such a conformation has been shown to be energetically favourable by

Extended Hückel Molecular Orbital (EHMO) calculations [158] and is observed in all solid

state structures of this type reported thus far (for representative examples see references [159-

169]).

Even at first glance, an interesting conformational difference between the allyl and the 2-Me-

allyl ligand in those four X-ray crystal structures is evident. Whereas the allyl ligand occupies

a position trans to the Nδ nitrogen atom of the (substituted) histidine, the 2-Me-allyl ligand is

located trans to the carboxylate oxygen atom of the amino acid. A slight alteration of the

electronic and steric properties of the allyl ligand by introduction of a methyl group on its

central carbon atom apparently results in a dramatic conformational change in the solid state.

Such a conformational difference between the allyl compounds and the 2-Me-allyl analogues

has not been observed before for pseudo-octahedral complexes of group VI elements. A

plausible explanation for this might be that in related group VI complexes with a tridentate

ligand reported thus far, the ligand itself possesses two or more identical donor atoms. In the

case of a tridentate ligand containing two chemically equivalent donor atoms the possibility

for observing a different conformational isomer for the allyl complex with respect to the 2-

Me-allyl complex is decreased from a statistical point of view, whereas this observation is

fully precluded in the case of three identical donor atoms.

The effect of the different conformation of the allyl complexes in comparison to the 2-Me-

allyl compounds is clearly reflected in some of the bond distances. As shown in Figures 4.1

and 4.2, the primary amino group in these four structures is always in a trans-position to a

carbonyl ligand. Therefore, unsurprisingly, the Mo(1) - N(11) bond distances of 5⋅⋅⋅⋅MeOH, 6,

7⋅⋅⋅⋅2MeOH and 8 are very similar. Only in 7⋅⋅⋅⋅2MeOH, the Mo-N(11) bond length is about

0.015-0.020 Å shorter with respect to the other three structures, but this might be due to

packing effects. The situation is completely different with the Mo(1) - O(9) and the Mo(1) -

N(1) distances. The Mo(1)-O(9) bond distance in 5⋅⋅⋅⋅MeOH and 7⋅⋅⋅⋅2MeOH is about 0.06-0.07

Å shorter in comparison to 6 and 8, whereas the Mo(1)-N(1) bond length in 5⋅⋅⋅⋅MeOH and

7⋅⋅⋅⋅2MeOH is about 0.040-0.055 Å longer compared to 6 and 8. These variations of the Mo(1)-

N(1) and Mo(1)-O(9) bond distances can be rationalised on the basis of the trans-influence of

the allyl and 2-Me-allyl ligands.


43

The two Mo-C(carbonyl) distances in both 5⋅⋅⋅⋅MeOH and 7⋅⋅⋅⋅2MeOH are nearly identical. In

both 6 and 8 however, Mo(1)-C(40) is about 0.02 Å shorter than Mo(1)-C(30), although it

must be noted that those bond lengths in 6 are within 3σ due to the relatively large esd’s. The

carbonyl ligand C(40)-O(40) in 6 and 8 is in a trans-position with respect to N(1) and the

slightly longer Mo(1)-C(40) bond might be caused by the π-acceptor properties of the

imidazole-group.

The allyl ligand is bound symmetrically to the molybdenum ion in 7⋅⋅⋅⋅2MeOH and almost

symmetrically in 5⋅⋅⋅⋅MeOH, with the Mo(1)-C(20) and Mo(1)-C(22) bond lengths in the latter

structure differing by 0.006 Å, but being within 3σ. In both structures with the 2-Me-allyl

ligand, Mo(1)-C(20) is slightly longer compared to Mo(1)-C(22). In 6 those two bond lengths

are within 3σ due to the larger standard deviations, but these are significantly different in 8.

These variations between the Mo-C bond distances of the allyl and 2-Me-allyl ligands cannot

be attributed with certainty to the steric and electronic differences between these two ligands,

because the donor atom that is in a trans-position relative to these ligands is not identical.

Apart from the hydrogen bonds of the two MeOH molecules with the carboxylate

group of the histidinato ligand in 7⋅⋅⋅⋅2MeOH, no other hydrogen bonds are present in that

structure. Also in 8 no hydrogen bonds exist in the solid state, and therefore the unit cell of

that compound consists of isolated molecules.

In addition to the hydrogen bond interaction of the disordered MeOH molecule with O(10) in

5⋅⋅⋅⋅MeOH, another intermolecular hydrogen bond is present in the structure between N(3) and

O(9) of a neighbouring molecule (N⋅⋅⋅O contact = 2.785 Å). Although the unit cell of

5⋅⋅⋅⋅MeOH contains a right-handed three-fold screw axis (space group P3121), the complex

molecules are not arranged in a helical fashion around this screw axis, but instead form a

three-dimensional hydrogen bonding network.

The hydrogen-bonding pattern in the lattice of 6 is more complicated, with the carboxylate

group of each crystallographically independent molecule being involved in one shorter and

one longer intermolecular hydrogen bond, also in this case resulting in a three-dimensional

network. The short hydrogen bonds are between molecules of the same crystallographic type,

whereas the longer hydrogen bonds form to the other crystallographically independent

molecule. Strong hydrogen bonds exist between N(3) and O(9) (N⋅⋅⋅O contact = 2.813 Å) and

between N(3)’ and O(10)’ ((N⋅⋅⋅O contact = 2.811 Å). Weaker hydrogen bonds are present

between O(10) and N(11)’ (N⋅⋅⋅O contact = 3.167 Å) and between O(10)’ and N(11) (N⋅⋅⋅O

contact = 2.913 Å).

Chapter 4

44

My interest concerned the 1H and 13C NMR spectra of AsPh4[Mo(L-His)(CO)3] (9) for

comparison with the NMR spectra of 5-8 (see next section). By slow evaporation of a MeOH /

H2O solution under a stream of argon, single crystals were obtained, which were subjected to

X-ray structure determination. An ORTEP representation of the anionic part of 9·H2O is

depicted in Figure 4.3, with relevant bond-length information given in Table 4.2.

Figure 4.3 ORTEP projection for the anionic part of the X-ray crystal structure 9·H2O

Table 4.2 Selected bond distances (Å) for 9·H2O

Mo(1)-N(1) 2.286(4) Mo(1)-C(30) 1.916(4)

Mo(1)-N(11) 2.290(4) Mo(1)-C(40) 1.917(5)

Mo(1)-O(9) 2.240(3) Mo(1)-C(50) 1.946(5)

The Mo atom in 9·H2O is coordinated in a distorted octahedral geometry by three carbonyl

ligands and by the carboxylato oxygen atom, the Nδ and the amine nitrogen atom of a

tridentate histidine. The complex displays two shorter Mo-C(carbonyl) bond distances around

1.917 Å (Mo(1)-C(30) and Mo(1)-C(40)) and a slightly longer one (Mo(1)-C(50) 1.946(5) Å).

It must be noted that Mo(1)-C(40) and Mo(1)-C(50) are just within 3σ, whereas this is not the

case with Mo(1)-C(30) and Mo(1)-C(50). The carbonyl ligand with the slightly longer Mo-


45

C(carbonyl) distance is trans relative to the histidine Nδ nitrogen atom. This bond length

elongation is likely caused by the better π-acceptor ability of the imidazole ring compared to

the amino and carboxylato group, both of which are pure σ-donors. The distances between the

Mo atom and the histidine donor atoms are all slightly longer in 9·H2O compared to the

structures 5⋅⋅⋅⋅MeOH, 6, 7⋅⋅⋅⋅2MeOH and 8, which is likely owing to the overall negative charge

of the complex of 9·H2O. Also the observation of the shorter average Mo-C(carbonyl) bond

length in 9·H2O compared to the other structures is consistent with a higher electron density at

the metal atom, as a consequence of the lower oxidation state.

Although AsPh4+ is a bulky cation, there are intermolecular hydrogen bond interactions in the

lattice of 9·H2O. The carboxylate group is involved in two hydrogen bonds, namely between

O(9) and N(3) of a neighbouring anionic complex (N⋅⋅⋅O contact = 2.805 Å) and between

O(10) and the oxygen atom (labelled O(60); not displayed in Figure 4.3) of the water

molecule (O⋅⋅⋅O contact = 2.751 Å), resulting in a zigzag pattern. Only one of the hydroxyl

groups of the water molecule is involved in this hydrogen bond, whereas the other one is not

hydrogen bonded. Furthermore, O(60) forms a weak hydrogen bond to one of the hydrogen

atoms of the coordinated NH2 group (N⋅⋅⋅O contact = 3.029 Å). The packing in 9·H2O is such,

that layers of AsPh4+ cations and layers of complex anions and water molecules alternate.

4.4 Behaviour in solution

All compounds crystallise uniformly and there is no sign of disorder of the complexes in the

solid state structures discussed so far, but the situation is different in solution. For 5 and 7,

two sets of signals are observed in both the 1H NMR and 13C NMR spectra. Parts of the 1H

and 13C NMR spectra of 5 are shown in Figure 4.4, together with the assignment of the

signals. In the 13C NMR spectra, the CO ligands show two pairs of resonances at 230.0 / 228.8

ppm and 228.2 / 227.9 ppm and also most of the resonances owing to the histidine-imidazole

group are doubled, except the one for Cδ (see Figure 4.4). All of the other signals in the 13C

NMR spectrum of 5, except the resonance originating from Cβ, also appear as two pairs (see

experimental part). In the 1H NMR spectrum of 5 the resonances also appear as two sets of

signals and these are especially clearly discernible for the signals owing to the imidazole CH

Chapter 4

46

hydrogen atoms and the NεH hydrogen atom (see Figure 4.4). For the assignment of the other

resonances in the 1H NMR spectrum, 2D-NMR spectra (1H-1H COSY, 1H-13C HMQC and1H-1H NOESY) were necessary. The allyl-hydrogen atoms are magnetically inequivalent,

resulting in observation of five resonances of intensity 1H for each set of signals. Hence, in

total ten resonances owing to the hydrogen atoms of the allyl ligand are observed in the 1H

NMR spectrum. For assignment of these, in particular 1H-1H NOESY spectra were very

informative, because positive crosspeaks in these spectra indicate chemical exchange between

the same hydrogen atom in the two different chemical environments [153]. Because two sets

of signals are observed in the 1H and 13C NMR spectra of 5, it can be concluded that two

isomers are present in solution.

Figure 4.4 Selected regions from the 1H and 13C NMR spectra of 5 (250 MHz; DMSO-d6)

The 1H and 13C NMR spectra of 7 are very similar to the spectra of 5, but these show extra

resonances owing to the methyl propionate substituent. In addition, the resonances originating

from the NεH hydrogen atom are no longer observed in the 1H NMR spectrum of 7. In the 13C

NMR spectrum the influence of the two different chemical environments extends to the CH2

carbon atom neighbouring the carbonyl carbon atom of the methyl ester. In the 1H NMR

spectrum on the other hand, all resonances owing to the methyl propionate substituent,

including those for the OCH3 group, differ in both chemical environments.


47

Although the conformation of the two isomers observed in solution for 5 and 7 is not clear yet

at this stage, the observation of positive crosspeaks in the 1H-1H NOESY NMR spectra

demonstrated that the isomers undergo chemical exchange. This means that these two isomers

are in equilibrium in solution, thus making determination of the activation barrier for

interconversion by variable temperature 1H NMR measurements possible. In DMSO-d6 at 293

K, the ratio between the two isomers for 5 and 7 is 53 / 47 and 57 / 43, respectively. Although

the use of the simplified equation 4.1 shown below is strictly limited to the case of equally

populated two-site systems, it can also be adapted in the case of moderately biased unequally

populated two-site systems, if the error is made larger than 0.5 kJ mol-1 [170]. In the case of 5

and 7, the deviation from the 1/1 ratio is rather small, and therefore the use of this formula is

allowed with an error of 1.0 kJ mol-1.

∆G≠ = 1.914 × 10-2 TC [9.972 + log (TC / δν)] (in kJ mol-1) (4.1),

with TC = coalescence temperature (in K) and δν the chemical shift difference (in Hz) at the

low temperature limit.

Figure 4.5 Region between 6 and 9 ppm in the 1H NMR spectrum of 5 at six different

temperatures (400 MHz; DMSO-d6). * = impurity

Chapter 4

48

Especially the resonances owing to the imidazole CH hydrogen atoms in the 1H NMR spectra

of 5 and 7 are very suitable for determining the activation barrier for interconversion. The

region between 6 and 9 ppm of the 1H NMR spectra of 5 in DMSO-d6 at several temperatures

is shown in Figure 4.5. By using Equation 4.1 and taking TC = 328 K and δν = 73.0 Hz (for

the CεH signals), the activation barrier for interconversion of 5 is calculated to be 66.7 ± 1.0

kJ mol-1. Similar measurements made for 7 yield ∆G≠ = 66.9 ± 1.0 kJ mol-1, which is almost

identical to the value for 5.

Now that the activation barrier for interconversion is known, my interest concerned

elucidating the exact structure of the two observed isomers for 5 and 7 in solution. The

imidazole CH hydrogen atoms of the two isomers in the 1H NMR spectra of 5 and 7

particularly show a large chemical shift difference and the ratio between the two isomers

shows a pronounced solvent dependence. In solvents that are known to coordinate very well,

like DMSO and DMF, the ratio between both isomers is close to 1 : 1, whereas it is about 4 :

1 in MeOH. (To keep the presentation of the results and the discussion fluent, the report of the

exact ratio between the isomers of 5 and 7 in several solvents other than DMSO is postponed

until the end of this section). Therefore, displacement of the imidazole group by a solvent

molecule appeared to be a possible cause for the observation of two isomers in solution. To

verify this, the 1H and 13C NMR spectra of AsPh4[Mo(L-His)(CO)3] (9) were recorded in

DMSO-d6. It was expected that if displacement of the imidazole group would be the reason

for the observation of two isomers in solution, the 1H and 13C NMR spectra of 9 would also

display two sets of signals. However, the fact that the Mo-Nδ bond distance in the X-ray

crystal structures of both 5⋅⋅⋅⋅MeOH and 7⋅⋅⋅⋅2MeOH is not the longest Mo-histidine donor atom

bond length made cleavage and reformation of the Mo-Nδ bond in principle unlikely. Both the1H and 13C NMR spectra of 9 show only one set of resonances, which means that Mo-Nδ bond

cleavage and reformation as the cause for the observation of two isomers in solution is

completely ruled out.

Allyl rotation, as observed in the complex Mo(η-Cp)(η-allyl)(CO)2 and analogues (see

Chapter 3), has been suggested by Beck to be the rotary motion observed spectroscopically in

5 [156]. However, based on results from EHMO calculations [158], which were published

four years after Beck’s suggestion, this rotary motion as the reason for observing two isomers

in solution can be in principle discarded.


49

However, this suggestion by Beck can be put to an experimental test. If rotation of the allyl

ligand would be the cause for the observation of two isomers in solution, it is expected that

the 2-Me-allyl complexes 6 and 8 show a higher activation barrier for this process, simply due

to the higher mass of the 2-Me-allyl ligand. Although Beck reported the 1H NMR spectrum of

6 to consist of only one set of signals [157], the operating frequency of the NMR spectrometer

in that study was not reported. Because the operating frequency influences the appearance of a

NMR spectrum at a given temperature in the case of a chemically exchanging system (see

equation 4.1 and its δν dependence), the NMR spectroscopic behaviour of this complex and

its closely related analogue 8 was investigated in detail.

Selected regions from the 1H NMR and 13C NMR spectra of 6 in DMSO-d6 are displayed in

Figure 4.6. The 1H NMR spectra of 6 shows broadened signals at room temperature, which is

indicative for occurring coalescence. Also in the 13C NMR spectrum of 6 several resonances

are broad, in particular those owing to the carbonyl and 2-Me-allyl ligands. To explore

whether coalescence occurs at room temperature, samples of 6 and 8 were cooled down in

MeOH. Because 6 is insoluble in MeCN, only 8 was also cooled down in this solvent. In all

cases two sets of signals appeared at temperatures lower than 273K, which substantiates the

initial assumption of the occurring coalescence at room temperature.

Figure 4.6 Selected regions from the 1H and 13C NMR spectra of 6 (400 MHz; DMSO-d6)

With the objective to determine the activation barrier for interconversion, compounds 6 and 8

were subjected to variable temperature 1H NMR spectroscopic investigations in CD3OD. This

solvent was chosen because the ratio between the isomers for 6 and 8 is 55 / 45 in both

compounds, which is close to 1 / 1, thus allowing the adaptation of Equation 4.1. Selected

parts of the 1H NMR spectra of 6 at various temperatures are shown in Figure 4.7. By

Chapter 4

50

adapting Equation 4.1, the activation barrier for interconversion between the isomers of 6 was

determined to be 58.6 ± 1.0 kJ mol-1. Similar variable temperature 1H NMR measurements for

8 yield a ∆G≠ = of 58.5 ± 1.0 kJ mol-1, which is almost identical to the value for the parent

complex 6.

Figure 4.7 Region between 6 and 9 ppm in the 1H NMR spectrum of 6 at six different

temperatures (400 MHz; CD3OD)

Because the activation barrier for interconversion between the isomers of 6 and 8 is about 8 kJ

mol-1 lower than the activation barrier for interconversion between the isomers of 5 and 7,

allyl-rotation as the cause for the observation of two isomers in 5, 6, 7 and 8 can be

completely ruled out.

The imidazole CH hydrogen atoms of 8 in CD3CN at 238 K resonate at 8.59 / 7.82 ppm and

6.93 / 6.77 ppm. These chemical shifts for the two isomers of 8 are very similar to those for 7

in CD3CN (8.55 / 7.81 ppm and 6.90 / 6.74 ppm) at room temperature (293 K). In the solid

state structures, the 2-Me-allyl ligand occupies the position trans relative to the carboxylate

and the allyl ligand is in a trans-position relative to the Nδ. However, the 1H NMR spectra of

the allyl compounds at ambient temperature are very similar to the NMR spectra of the 2-Me-

allyl ligand complexes at lower temperature. This makes it very likely that the two observed

isomers in solution have the conformations shown in the solid state structures, viz. the allyl or


51

2-Me-allyl ligand trans to the carboxylate and trans to the Nδ. The large chemical shift

difference between the resonances for the imidazole hydrogen atoms of the two isomers can

be explained in this way by the trans-influence of the allyl or 2-Me-allyl ligand. When the

allyl or 2-Me-allyl ligand is trans to the Nδ, the Mo-Nδ bond is stronger and shorter and, thus,

the imidazole-hydrogen atoms resonate downfield compared to the isomer with a carbonyl

ligand in a trans-position relative to the Nδ. The most probable rotary motion to interconvert

these species involves a restricted trigonal twist around the Mo-N(amine) axis, as observed

previously for similar types of pseudo-octahedral compounds which bear the Mo(η-

allyl)(CO)2 moiety [159, 160, 164, 171]. As a matter of fact, the activation barriers for

interconversion between the isomers of 5-8 are in the range reported for trigonal twists of

analogous pseudo-octahedral Mo complexes [160, 164].

From now on, the isomer with the allyl or 2-Me-allyl ligand trans relative to the histidine-Nδ

atom will be called a and the isomer with the allyl or 2-Me-allyl ligand trans relative to the

carboxylate oxygen atom will be denoted b, as visualised in Scheme 4.2.

Scheme 4.2 Conformation of the observed isomers a and b in solution for 5-8. Note that

the terminal C atoms of the allyl ligand are directed towards the CO ligands

N

N

H2N

O O

MoOC

OC

R

R'

R N

N

H2N

O O

MoOC

R'

CO

a b

For completeness, the ratio between the isomers a and b for the compounds 5-8 as well as the

determined activation barriers for a / b interconversion is summarised in Table 4.3. By using

the formula ∆G0 = -RTlnK, the ratio between the isomers is also converted into the energy

difference between the two species. The absolute value is reported, which means that the most

abundant species has the lower G0.

Chapter 4

52

Table 4.3 Summary of thermodynamic and kinetic parameters obtained from 1H NMR

spectroscopic investigations on 5-8 in several solvents

Compound Solvent a / b ratio ∆G0[a, b] ∆G≠[a]

5 DMSO-d6 57 / 43[c] 0.7 66.7 ± 1.0

5 MeOH-d4 79 / 21[c] 3.3

6 MeOH-d4 55 / 45[d] 0.4 58.6 ± 1.0

7 DMSO-d6 53 / 47[c] 0.3 66.9 ± 1.0

7 MeCN-d3 53 / 47[c] 0.3 66.7 ± 1.0

8 MeOH-d4 55 / 45[d] 0.4 58.5 ± 1.0

8 MeCN-d3 29 / 71[d] 1.8

[a] in kJ mol-1; [b] absolute value; the most abundant isomer has the lowest G0; [c] at 300 K; [d]

at 243 K

From this table, it can be concluded that the thermodynamic energy difference between the

isomers a and b of 5 and 7 in identical solvents is very similar and the same is also observed

for the isomers a and b of 6 and 8 in MeOH-d4. Apparently, introduction of a methyl

propionate substituent on the Nε atom does not influence the thermodynamic energy

difference between the isomers a and b for the allyl and 2-Me-allyl complexes significantly.

Both, the allyl and 2-Me-allyl complexes show a solvent dependence. In the polar protic

solvent MeOH-d4, the ratio between the isomers is shifted considerably in favour of the a

conformer in comparison to the polar aprotic solvents MeCN-d3 and DMSO-d6.

4.5 Results from Density Functional Theory calculations

In order to ensure that the chosen theoretical methods are capable of correctly describing the

energetics of molybdenum allyl dicarbonyl complexes, the relative energies of the endo and

exo isomers of Mo(η-Cp)(CO)2(η-allyl) have been computed (see Scheme 3.1). It is known

from NMR studies that both conformers exist as an equilibrium mixture at room temperature

(endo:exo ratio 1:4, which corresponds to ∆G0 = 3.3 kJ mol-1) [154]. At the

BPW91/II"//BP86/SDD level, the endo form is computed 1.6 kJ mol-1 less stable than the exo


53

isomer, which corresponds, according to a Boltzmann distribution at room temperature, to a

1:2 ratio, which is in good qualitative agreement with experimental findings. The difference

between the calculated and computed thermodynamic energy difference of 1.7 kJ mol-1 is

within the accuracy of the calculations. At the BPW91/II//BP86/ECP1 level, which makes use

of larger basis sets on the ligands (cf. reference [172]), an almost identical energy difference

between the exo and endo isomer is obtained, namely 1.7 kJ mol-1. Because the results appear

not to depend significantly on the employed basis set, the smaller basis set

(BPW91/II"//BP86/SDD) has been used for the following calculations in order to save CPU-

time.

Subsequently, the energies of all six regio-isomers of Mo(L-His)(CO)2(η-allyl) (5) and Mo(L-

His)(CO)2(η-2-Me-allyl) (6) were computed. These regio-isomers differ in the relative

orientation of the histidinate and the (η-allyl)(CO)2 moieties in the pseudo-octahedral

arrangement around the Mo-ion, as well as in the orientation of the allyl or 2-Me-allyl ligand

(endo or exo). The resulting structures are depicted schematically in Scheme 4.3.

Scheme 4.3 Schematic representation of the DFT-optimised minima for 5 and 6.

In parentheses: relative energies in kJ mol-1

R

N NH2O

OCOC

MoN NH2O

COOCMo

N NH2O

COMo

OC

N NH2O

OCOC

Mo

N NH2O

COOCMo

N NH2O

COMo

OC

R RR

R R

N NH2 = His

5a (0.0)6a (0.0)

5c (6.3)6c (5.9)

5b (0.4)6b (0.8)

5d (30.5)6d (24.6)

5f (25.9)6f (18.8)

5e (33.9)6e (26.3)

O

5 R = H6 R = Me

Chapter 4

54

As shown in Scheme 4.3, the endo isomers 5d-f are much higher in energy than the exo

minima, of which two, 5a and 5b, are quite close in energy (within less than 1 kJ mol-1), and

the third, 5c, is significantly less stable. These data support the identification of 5a and 5b as

the two isomers that are observed in solution.

Essentially the same results are obtained for Mo(L-His)(η-2-Me-allyl)(CO)2 (6) (see Scheme

4.3). Isomers 6a and 6b are computed to be most stable and are, thus, most likely the species

found in solution. A minor inconsistency between the experimental findings and the results

from the calculations is that the latter find the energy difference between 6a and 6b higher

than the energy difference between 5a and 5b. This is in contrast with the NMR spectroscopic

data, which show that isomer b of 6 is stabilised with respect to isomer b of 5 in identical

solvents. However, it must be noted that the differences in energy between 6a and 6b are

rather small and the theoretical level chosen is certainly not accurate enough to assess energy

differences of a few tenths of a kJ mol-1, but it should be adequate for the more qualitative

purposes of this computational study.

As derived from variable temperature 1H NMR spectroscopic investigations, the two isomers

observed for 5 and 6 interconvert with energy barriers of 66.7 ± 1.0 kJ mol-1 (for 5 in DMSO-

d6) and 58.6 ±1.0 kJ mol-1 (for 6 in MeOH-d4). In an attempt to rationalise the experimental

findings and to provide further support for the structural assignments, a plausible path was

computed for this interconversion, namely the pseudo-rotation of the histidinate and the (η-

allyl)(CO)2 fragments through pseudo-trigonal prismatic transition states (path A, upper part

of Scheme 4.4). For 5, the computed barrier via this path of 66.1 kJ mol-1 is virtually identical

to the experimental value (66.7 kJ 7 ± 1.0 mol-1 in DMSO-d6), but a substantially higher value

of 75.3 kJ mol-1 is obtained for the 2-Me-allyl derivative 6. The higher barrier for 6 in

comparison to 5 on path A is due to the increased steric interference of the methyl group as it

passes the histidinate NH2 moiety in TS6ab (closest NH...HC contact 2.08 Å vs. 2.22 Å in

TS5ab). It is in sharp contrast with the experimental data, which show a lower barrier for 6

than for 5. Thus, a simple pseudo-rotation via path A cannot be reconciled with the

experimental observations. A similar path with two consecutive pseudo-rotations according to

the sequence a → c → b was computed to be less favourable for 5 (highest point TS5bc, 69

kJ mol-1, not shown in Scheme 4.4) and even more so for 6 (TS6bc, 95 kJ mol-1).


55

Scheme 4.4 Schematic paths A and B for interconversion between isomers a and b of 5 and

6. The highest point on each path is highlighted. In parentheses: DFT-

computed relative energies in kJ mol-1

RN NH2O

OCOC

Mo

N NH2O

COMo

OC

N NH2O

COOC

MoRR

R

N NH2O

OCOC

Mo

N NH2O

CO

Mo

OCR

N NH2O

OCOC

Mo

R

R

N NH2O

COOC

Mo

R

N NH2O

CO

Mo

OC

‡

‡

‡

‡

path A

path B

5a (0.0)6a (0.0)

TS5ab (66.1)TS6ab (75.3)

5b (0.4)6b (0.8)

5d (30.5)6d (24.6)

5e (33.9)6e (26.3)

TS5ad (51.9)TS6ad (43.9)

TS5de (65.3)TS6de (59.8)

TS5be (38.5)TS6be (33.9)

Closer inspection of all isomers of 5 and 6 reveals that the endo isomers of the latter are

generally lower in energy than those of the former. It was assumed that the lower barrier for 6

might involve some of these endo isomers as intermediates during the interconversion process

and, indeed, a suitable reaction path was found computationally (path B, lower part in Scheme

4.4). This path consists of 1) rotation of the coordinated allyl group affording the endo species

d, 2) pseudo-rotation via a quasi-trigonal prismatic transition structure affording e, and 3)

rotation of the allyl moiety yielding the exo isomer b. In this sequence, the highest total

activation energy is required for step 2), which is therefore the rate-determining step. The

corresponding barrier of 65.3 kJ mol-1 in the case of 5 is similar to, and in fact even slightly

Chapter 4

56

lower than, that along path A. More importantly, for 6 the highest point on path B, 59.8 kJ

mol-1, is lower than that for 5 by circa 6 kJ mol-1. This finding is now fully consistent with the

experimental data, which find the activation barrier for interconversion for 6 to be 58.6 ± 1.0

kJ mol-1 in MeOH-d4.

Path B thus represents a viable reaction sequence consistent with the experimental results,

substantiating the assignment of the two isomers for each 5 and 6. The interconversion

between these isomers is indicated to be a complex process, involving possibly, and for 6

most likely, at least two intermediates with an endo orientation of the allyl group.

It should be noted that it cannot be fully excluded that other paths may also play a role, for

instance involving the remaining isomers c and f. It is unlikely, however, that the general

conclusion regarding the highest barriers for 5 and 6 would be invalidated. On paths A and B,

the allyl groups are always bonded in a η3- fashion, but another possibility involves η1- bound

allyl intermediates. In order to test if such species with the allyl ligand bound in an η1-mode

could be competitive, a representative isomer (5g) has been optimised. The resulting structure

is characterised by an α-agostic Mo-H interaction [173] (see Scheme 4.5) and is 130 kJ mol-1

higher in energy than 5a. It is conceivable that solvents with strongly σ-donating properties

might stabilise an η1-form by coordination to the Mo atom at the empty site that is occupied

by the α-agostic H atom in 5g. Such a scenario, however, appears to be unlikely on the basis

of the thermodynamic stabilisation brought about by a η3-bound allyl ligand with respect to

the η1-binding mode and therefore this possibility has not been further explored

computationally.

Scheme 4.5 Conformation of a hypothetical isomer of 5, with an η1-bound allyl ligand

N NH2O

OCOC

Mo

5g

H


57

4.6 Electrochemical investigations

In order to investigate whether both the allyl and 2-Me-allyl compounds could serve as a

marker based on electrochemical detection, their electrochemical behaviour was investigated.

Because the parent complexes 5 and 6 are insoluble in MeCN and CH2Cl2 (the most common

solvents used for electrochemistry), only the electrochemical properties of 7 and 8 were

studied. In section 4.2, it has been demonstrated that the properties of 7 and 8 are very similar

to those of the parent complexes.

The cyclic voltammogram (CV) of 7 in MeCN (scan rate of 0.1 V s-1, 25° C; 0.1 M NBu4PF6

as supporting electrolyte) shows a wave with a reversible appearance at E1/2 = +86 mV vs.

Fc/Fc+. This value is in the range reported for several analogous pseudo-octahedral Mo-

complexes [174, 175]. Controlled potential coulometry (at +0.5 V vs. Fc/Fc+) reveals that the

wave arises from an one-electron oxidation of 7, but this oxidation does not proceed via an

uncomplicated reversible single-electron process. At low scan rates, the peak separation

increases with increasing scan rates from 85 mV at 0.025 V s-1 to 140 mV at 0.8 V s-1, but at

even faster scan rates (up to the highest value measured of 25 V s-1) it remains almost constant

(±10 mV). It should be noted that electrodes of different sizes were used to minimise peak

shifts resulting from the limited electrical conductance of the solution. Within the total range

of scan rates there was only a small shift of E½ (that is (Ep,a - Ep,c)/2) of approximately 0.015

V towards positive potentials with increasing scan rate.

This behaviour of the cyclic voltammograms cannot be explained by electrode kinetics, but it

can be accommodated in a “square scheme” shown in Scheme 4.6, taking into account the

presence of two isomers both in the oxidised and reduced form of 7.

In such schemes, limited shifts of the peak potentials are possible at lower scan rates, the size

of which is determined by the difference between the two reduction potentials. At higher scan

rates the isomerisation processes are "frozen" during the time of a scan and consequently the

measured cyclic voltammogram corresponds to those of two individual uncoupled species.

Thus, multiple peaks on the anodic or cathodic waves are expected at high scan rates. These

were, however, not found, possibly because the difference between the two reduction

potentials is relatively small (≤ 0.1 V) and, in addition, the peaks tend to broaden at higher

Chapter 4

58

scan rates because of capacitive currents. In order to detect multiple peaks at lower scan rates,

voltammetric experiments were performed at -30°C and –60°C in propionitrile. In the CV

broad shoulders on the anodic and cathodic peaks appeared but the presence of two species

could be much better detected by square wave voltammetry (SWV). The square wave

voltammograms of 7 at +25, -20, -40 and -60° C are shown in Figure 4.8.

Scheme 4.6 Square scheme showing the redox reactions and rearrangements of 7

N

N

H2N

O O

MoOC

OC

R

N

N

H2N

O O

MoOC

CO

R

N

N

H2N

O O

MoOC

OC

R

N

N

H2N

O O

MoOC

CO

R ++

∆G02

∆G04

∆G03 ∆G0

1

K4

K2

Figure 4.8 Square wave voltammograms of 7 at four different temperatures (10-3 M;

EtCN; 0.1 M NBu4PF6)


59

At –60° C, the anodic and cathodic scans are fully symmetric, which shows that the

equilibrium composition of the reduced form (the starting material 7) is maintained after the

oxidation for the time period of equilibration at + 0.6 V and during the scan, i.e. for at least a

few seconds. Because the equilibrium constants for the oxidised and reduced forms are quite

different (vide infra), the isomerisation reaction must be correspondingly slow at –60°C. At –

20°C, anodic and cathodic scans are different and it is observed that oxidised 7 exhibits a

preference for the state with the lower redox potential. At even higher temperatures, two

components are no longer discernible by SWV and a situation analogous to coalescence in

NMR experiments is reached. Curve fitting (COOL kinetic software) to the model of

reversible oxidation of two independent species with equal diffusion coefficients was

performed. The results from this fitting procedure at three different temperatures are displayed

in Table 4.4 and the reactions to which the ∆G0 and K values correspond are shown in

Scheme 4.6.

Table 4.4 Thermodynamic parameters obtained from the fitting procedure on the square

wave voltammograms of 7 at –20, -40 and –60°C[a]

T[b] E1(1/2)[c] E2(1/2)[c] ratio ∆G10 [d] ∆G2

0 [d] K2 ∆G30 [d] ∆G4

0 [d] K4

- 60 0.015 0.100 40 / 60 - 40.1 -0.72 1.5 +48.3 -7.5 68

- 40 0.022 0.103 43 / 57 - 40.7 -0.55 1.3 +48.5 -7.3 43

- 20 0.030 0.109 45 / 55 - 41.5 -0.42 1.2 +49.1 -7.2 31

[a] COOL kinetic software used; processes to which the K and ∆G0 values correspond are

shown in Scheme 4.6; [b] in °C; [c] in V, vs. Fc/Fc+; [e] in kJ mol-1

At –60°C the values E1(1/2) = +15 mV and E2(1/2) = +100 mV vs. Fc/Fc+ were obtained, with

relative contributions of 40% and 60% for the species with E1(1/2) and E2(1/2), respectively.

These values were independent of the scan frequency (20 - 80 Hz, step increment 1 mV) and

of the direction of the scan. The ratio between the species at the three temperatures

corresponds well to the 1H NMR data (59 / 41 at –30°C in MeCN), indicating that the same

two species are under investigation by the two methods.

Chapter 4

60

It should be noted that in reduced 7 the form with the higher redox potential is slightly

favoured, whereas the oxidised form exhibits a strong preference for the species with the

lower redox potential. The temperature dependence of ∆G20 and ∆G4

0 is only weak and

analysis in terms of the van 't Hoff isochore yields standard reaction enthalpies ∆H20 = -2.5 kJ

mol-1 and ∆H40 = -8.3 kJ mol-1 for the reduced and the oxidised forms of 7, respectively.

Small values are expected because the same types of bonds exist in all species. The difference

of about 6 kJ mol-1 might be attributed to increased solvation of the oxidised species as a

consequence of their positive charge. The standard reaction entropies (∆S) for the reduced and

oxidised form are -8.2 J mol-1 K-1 and -4.3 J mol-1 K-1, respectively.

The square wave voltammogram of 8 (1× 10-3 M, propionitrile, 0.2 M TBAPF6) at 25 oC

exhibits a reversible wave with a half-width of 100 mV and a peak position of -0.025 V vs

Fc/Fc+. Controlled Potential Coulometry (at 0.6 V vs Fc/Fc+) shows the wave to arise from a

one-electron oxidation of 8. The voltammograms before and after oxidation are identical,

which demonstrates the stability of 8 in the oxidised and reduced form.

Square wave voltammograms of 8 at four different temperatures are shown in Figure 4.9. On

lowering the temperature gradually to -40 oC, no significant change occurs. The peak

positions shift by 0.35 mV / oC in the direction of negative potentials, but at a fixed

temperature both anodic and cathodic scans yield narrow peaks (half width 100 ± 4 mV) at

almost the same peak position (± 5 mV). Therefore, at -40 oC there is no evidence from the

voltammograms for the presence of more than one species.

Voltammograms recorded at -55 oC and at higher scan rates (60 -100 Hz) exhibit a shoulder

on the anodic side which is more developed on anodic than on cathodic waves. At -75 oC, the

anodic and cathodic voltammograms are quite symmetric, i.e. the shoulders on both scans are

about equally pronounced. This demonstrates that only at -75 oC the equilibrium composition

of the reduced form is maintained after oxidation during the time required for a cathodic scan.

The same behaviour for 7 was observed already at -40 oC. A closer analysis (by the curve-

fitting procedure, as described for 7) shows that even at -75 oC the equilibrium-composition of

8 is not perfectly "frozen", as can be seen from the scan rate and scan direction dependence of

the apparent composition shown in Table 4.5 (this dependence was not observed with 7).


61


EtCN; 0.2 M NBu4PF6).

Table 4.5 Results from the curve fitting procedure on square wave voltammograms of 8

at various scan frequencies at –75 °C[a]

Scan[b] anodic scans cathodic scans

frequency conc. ratio[c] species ratio[d] conc. ratio[c] species ratio[d]

10 0.55 35 / 65 0.45 31 / 69

20 0.59 37 / 53 0.45 31 / 69

50 0.71 42 / 58 0.53 35 / 65

100 0.71 42 / 58 0.50 33 / 67

[a] COOL kinetic software used; [b] in Hz; [c] Concentration ratio values obtained from curve fitting;[d] Relative contributions calculated from the concentration ratio.

Chapter 4

62

The most favourable conditions for application of the model of two independent species are

anodic scans with high scan rates, and for these the concentration ratio approaches a constant

value of 0.71. This value is assigned as the equilibrium constant K2 of the isomers in their

reduced form at -75 oC, i.e. ∆G20 = +0.57 kJ mol-1 (see Scheme 4.6 for representation of the

isomers, but the allyl ligand is in this case a 2-Me-allyl ligand). Together with the reduction

potential difference for the two isomers of 80 mV found in the curve fitting procedure this

yields ∆G40 = -8.3 kJ mol-1 (and K4 = 150) for the two isomers in their oxidised form, i.e.

with 8 there is a preference for the species with the lower reduction potential in their reduced

and (more pronounced) oxidised form.

A small inconsistency is that a ratio between the a / b isomers of 31 / 69 is obtained from 1H

NMR spectroscopic investigations at –30°C in MeCN, which differs slightly from the 42 / 58

ratio obtained from electrochemical investigations at –75° C. However, it must be noted that

EtCN and MeCN are not identical solvents, and it is expected that at lower temperatures the

ratio will be shifted slightly more towards the thermodynamically favoured b isomer.

In a control experiment cathodic scans in a solution of 8+ were performed in the same

temperature range (after bulk electrolysis of 8 at -10 oC). Only a single narrow peak was

detected (half width around 100 mV at all temperatures) with the peak position shifting

steadily by 0.35 mV / oC. This is in agreement with the high value of K4, i.e. the oxidised

form consists of only one species. Any other isomer of 8 possibly present must therefore have

either a very similar reduction potential (difference < 30 mV) or must be facile

interconvertible into the measured form even at -75 oC (half-life of interconversion less than a

few seconds).

In summary, by performing square wave voltammetric experiments at low temperatures it has

been demonstrated that 7+ and 8+ display, like their neutral precursors 7 and 8, fluxional

behaviour in solution. The electrochemical properties of several other pseudo-octahedral Mo

complexes containing the (η-2-R-allyl)(CO)2 core (R= H or Me) have been investigated [174,

175], but this is the first report of the observation of two isomers in the oxidised form. Both 7+

and 8+ were shown to exhibit a strong preference for the conformation in which the allyl or 2-

Me-allyl ligand occupies a trans-position relative to the carboxylate oxygen atom. The energy

difference between the thermodynamically preferred conformation and the one in which the


63

Nδ is trans with respect to the allyl or the 2-Me-allyl ligand was determined to be 7.5 kJ mol-1

for 7+ and 8.3 kJ mol-1 for 8+. These large energy-differences imply that these compounds

exist almost exclusively in the isomer with the allyl or 2-Me-allyl ligand trans relative to the

carboxylate oxygen atom at temperatures at which exchange between the two isomers takes

place (i.e. higher than –40° C in the case of 7+ and higher than –70° C for 8+). It should be

noted that these measurements do not give an indication whether the allyl and 2-Me-allyl

ligands in 7+ and 8+ still have their terminal carbon atoms oriented towards the carbonyl

ligands, or whether they have undergone a 180° rotation upon oxidation.

Below the infrared and electronic spectroscopic properties of 5-8, 7+ and 8+ are presented and

compared mutually.

4.7 Electronic and infrared spectroscopic investigations

The electrochemical investigations in the previous section demonstrated good stability of the

one electron oxidised complexes 7+ and 8+. Infrared, UV-Vis and circular dichroism spectra

of these cationic compounds were obtained by generating these electrochemically in a

coulometry cuvet (for the UV-Vis and CD spectra) or in an OTTLE-cell (Optically

Transparent Thin Layer Electrochemical-cell; for the infrared spectra). In this section, first the

infrared spectra of 5-8, both in solution and as a KBr pellet, will be discussed, followed by

comparison of the solution infrared spectra with those of 7+ and 8+. Thereafter, the UV-Vis

and CD spectra of 5-8 will be presented, followed by comparison of these UV-Vis and CD

spectra with those of the oxidised complexes 7+ and 8+.

Carbonyl stretching frequencies for 5-8 in solution and as a KBr pellet are presented in Table

4.6. Although only two carbonyl ligands are present in 5 and 6, three carbonyl stretching

vibrations in the infrared spectra of these compounds as a KBr pellet are observed, which is

likely caused by lattice effects. From 1H NMR spectroscopic investigations it is known that

the a / b ratio for 5 and 6 is different in MeOH (79 / 21 and 55 / 45 respectively), but this is

not reflected in the position of the two carbonyl stretching vibrations. Also in the case of 7

and 8 it can be assumed that the a / b ratio for these two compounds in CH2Cl2 is not

Chapter 4

64

identical, although the exact ratio between the isomers has not been determined in this

solvent. Because the νCO vibrations for the allyl and 2-Me-allyl compounds are located at

similar wavenumbers, it appears that the isomers a and b display similar CO stretching

vibrations. However, these νCO vibrations for the isomers a and b cannot be identical, because

the carbonyl ligands in both isomers are not symmetry-related.

Table 4.6 Selected vibrational data for 5-8.

Complex νCO (KBr)[a] Solvent νCO (solution)[a]

5 1805, 1835, 1932 MeOH 1847, 1939

6 1808, 1842, 1928[b] MeOH 1848, 1939

7 1824, 1927 CH2Cl2 1835, 1933

8 1827, 1929 CH2Cl2 1835, 1932

[a] in cm-1; [b] unit cell contains two crystallographically independent molecules

The changes in the infrared spectrum observed during oxidation of 8 in CH2Cl2 are depicted

in Figure 4.10. Upon oxidation the νCO vibrations shift from 1832 and 1932 cm-1 to 2002 and

2062 cm-1. This shift to higher wavenumbers is consistent with a decrease in the electron-

availability of the molybdenum atom in 8+ with respect to 8.

Figure 4.10 Infrared spectroscopic changes monitored during the transition from 8 to 8+ (in

CH2Cl2 by using the OTTLE-cell)


65

It is observed that the absorbance-maximum of the νCO vibrations in 8+ is only slightly lower

than that of the corresponding vibrations in 8. Usually, the intensity of the carbonyl stretching

vibrations becomes significantly lower upon increase of the metal ion’s valency. This is due

to the energy-lowering of the metal d-orbitals upon increase of the oxidation state, in this way

resulting in decreased MO-energy stabilisation brought about from the interaction of metal d-

orbitals of the appropriate symmetry with the energetically higher carbonyl π*-orbitals. The

reason why carbonyl ligands display infrared vibrations of high intensity is not owing to the

high polarity of the bonds involved, as is e.g. in the case of the vibrations of the

hexafluorophosphate anion. As a matter of fact, the polarity of the bonds in carbon monoxide

is low, both in the free (uncomplexed) form and in metal carbonyl compounds [176]. The high

intensity of metal-bound carbonyl vibrations is a consequence of the retrodative bonding, or

backbonding. The filled π-orbitals of a coordinated carbonyl ligand are concentrated on the

oxygen atom, whereas the vacant π*-orbitals have the largest coefficients on the carbon atom

[176], and these π*-orbitals overlap appreciably with the metal d-orbitals. Upon excitation of

the carbonyl ligand in its first excited vibrational state, the C-O bond elongates, resulting in an

energy-lowering of the carbonyl π*-orbitals because the magnitude of the antibonding

interaction between the carbon atom and oxygen atom is reduced. This leads to an increase of

the amount of metal to carbonyl π-donation and, thus, to charge-separation. This charge

separation gives rise to an oscillating dipole, resulting in intense infrared active vibrations.

The rearrangement of the orbitals upon stretching of a carbonyl C-O bond has been termed

orbital following [155, 177, 178]. When the metal d-orbitals are lower in energy as a

consequence of an increase of the metal ion’s oxidation state, the amount of orbital following

will be less, which results in a decreased intensity of the νCO stretching vibration.

When the carbonyl stretching vibrations of 8 and 8+ are examined more carefully, it is

observed that the vibrations of 8 are broader than those of 8+ and, thus, upon oxidation the

intensity of the νCO vibrations indeed decreases significantly, although the absorbance-

maximum decreases only slightly.

From electrochemical investigations, it has been derived that 8+ exists predominantly in the

isomer with the 2-Me-allyl ligand in a trans-position relative to the carboxylate oxygen atom

(K = 160 implies less than 1% of the other isomer), whereas both isomers are present in a

ratio close to 1 / 1 in the neutral form. Because the νCO vibrations owing to each isomer are

expected to be slightly different, the broader νCO vibrations of 8 compared to those of 8+

might very well be due to overlap of the vibrations from both isomers.

Chapter 4

66

As can be observed in Figure 4.10, also the νC=O vibration of the coordinated carboxylate

shifts from 1642 cm-1 to 1689 cm-1 upon oxidation of 8 to 8+, whereas the νC=O vibration of

the methyl ester is unaffected. The shift of the νC=O vibration to higher energy indicates that

the Mo-O bond becomes much stronger upon oxidation, in this way increasing the electron

donation from the carboxylate group.

The infrared spectral changes upon oxidation of 7 are almost identical to those observed for

the transition from 8 to 8+. The νCO vibrations at 1835 and 1933 cm-1 in 7 shift to 2006 and

2065 in 7+, while at the same time becoming much sharper, and also the νC=O vibration of the

carboxylate moiety shifts from 1642 to 1687 cm-1. Thus, from the results from infrared

spectroscopic investigations, it appears that the changes occurring upon oxidation of 7 and 8

are very similar.

Selected UV-Vis spectroscopic data for the complexes 5-8 in several solvents is listed in

Table 4.7 and, as representative examples, selected regions from the UV-Vis spectra of 7 and

8 in MeCN are depicted in Figure 4.11.

Table 4.7 Summary of UV-Vis spectroscopic data for the neutral complexes

Complex Solvent λmax[a] ε(λmax)[b] a / b ratio

5 MeOH 385 8.2 79 / 21

7 MeOH 385 8.2 79 / 21

7 MeCN 375 8.4 53 / 47

8 MeOH 366 9.4 55 / 45

8 MeCN 353 10.8 21 / 79

[a] in nm; [b] in 102 M-1 cm-1

In addition to the maxima around 280 nm owing to the π-π* transition of the imidazole-ring,

the complexes display a MLCT band around 370 nm. The general appearance of the UV-Vis

spectra of the allyl and 2-Me-allyl complexes is very similar and these spectra of 7 and 8 only

differ significantly in the position of λmax. For both types of compounds it appears that the

position of λmax for this MLCT transition depends on the solvent and, thus, also on the ratio

between the isomers. The data summarised in Table 4.7 show that the absorbance maximum


67

of this MLCT transition shifts to lower energy when the abundance of isomer a increases.

Isomer a has the conformation in which the Nδ nitrogen atom is in a trans position relative to

the allyl or 2-Me-allyl ligand, whereas the Nδ nitrogen atom of isomer b is in a trans position

relative to a carbonyl ligand. Both the allyl (or 2-Me-allyl) ligand and the carbonyls are

ligands that only form stable complexes when the binding of these is stabilised by

backbonding. It is expected that in either conformational isomer the metal d-orbitals are

orientated suitably for d-π* transitions (MLCT), because backbonding takes place in all three

directions (x, y, z; chosen arbitrarily). Apparently, this MLCT transition requires less energy

when the imidazole is trans relative to the allyl or 2-Me-allyl ligand. Furthermore, it also

seems that the molar extinction coefficient of the MLCT absorbtion band is higher when the

λmax of this band is situated at higher energy, but this effect might also be due to the fact that

the intraligand π-π* transitions still have a small intensity in that region. The exact identity of

the MLCT transition in the UV-Vis spectra of 7 and 8 is not clear, because the imidazole

moiety, the carbonyl ligands and the allyl (or 2-Me-allyl) ligand each have π*-orbitals that are

relatively low in energy and can act as π-acceptors.

Figure 4.11 Selected regions from the UV-Vis spectra of 7 and 8 in MeCN

Circular dichroism spectra of complexes 7 and 8 were recorded in several solvents, the results

of which are summarised in Table 4.8. As representative examples, the CD spectra of 7 and 8

in CH3CN are depicted in Figure 4.12.

Chapter 4

68

Table 4.8 Circular dichroism spectroscopic data for 7 and 8 in several solvents.

Complex Solvent λmax[a] θ (λmax)[b]

× 10-3

λmax[a] θ (λmax)[b]

× 10-3

a / b ratio

7 MeOH 322 -3.9 413 -6.7 79 / 21

7 MeCN 320 -7.7 409 -5.0 53 / 47

7 DMSO 320 -8.5 406 -4.7 53 / 47

8 MeOH 324 -8.7 411 -5.0 55 / 45

8 MeCN 323 -12.2 410 -1.7 21 / 79

8 DMSO 320 -12.6 408 -1.4 n. d.[c]

[a] in nm; [b] units of θ are 102 × ° M-1 cm-1; [c] n. d. = not determined.

Figure 4.12 Selected regions from the circular dichroism spectra of 7 and 8 in MeCN.

Complexes 7 and 8 exhibit two negative Cotton effects in MeOH, MeCN and DMSO, with

the maxima located around 320 nm and 410 nm. The molecular ellipticity at the two maxima

varies with the solvent, and thus, also with the ratio between the isomers a and b. For

example, in the CD spectrum of 7 in MeOH, the Cotton effect at 320 nm has a lower negative

maximum than the Cotton effect at the higher wavelength. In contrast, the situation is

opposite in the CD spectra of 7 in MeCN and DMSO, in which the Cotton effect at the lower

wavelength displays a much higher negative maximum compared to the negative maximum at

around 410 nm. From the CD spectroscopic data for 7 and 8 in Table 4.12, a general trend can

be derived. When the abundance of isomer a increases, the negative maximum of the Cotton


69

effect at around 320 nm decreases, whereas that of the Cotton effect at around 410 nm

increases. It is not clear which electronic transitions give rise to the Cotton effects in the CD

spectra of 7 and 8. By comparing the CD spectra of 7 and 8 with their UV-Vis spectra, it can

be concluded that the optically active transitions in the CD spectra are not situated at the same

wavelength as the MLCT transitions in the UV-Vis spectra of these complexes. However, it is

likely that the bands in the UV-Vis spectra corresponding to the Cotton effects in the CD-

spectrum are obscured by the more intense MLCT absorbtion bands.

The changes in the UV-Vis spectrum during the oxidation of both 7 and 8 in MeCN are

depicted in Figure 4.13. Upon oxidation of 7, the MLCT transition at 375 nm disappears

gradually and this is accompanied by the formation of a new absorbtion band at 348 nm,

which has a molar extinction coefficient around 1370 M-1 cm-1. Furthermore, a new band

appears at higher wavelengths with λmax = 622 nm (ε around 70 M-1 cm-1). The molar

extinction coefficients after the oxidation cannot be determined very reliably because of

solvent-effervescence as a result of degassing the solution with argon. Two isobestic points

are observed at 383 nm and 456 nm, which demonstrates the reversibility of the

electrochemical process. The band with λmax = 622 nm is assigned as a d-d transition of a

lower lying orbital with d-character to the half-filled orbital with d-character. The compound

under investigation is not a pure Werner-type coordination complex, but also contains

organometallic fragments and therefore the d-orbitals mix to a large extent with π and π*

orbitals of the organometallic ligands, resulting in orbitals that do not have pure d-character.

Figure 4.13 UV-Vis spectroscopic changes monitored during the oxidation of 7 to 7+ and 8

to 8+ in MeCN at –20° C.

Chapter 4

70

Upon oxidation of 8 to 8+, a new absorbtion band appears in the UV-Vis spectrum at around

330-335 nm, which is considerably stronger than the MLCT-band present in the UV-Vis

spectrum of 8. However, the exact position of this new band as well as its molar extinction

coefficient cannot be determined reliably because it overlaps with the π-π* transition of the

imidazole group. Furthermore, a new band appears after oxidation with λmax = 546 nm, with a

molar extinction coefficient around 64 M-1 cm-1. Like in the UV-Vis spectrum of 7+, this new

band is also assigned as the transition from a lower lying orbital with d-character to the orbital

with d-character that is half-filled. Also in this case, two isobestic points are observed, located

at 379 nm and 437 nm.

The UV-Vis spectra of 7 and 8 differ in the position of the λmax of the MLCT transition,

which is situated approximately 20 nm lower for 8 than for 7. Also after the oxidation, the

λmax of the newly formed band around 340 nm is located about 15-20 nm lower in the UV-Vis

spectrum of 8+ compared to the spectrum of 7+. From electrochemical investigations it is

known that both complex cations exist almost exclusively in the isomer with the allyl ligand

or 2-Me-allyl ligand in a trans position relative to the carboxylate oxygen atom. However, the

fact that the transition with d-d character occurs at lower wavelength for 8+ compared to 7+

(λmax = 546 nm and λmax = 622 nm, respectively) indicates that the energy levels of the

orbitals with d-character are not identical in both compounds.

Because the CD spectrometer is a single-beam apparatus, spectra during the oxidation of 7

and 8 could not be obtained because of the long spectral acquisition times. CD spectra of 7+

and 8+ were recorded after the complexes 7 and 8 were quantitatively transformed into their

oxidised derivatives by controlled potential coulometry and the resulting spectra are shown in

Figure 4.14.

The two negative Cotton effects at around 320 nm and 410 nm observed for the neutral

compounds 7 and 8 are no longer present after the oxidation. Instead a much more intense

negative Cotton effect is observed with its maximum situated at 352 nm (for 7+) and 345 nm

(for 8+). Interestingly, the negative maximum in the CD spectrum of 7+ is located at nearly the

same wavelength as the intense band in the UV-Vis spectrum of this compound (352 nm in

the CD-spectrum and 348 nm in the UV-Vis). For 8+, however, the maxima observed in the

UV-Vis and CD spectra differ to some extent, being 345 nm in the CD spectrum and 330-335

nm in the UV-Vis spectrum.


71

Figure 4.14 Selected regions from the CD spectra of 7+ and 8+ in MeCN at 0° C.

The detector used for the CD measurements (see experimental part) has been specified by the

manufacturer to be suitable in the range 165-850 nm, but the spectral-noise becomes already

relatively high at 550 nm. Despite the spectral noise, it can be observed that the CD spectrum

of 8+ displays a second considerably weaker negative Cotton effect with its maximum located

at 553 nm. The position of the negative maximum is nearly identical to the λmax of the band in

the UV-Vis spectrum of 8+ at 546 nm. Thus, it appears that the transition with d-d character in

the UV-Vis spectrum of 8+ is also optically active. The related transition in the UV-Vis

spectrum of 7+ has its maximum at 622 nm and a Cotton effect associated with this electronic

transition could not be detected because of the high spectral noise in the CD spectrum at this

wavelength.

4.8 EPR spectroscopic investigations

The results from the low temperature electrochemical investigations on 7 and 8 clearly

demonstrated that the oxidised complexes 7+ and 8+ exhibit a strong preference for the

conformation in which their allyl or 2-Me-allyl ligand occupies a position trans relative to the

carboxylate oxygen atom. The high values for K4 imply that at temperatures at which

exchange occurs (higher than –40°C), only at the most about 1% of the complexes 7+ and 8+

exists in the form with their allyl ligand trans to the Nδ. The UV-Vis spectro-electrochemical

investigations showed that the energy levels of the orbitals with d-character of 7+ and 8+ are

not identical, because the transitions with d-d character do not occur at identical wavelengths.

In order to gain more insight into the electronic structure of 7+ and 8+, EPR spectroscopic

Chapter 4

72

investigations were performed on these compounds. The spectra were recorded on frozen

MeCN solutions after generation of 7+ and 8+ by controlled potential coulometry at –20° C. In

view of the long times required for generation of 7+ and 8+ by coulometry (10-15 minutes), it

is likely that the oxidised compounds will exist in their equilibrium composition. Because

freezing of the solution in liquid nitrogen takes place rapidly, it is expected that the frozen

solution subjected to EPR spectroscopic investigations provides a snapshot of the complexes

in solution at –20° C. The X-band EPR spectra of 7+ and 8+ are not very informative, because

they consist of resonances that are not well separated. In addition, an isotropic signal at

around g = 2.00-2.01 was found to be present in varying ratios. The spectra of 7+ and 8+ have

a different appearance but more information cannot be derived from these spectra. With the

objective to obtain well-resolved signals, Q-band EPR spectra of both 7+ and 8+ were

recorded. The resulting Q-band spectra are shown in Figure 4.15 (for 8+) and Figure 4.16 (for

7+), together with the results from the simulation.

No resonances were detected in the Q-band EPR spectrum of 8+ at around 1180 mT and 1220

mT. The resonances in both Q-band EPR spectra are well resolved and at first glance it

appears that the spectrum of 7+ consists of two subspectra, with one of those being identical to

the spectrum observed for 8+. Therefore, the spectrum of 8+ was first simulated and it could

only be fitted with two components. One of the components is an isotropic spectrum with a g-

value around 2.007, whereas the other subspectrum is an axial spectrum with g⊥ = 2.010 and

g|| = 2.005. Saturation experiments indeed revealed that the observed resonances do not

belong to the same species. Further evidence that the isotropic spectrum and the axial

spectrum are not related is obtained from the fact that the ratio between these resonances

varies per coulometry and subsequent EPR spectroscopic measurement. Because this isotropic

signal has a g-value around 2.007, it appears that this resonance originates from an organic

radical, probably resulting from decomposition of the oxidised complex. It has to be noted

that 7+ and 8+ are very sensitive towards traces of dioxygen, and it is conceivable that small

amounts of dioxygen have come into contact with the sample during its preparation, because

the solution has to be transferred from the coulometry vessel via a syringe into the thin Q-

band tube. The axial spectrum likely originates from the complex 8+, because it is a typical

spectrum for a transition metal complex. The ratio between the isotropic signal and the axial

signal in the Q-band spectrum is 8 / 92.


73

Figure 4.15 Q-band EPR spectrum of 8+, generated by coulometry, as a frozen MeCN

solution (0.1 M NBu4PF6), together with the simulated spectrum. T = 50 K;

Power: 9.62 × 10-3 mW; Frequency: 34.009702 GHz; Modulation: 0.51 mT.

Simulation: sub 1) gx = 2.0076, gy = 2.0070, gz = 2.0069; W = 15.0 G; sub 2)

gx = gy = 2.0097, gz = 2.0048; Wx = Wy = 16.0 G, Wz = 10.0 G

Chapter 4

74

Figure 4.16 Q-band EPR spectrum of 7+, generated by coulometry, as a frozen MeCN

solution (0.1 M NBu4PF6), together with the simulated spectrum. T = 50 K;

Power: 9.62 × 10-3 mW; Frequency: 33.9940100 GHz; Modulation: 0.82 mT.

Simulation: sub 1) gx = 2.0076, gy = 2.0070, gz = 2.0069; W = 15.0 G; sub 2)

gx = gy = 2.0097, gz = 2.0048; Wx = Wy = 16.0 G, Wz = 10.0 G; sub 3)gx =

2.0592, gy = 2.0025, gz = 1.9873; Wx = 90.0 G, Wy = 40.0 G, Wz = 54.0 G; sub

4) gx = gy = gz = 2.0138; W = 669.8 G


75

The Q-band EPR spectrum for 7+ had to be simulated with four different components. For two

of these species, the g-values obtained from the simulation procedure for 8+ were taken as

such and, surprisingly, this fits well. Furthermore, a third component with a rhombic spectrum

was simulated, having g-values: gx = 2.059, gy = 2.003 and gz = 1.987. The fourth

subspectrum is a very broad isotropic spectrum with a g-value of 2.014 and a line-width of

670 Gauss. This seems to be a signal originating from inhomogeneity, and the observed g-

value of 2.014 excludes that it originates from the organic radical and the species displaying

the axial spectrum. Because its g-value is approximately the average of the three g-values for

the observed rhombic spectrum (gav = 2.016), it is very conceivable that this broad resonance

originates from the same species that displays the rhombic spectrum. If the broad resonance

and the rhombic subspectrum are assumed to originate from the same species, a ratio between

the organic radical, the axial spectrum and the rhombic spectrum of 0.1% / 0.8% / 99.1% is

obtained.

The results appear to be very puzzling, because only one signal originating from a metal

complex species is observed for 8+, whereas two spectra owing to transition metal compounds

are observed for 7+. Interestingly, the g-values observed for the molybdenum species in the

spectrum of 8+ are identical to those of the minor component in the EPR spectrum of 7+. The

electrochemical investigations at low temperatures showed that both the allyl and 2-Me-allyl

compounds have a very strong preference in the oxidised form for the conformation in which

the allyl or 2-Me-allyl ligand is trans to the carboxylate oxygen atom. The results from EPR

spectroscopy can therefore not be explained by taking into account the presence of the

isomers with the allyl or 2-Me-allyl ligand trans to the oxygen atom and trans to the Nδ

nitrogen atom.

In order to get additional information on the relative energies of the isomers of 7+ and 8+, DFT

calculations were made on these compounds. It was hoped that these calculations could

correlate the results from the electrochemical and EPR spectroscopic investigations.

4.9 Density Functional Theory calculations on the oxidised complexes

The relative energies for the isomers a-f of 5+ and 6+ (see Scheme 4.3 for a schematic

depiction of the structures) have been calculated, with geometry optimisations starting from

the structures of the respective neutral isomers. It was observed computationally that upon

oxidation the bonding between the Mo atom and the histidine donor atoms for each isomer is

Chapter 4

76

reinforced. In particular, the bond between the Mo atom and the carboxylate oxygen atom

becomes much stronger and it was calculated to compress by approximately 0.12 Å. In

addition, the decrease of the electron density at the metal ion results in decreased backbonding

and the Mo-C (carbonyl as well as allyl) distances are computed to elongate by 0.02 to 0.09

Å. These computational findings are in good agreement with the results from the infrared

spectro-electrochemical investigations.

Table 4.9 DFT-computed relative energies (in kJ mol-1) for the isomer a-f of 5 and 6 and

their one-electron oxidised analogues 5+ and 6+

isomer a b c d e f

complex

5[a] 0.0 0.4 6.3 30.5 33.9 25.9

5+ 5.5 0.0 15.3 13.8 3.2 25.3

6[a] 0.0 0.8 5.9 24.6 26.3 18.8

6+ 18.8 15.2 34.0 22.5 0.0 33.6

[a] The relative energies for these compounds have already been presented in Scheme 4.3

The resulting energies for the isomers of 5+ and 6+ are shown in Table 4.9, together with those

for the isomers of the neutral compounds. A completely different energetic sequence is

obtained for the cationic complexes in comparison to the neutral compounds. On going from 5

to 5+, the isomers a and b switch their order of stability and the energy difference between 5a

and 5b of 5.5 kJ mol-1 is in good agreement with the value obtained from the variable

temperature electrochemical investigations on the methyl propionate derivative 7+ (7.5 kJ

mol-1). For 6+, the energetic sequence differs from that of 5+, because the e isomer is in this

case by far energetically favoured. Isomer e has the same arrangement of the ligands around

the Mo atom as b, but it differs from b in the orientation of the allyl or 2-Me-allyl ligand (see

Scheme 4.3). The energy difference between 6e+ and 6a+ can be regarded as the actual

difference that is determined from the low temperature electrochemical experiments. It is

conceivable that the 2-Me-allyl ligand of isomer 6b undergoes a 180° rotation upon oxidation.

In fact, DFT-calculations on the activation barrier for interconversion between the neutral


77

isomers 6a and 6b indicated the presence of endo-intermediates during the interconversion

process. It must be noted that the variable temperature electrochemical investigations only

allow the determination of the energy-differences between the conformations with their allyl

or 2-Me-allyl ligand trans to the O-atom and trans to the Nδ. A rotation of the allyl or 2-Me-

allyl ligand will not be observable by electrochemistry. A small inconsistency between the

computational results for the isomers of 6+ and the results from the electrochemical

investigations on 8+ is that the former find the energy difference between 6a+ and 6e+ to be

18.8 kJ mol-1, whereas the latter find an energy difference between the conformations with

their 2-Me-allyl ligand trans to the O-atom and trans to the Nδ of 8.3 kJ mol-1.

Closer inspection of the energy profiles of the isomers of 5+ reveals that the e isomer is only

3.2 kJ mol-1 higher in energy than isomer b. Because this difference is not high, it is expected

that also a fraction of the complexes exists in this conformation.

The results from the DFT calculations on the cationic complexes provide an explanation for

the observed EPR spectra of 7+ and 8+. The EPR spectrum of 8+, a complex which only differs

from 6+ by the presence of a methyl propionate substituent on the Nε-atom, shows only one

spectrum originating from a transition metal complex, whereas two metal complex spectra are

observed for 7+. In addition, the minor component in the EPR spectra of 7+ has identical g-

values as the only component observed for 8+. On the basis of the results from DFT

calculations, it might be that the axial EPR spectrum observed for 8+ and the minor metal

complex component in the EPR spectrum of 7+ are owing to their e isomers. The major

component in the Q-band EPR spectrum of 7+, the rhombic spectrum, might originate from

isomer 7b+.

This possible explanation for the EPR spectroscopic data raises two new questions. First of

all, what the electronic reason for rotation of the allyl ligand in the oxidised form is. The

second issue concerns the similarity of the g-values for the isomers 6e+ and 5e+, because this

suggests an identical composition of the SOMO (singly occupied molecular orbital) for these

isomers.

An inspection of the SOMOs obtained from the DFT calculations reveals that in the case of

5b+ both carbonyl ligands are involved in backbonding, whereas the allyl ligand is left out of

a backbonding interaction. The SOMOs of 5e+ and 6e+ are nearly identical and in both cases

strong backbonding takes place to the allyl or 2-Me-allyl ligand, whereas only one of the

Chapter 4

78

carbonyl ligands is involved in a backbonding interaction. These observations from the DFT

calculations answer the above mentioned issues. The reason for rotation of the allyl ligand is

that it becomes involved in a very strong backbonding interaction in the SOMO. Considering

the large similarity of the SOMOs of 5e+ and 6e+, it appears very plausible that the unpaired

electron experiences an identical environment and, consequently, displays an identical EPR

spectrum. The SOMO for isomer 5b+ is completely different and thus a different EPR

spectrum is expected for 7b+.

4.10 Concluding remarks

By combining the results from X-ray crystallography, spectroscopic investigations, low

temperature electrochemical investigations and DFT-calculations, the fluxional behaviour in

solution of the complexes 5, 6, 7 and 8 as well as their oxidised analogues 7+ and 8+ has been

elucidated in detail. The resulting thermodynamically favoured isomers for the allyl and 2-

Me-allyl complexes in the neutral and one-electron oxidised forms are shown in Scheme 4.7.

In the neutral form, the allyl and the 2-Me-allyl complexes exist in their isomeric forms a and

b in a ratio close to 1/1. The isomers a and b are in equilibrium and the netto process for

interconversion between these conformers is a restricted trigonal twist. However, the results

from DFT-calculations indicate that the interconversion does not proceed via an

uncomplicated pseudo-rotation but instead likely involves some endo-isomers as

intermediates. The computed activation barriers for interconversion are in good agreement

with the experimentally obtained energy barriers from variable temperature 1H NMR

spectroscopic investigations. After oxidation, the complexes strongly prefer the conformation

in which their allyl or 2-Me-allyl ligand occupies a trans-position relative to the carboxylate

O-atom, as revealed by low temperature electrochemical investigations on 7 and 8. By

combining the results from EPR spectroscopic investigations on 7+ and 8+ and DFT-

calculations for 5+ and 6+, the orientation of the allyl or 2-Me-allyl in the oxidised complexes

could be elucidated. The 2-Me-allyl complex 8+ likely exists only in the isomer with an endo

orientation of the 2-Me-allyl ligand, with this ligand being trans relative to the carboxylate

oxygen atom (isomer 8e+). The results for 7+ indicate that this compound exists as a mixture

of two isomers, with the main component being isomer 7b+, but a small fraction of the

complexes have an endo-orientation of the allyl ligand (isomer 7e+). The reason for rotation of

the allyl or 2-Me-allyl ligand could be explained by an inspection of the SOMOs obtained


79

from the DFT calculations. If the allyl or 2-Me-allyl ligand is rotated (the endo-

conformation), backbonding is taking place to this ligand in the SOMO, whereas the allyl

ligand is not involved in a backbonding interaction in the SOMO in the case of the exo-

orientation of this ligand.

Scheme 4.7 Overview of the favoured isomers of the allyl and 2-Me-allyl complexes in

solution in the neutral and one-electron oxidised form

N

N

H2N

O O

MoOC

OC

R

N

N

H2N

O O

MoOC

CO

R

N

N

H2N

O OMoOC

OC

R

N

N

H2N

O OMo

OCCO

R

N

N

H2N

O O

MoOC

CO

R

N

N

H2N

O O

MoOC

CO

R +

N

N

H2N

O O

Mo

+

OCCO

R +

oxidation

minor major

oxidation

Complexes with the allyl ligand Complexes with the 2-Me-allyl ligand

a b

b e e

a b

This is the first report of such a detailed study of the fluxional behaviour of pseudo-octahedral

molybdenum complexes containing (η-allyl)(CO)2 and (η-2-Me-allyl)(CO)2 fragments.

Although the complexes only differ in the presence of a methyl group on the central carbon

atom of the allyl ligand, this has significant impact on the fluxional behaviour and the

conformational preferences in solution, both in the neutral form and the one-electron oxidised

form.

In the next chapter, the use of the Mo(His)(η-allyl)(CO)2 moiety as a marker for biomolecules

is investigated.

80

5Markers based on the complex Mo(His)(ηηηη-allyl)(CO)2


In the previous chapter, it has been demonstrated that Mo(L-His-Nε-C2H4CO2Me)(η-

allyl)(CO)2 and the 2-Me-allyl analogue each display a reversible one-electron oxidation at a

potential similar to the Fc/Fc+ couple. Although 7 and 8 exhibit more complicated

electrochemical behaviour at temperatures much lower than 0° C, their oxidation-wave at

room temperature has a perfect reversible appearance, thus making these compounds suitable

for use as an electrochemical marker. These complexes can also serve as an infrared

spectroscopic handle, because the carbonyl ligands in 7 and 8 show C-O stretching vibrations

in the solution infrared spectrum at around 1835 and 1932 cm-1.

In addition, 7 and 8 are much less reactive towards dioxygen than Mo(η-Cp)(η-allyl)(CO)2

and the derivatives thereof described in Chapter 3. In the solid state, 7 and 8 are stable for

weeks in air, and also under an aerobic atmosphere in solution, these compounds appear not to

decompose significantly during a period of 24 hours. However, when the complexes are

subjected to 1H NMR spectroscopic investigations after this period, broadened resonances are

observed, which is indicative for the presence of small amounts of paramagnetic

decomposition products.

Markers based on the complex Mo(His)(η-allyl)(CO)2

81

Because both compounds are about equally stable towards dioxygen, they could in principle

both serve as markers. However, my preference was to use the allyl derivative as a labelling

reagent, because the NMR spectra of this type of complexes display two distinct sets of

resonances for the isomers a and b. The isomers of the 2-Me-allyl derivatives on the other

hand have a lower activation barrier for interconversion, which would complicate the NMR

spectroscopic investigations of the bioconjugates due to the broad resonances as a

consequence of occurring coalescence at room temperature.

Because the ligand histidine is chiral, my interest concerned using both the complexes with L-

histidine and D-histidine as labels. These compounds as such are enantiomers, but when these

are coupled to amino acids and peptides of L-configuration, the thereby obtained

bioconjugates are diastereomers. It appeared interesting to compare the circular dichroism

spectra of the bioconjugates containing the D-histidine and L-histidine complexes and to

investigate whether the pairs of diastereomeric compounds would display other spectroscopic

differences.

In the following section, the synthesis of the markers Mo(His-Nε-C2H4COOH)(η-allyl)(CO)2

with both L-histidine and D-histidine is described, followed by the coupling of both to L-

phenylalanine methyl ester and the dipeptide H-L-Phe-L-Leu-OMe via peptide synthesis

methods in solution.

5.2 Synthesis of the enantiomeric markers and diastereomeric bioconjugates

Analogous to the synthesis of 5, the reaction of Mo(η-allyl)Br(CO)2(NCMe)2 with D-

histidine⋅HCl and two equivalents of KOH in a MeOH / H2O mixture yields Mo(D-His)(η-

allyl)(CO)2 (10) in 90% yield. By reacting 10 with 3-bromo-methyl propionate under identical

conditions as described for the synthesis of the L-histidine analogue 7, the complex Mo(D-

His-Nε-C2H4CO2Me)(η-allyl)(CO)2 (11) is obtained. Because the complexes are enantiomers,

the UV-Vis, infrared and NMR spectra of 10 and 11 are identical to those of 5 and 7. The

compounds with L-histidine only differ spectroscopically from the D-histidine analogues in

that their CD-spectra are mirror images (vide infra).

Chapter 5

82

Scheme 5.1 Route of synthesis for the markers Mo(His-Nε-C2H4COOH)(η-allyl)(CO)2

N

N

H2N

OO

Mo CO

CO

OCH3O

N

N

H2N

OO

Mo CO

CO

OHO

L-His 7

D-His 11

L-His 12

D-His 13a

Reagents and conditions: a) NaOH in H2O / MeOH; stirring overnight at RT; addition of diluted HCl.

Reacting 7 or 11 with five equivalents of KOH in a MeOH / H2O mixture overnight results in

hydrolysis of the methyl ester, as shown in Scheme 5.1. The resulting carboxylate derivative

and its corresponding acid following pH adjustment to neutrality with diluted hydrochloric

acid are only soluble in DMSO, DMF, alcohols and water. Consequently, the acids Mo(His-

Nε-C2H4COOH)(η-allyl)(CO)2 (12 for L-His; 13 for D-His) were never obtained in pure form,

but instead as mixtures with salts, but these do not have any influence on coupling reactions.

The best work-up was found to be adjustment of the pH to 7 after hydrolysis of the methyl

ester, followed by removal of the solvent in vacuo, yielding the acids as mixtures with KCl.

Care should be taken to avoid the pH from becoming too low, because the compounds are

somewhat sensitive to acidic conditions, which might be rationalised on the basis of

protonation and dissociation of the allyl ligand.

The 13C NMR spectra and FAB-positive and negative mass spectra of 12 and 13 are

consistent with their formulation. Prior to the coupling reactions of 12 and 13, the complex /

salt mixture was suspended in a mixture of DMF and dipea and subsequently filtered to

remove most of the salts. To this DMF / dipea solution of either 12 or 13 was added

approximately equimolar amounts of phenylalanine methyl ester hydrochloride or the

dipeptide H-Phe-Leu-OMe and one equivalent of TBTU, as shown in Scheme 5.2. After

stirring at room temperature for 30 minutes, the solvent was removed in vacuo. The residue

was treated with EtOH, subsequently filtered to remove insoluble material (salts and starting

material), followed by removal of the solvent in vacuo. After purification by preparative

HPLC, the bioconjugates 14-17 shown (see Scheme 5.2) were obtained in 45-55 % yield in

pure form.


83

Scheme 5.2 Synthesis and constitution of the amino acid and dipeptide derivatives

N

N

H2N

OO

MoCO

CO

OHO

N

N

H2N

OO

MoCO

CO

AA-OMeO

14 L 15 D 16 L17 D

His AA Phe PhePhe-LeuPhe-Leu

a

Reagents and conditions: a) TBTU and either H-Phe-OMe or H-Ala-Phe-OMe, DMF / dipea mixture,

stirring at RT for 30 mins.

Evidence for the proposed constitution of these compounds is first of all obtained from the

results from elemental analysis, which fit well. Furthermore, the FAB mass spectra show

clusters with the appropriate [M+H]+ peak and these clusters have an isotopic distribution in

good agreement with the calculated one. In addition, all the resonances in the 1H and 13C

NMR spectra can be assigned unambiguously (see experimental part and section 5.4).

By slow evaporation of a MeOH / H2O solution of 14, single crystals suitable for X-ray

analysis of this compound were obtained. In the next section the X-ray crystal structures of

10⋅MeOH and 14 are described, and these are compared mutually as well as with the

structures 5⋅MeOH and 7⋅2MeOH

5.3 X-ray crystallography

In a similar manner as for 5⋅MeOH, single crystals of 10⋅MeOH were obtained, which were

subjected to X-ray structure determination. ORTEP plots for 5⋅MeOH and 10⋅MeOH are

shown in Figure 5.1. As expected, these two complexes are each other’s mirror image, and the

Mo-donor atom bond lengths for both structures are very similar, i.e. identical or within 1σ, as

summarised in Table 5.1. Also the hydrogen bonding pattern for 10⋅MeOH is identical to that

observed for 5⋅MeOH.

Chapter 5

84

Figure 5.1 ORTEP plots for 10⋅MeOH (left) and 5⋅MeOH (right), visualising the mirror-

image relationship.

The Nδ nitrogen atom of the histidinate ligand occupies a trans-position relative to the allyl

ligand in the structures 5⋅MeOH, 7⋅2MeOH and 10⋅MeOH and also for 14 an identical solid

state conformation is observed. An ORTEP projection for 14 is shown in Figure 5.2, with Mo-

donor atom bond distances listed in Table 5.1.

Table 5.1 Selected bond lengths (Å) for 5⋅MeOH, 7⋅⋅⋅⋅2MeOH, 10⋅⋅⋅⋅MeOH and 14

5⋅⋅⋅⋅MeOH 7⋅⋅⋅⋅2MeOH[a] 10⋅⋅⋅⋅MeOH[a] 14

Mo(1) - N(1) 2.217(2) 2.216(2) 2.214(3) 2.2009(12)

Mo(1) - O(9) 2.222(2) 2.214(2) 2.222(3) 2.2037(11)

Mo(1) - N(11) 2.272(2) 2.256(2) 2.273(3) 2.2866(12)

Mo(1) - C(30) 1.945(2) 1.944(3) 1.944(4) 1.947(2)

Mo(1) - C(40) 1.948(2) 1.945(3) 1.948(5) 1.948(2)

Mo(1) - C(20) 2.326(2) 2.328(3) 2.324(4) 2.317(2)

Mo(1) - C(21) 2.212(3) 2.209(2) 2.214(5) 2.2006(14)

Mo(1) - C(22) 2.320(2) 2.328(3) 2.329(5) 2.308(2)

[a] Bond distances for these structures have already been presented in Table 4.1


85

Figure 5.2 ORTEP projection for Mo(L-His-Nε-C2H4C(O)-Phe-OMe)(η-allyl)(CO)2 (14)

The standard deviations for the Mo-donor atom bond distances of 14 are very small on

account of the high quality of the structure of 14 (R1 for all data = 0.0274). Compared to the

other three structures shown in Table 5.1, the Mo(1)-N(1) and Mo(1)-O(9) bond distances in

14 are about 0.01-0.02 Å shorter, whereas the Mo(1)-N(11) bond length in 14 is significantly

elongated in comparison to the crystal structures of the other three compounds. In addition,

the Mo-C(allyl) bond distances in 14 are considerably shorter than those observed for the

other three compounds. The differences in bond-lengths for 14 with respect to the structures

5⋅MeOH, 7⋅2MeOH and 10⋅MeOH are likely caused by packing effects, since the bulky Phe-

OMe substituent has to be accommodated in the lattice of 14.

Intramolecular and strong intermolecular hydrogen bonds are not present in the solid state.

Weak intermolecular hydrogen bonds exist between N(11) and O(13) (N···O contact = 3.051

Å) and between the amide NH moiety (N(5)) and O(10) (N···O contact = 3.082 Å).

Furthermore, a weak polarisation exists between N(5) and O(40) of a different molecule

(N···O contact = 3.188 Å), but in view of the low polarity of carbonyl ligands, this interaction

cannot be regarded as a hydrogen bond.

In the next section, the solution structures of the compounds 14-17 are investigated by NMR

spectroscopy.

Chapter 5

86

5.4 NMR spectra of the phenylalanine and dipeptide derivatives

Although the X-ray crystal structure of 14 revealed the conformation of isomer a in the solid

state, isomers a and b are both present in DMSO solution (see Scheme 4.2), as can be

concluded from the 1H and 13C NMR spectra. The region between 6 and 9 ppm of the 1H

NMR spectrum of 14 is displayed in Figure 5.3, together with the assignment of the

resonances for each conformer. The two sets of signals are clearly discernible, similar to what

is observed for the compounds 5 and 7 (see previous chapter). The effect of the two isomers

extends to the phenylalanine NH hydrogen atom, which is spatially comparable to what is

observed for 7, for which the CH3 group of the methyl ester shows a different resonance in

each isomeric form.

Figure 5.3 Selected region from the 1H NMR spectrum of 14 (500 MHz; DMSO-d6)

Also compounds 15-17 exist as mixtures of isomers a and b in DMSO, as can be concluded

from their 1H NMR spectra. The ratio between the isomers a and b in DMSO-d6 is in all four

compounds 53 / 47, which is identical to the a / b ratio of 7. Apparently, this ratio is not

affected by the size of the substituent and also not by the chirality of the histidinate ligand. To

investigate whether the presence of bulky substituents has any effect on the activation barrier

for interconversion, the complexes 14-17 were subjected to variable 1H NMR spectroscopic

investigations. Also in this case no influence was observed: all four compounds show an

activation barrier for interconversion of 67.0 ± 1.0 kJ mol-1, virtually identical to the

activation energy for 7 (66.9 ± 1.0 kJ mol-1).


87

The 13C NMR spectra were in particular very useful for further confirming the constitution of

these compounds. Because the number of carbon atoms in these compounds is not very high,

all the resonances in the 13C NMR spectra were assigned easily. The 1H NMR spectra on the

other hand are rather difficult to interpret, since many signals overlap and the compounds

exist as mixtures of isomeric forms a and b. However, all the signals in the 1H NMR spectra

of 14-17 were assigned by using 2D NMR techniques (1H- 1H COSY and 1H-13C HMQC).

Although compounds 14 and 15 as well as 16 and 17 are diastereomers, large

differences are not observed in the 1H and 13C NMR spectra and, in fact, the diastereomeric

compounds cannot even be distinguished on the basis of the NMR spectroscopic data.

Probably the chemical properties of these diastereomers are so similar that the chemical shifts

of some resonances of these differ by only a few hundredths of a ppm.

In the next section, it is investigated whether the coupling of 12 and 13 is compatible with

solid phase peptide synthesis methods.

5.5 Solid phase synthesis

Because the bioconjugates 14-17 exhibit good stability towards dioxygen, comparable to that

of the compounds 7 and 11, investigations were made whether the complexes 12 and 13 can

be tethered to larger peptides via SPPS. The synthesis of [Leu]-enkephalin on the resin was

already discussed in detail in Chapter 2. Because the phenolic OH group of the tyrosine-

residue might have a negative influence on the coupling reaction, the complexes 12 and 13

were first reacted with resin-bound Fmoc-deprotected [Leu]-enkephalin that still contains the

2-ClTrt group. The coupling reactions were performed overnight with five equivalents of the

enantiomeric metal complexes, in order to ensure that these reactions proceed in high yield.

After washing the resin the next day to remove side products and excess starting material,

already by visual inspection it could be concluded that the coupling reactions of the metal

complexes had been successful, because the normally off-white resin had turned yellow. The

[Leu]-enkephalin metal complex bioconjugates were cleaved from the resin by treatment with

a saturated NH3 solution in MeOH for 48 hours. After that period, the resin was removed by

filtration, and the yellow filtrate was concentrated to an oil in vacuo. Purification by

preparative HPLC afforded about 100 mg of the [Leu]-enkephalin derivatives 18 and 19

depicted in Scheme 5.3 in highly pure form.

Chapter 5

88

Scheme 5.3 Molecular structure of the [Leu]-enkephalin conjugates obtained via SPPS

N

N

H2N

OO

MoCO

CO

HN

NH

O

O

HN

O

NH

O

N

O

O

R

NH2

O

18 L 19 D 20 L21 D

His R

2-ClTrt2-ClTrt H H

Subsequently, the complexes 12 and 13 were reacted in an analogous manner with resin-

bound [Leu]-enkephalin, from which the 2-ClTrt and Fmoc protecting groups were already

previously removed (see Chapter 2 for details). After an identical work-up as described for the

2-ClTrt derivatives above, about 65 mg of each compound 20 and 21 (see Scheme 5.3) was

obtained in pure form.

Evidence for the proposed constitution of 18-21 was first of all obtained from the ESI-positive

mass spectra. In the ESI-MS spectra of 18 and 19 a cluster around m/z = 1256.5 is observed,

which is assigned as the [M+Na]+ peak and the isotopic distribution of this cluster is

consistent with the expected calculated one. Furthermore, a signal of high intensity at m/z =

277 is observed, which is owing to the 2-ClTrt protecting group. In the ESI-positive mass

spectra of 20 and 21 clusters of peaks around m/z = 958 and m/z = 980 are observed, which

are assigned as [M+H]+ and [M+Na]+ peaks, respectively. For 20 also an ESI-positive high-

resolution mass spectrum was recorded, yielding an exact mass for the [M+Na]+ peak of

980.2826, which is in good agreement with the calculated mass (980.2817).

Further evidence for the constitution of the compounds 18-21 is obtained from the NMR

spectra, which are discussed in the following section.


89

5.6 NMR spectra of the [Leu]-enkephalin bioconjugates

Especially the 13C NMR spectra of the compounds 18-21 are very informative. Although each

compound has a large number of carbon atoms (61 for 18 and 19 and 42 in the case of 20 and

21), all the observed resonances in the 13C NMR spectra were assigned easily by comparing

the spectra with those of 14-17. The 13C NMR spectra of 18 and 19 were measured in DMSO-

d6, whereas those of the more interesting bioconjugates 20 and 21 were recorded in CD3OD,

because the compounds can be recovered after the measurement in that way. The ratio

between the isomers a and b is solvent dependent, as shown in the previous chapter for

compound 7, being around 5 / 4 in DMSO and around 4 / 1 in CD3OD. Therefore, resonances

owing to both isomers in the 1N NMR spectra of 18 and 19 can be assigned reliably, whereas

signals due to the major isomer could only be detected with certainty in the 13C NMR spectra

of 20 and 21.

The aromatic region of the 13C NMR spectra of 18 and 19 is expected to be comprised of 25

resonances: five for the histidinate (resonance owing to Cδ is identical for both isomers), four

for the Phe-residue, four for the unsubstituted aromatic rings of the 2-ClTrt group, six for the

Cl-substituted aromatic ring of the 2-ClTrt group, and six for the tyrosine aromatic ring,

because this ring contains a bulky unsymmetrical protecting group. Exactly 25 resonances are

observed in the aromatic region of the 13C NMR spectra of 18 and 19, which further

substantiates the identity of these compounds.

The 1H NMR spectra of 18-21 were recorded in DMSO-d6, because the spectra in CD3OD

will be devoid of resonances owing to NH and OH groups because of H / D exchange in that

solvent. The 1H NMR spectra of 18 and 19 are virtually identical, whereas these of 20 and 21

are slightly different. The region between 6 and 9.5 ppm in the 1H NMR spectra of 19, 20 and

21 is depicted in Figure 5.4, together with the assignment of the resonances.

Chapter 5

90

Figure 5.4 Selected regions from the 1H NMR spectra (500 MHz; DMSO-d6) of 19-21,

together with the assignment of the resonances.


91

In the 1H NMR spectrum of 19, several resonances in between 7.4 and 7.2 ppm are observed

due to the 2-ClTrt protecting group and the 1H NMR spectra of 20 and 21 show resonances

owing to the Tyr-OH at around 9.2 ppm. This OH hydrogen atom appears as two signals, with

the same ratio between them as between the CδH resonances of the histidinate (55 / 45), which

indicates that the OH hydrogen atom experiences a different chemical environment in each

isomeric form. This seems to be peculiar, because in the 1H NMR spectra of the other

bioconjugates discussed thus far the effect of the different isomers only extends to the first

amide hydrogen atom. Also in the spectra of 18-21 the first NH hydrogen atom (the Tyr-NH)

shows up as two resonances. In the NMR spectrum of both 20 and 21, the Tyr-OH resonance

at a higher chemical shift has the higher intensity and, thus, this hydrogen atom is deshielded

in isomer a. Isomer a has the Nδ nitrogen atom in a trans position to the allyl ligand and the

carboxylate trans relative to a carbonyl ligand, whereas the situation is the opposite in

isomeric form b, for which the Tyr-OH resonance is located is upfield. The observation of

two different resonances for the Tyr-OH hydrogen atom can be explained by the presence of a

hydrogen bond between the Tyr-OH group and the histidinate carboxylate moiety. When the

carboxylate is trans to a carbonyl ligand (isomer a), the Mo-O bond length is slightly

elongated and weakened compared to the conformation of isomer b, which has the

carboxylate trans relative to the allyl ligand. Because the carboxylate moiety has more

electron density in isomeric form a, it acts as a stronger hydrogen bond acceptor in this

conformer. As a result of the stronger hydrogen bond between Tyr-OH moiety and the

carboxylate in isomeric form a, the resonance of the Tyr-OH hydrogen atom shifts to higher

frequency with respect to conformer b. The resonances of the Tyr-OH hydrogen atom differ

in 20 and 21, being 9.19 and 9.15 ppm for the former and 9.25 and 9.19 for the latter. Thus,

apparently, the hydrogen bond between the Tyr-OH and the carboxylate is stronger in the case

of the bioconjugate 21 containing the D-histidine complex.

Further evidence that the Tyr-OH is hydrogen bonded with the coordinated

carboxylate group is obtained from a different spectral feature. In the 1H NMR spectrum of 21

the histidine CδH resonates at 6.81 / 6.68 ppm, whereas the resonances of this hydrogen atom

are located about 0.12 ppm downfield in the 1H NMR spectra of all the other bioconjugates.

The phenol ring of the tyrosine is expected to be in the proximity of the CδH hydrogen atom

in order to make a hydrogen bond between the OH moiety and the carboxylate group possible.

The shift to lower frequency of the CδH hydrogen atom can be explained in this way by the

magnetic anisotropy effect of the tyrosine phenol ring. Relationships between the magnitude

Chapter 5

92

of the upfield shift and the distance from the centre of a phenyl ring have been derived by

theoretical calculations and these were confirmed by spectroscopic investigations [179, 180].

In addition to the distance of the aromatic ring to the hydrogen atom also the orientation of the

ring has an influence and the largest upfield shift is observed when the imaginary axis joining

the hydrogen atom and the centre of the phenyl ring is perpendicular to the ring. In fact, the

space is divided into specific segments in the above-mentioned relationship and several

different distances and orientations would result in an upfield shift of 0.12 ppm. Furthermore,

it should be noted that this relationship applies to phenyl rings, whereas the aromatic ring

causing the downfield shift is a phenol ring. Thus, it can only be concluded that the phenol

ring is in the proximity of the CδH hydrogen atom, but the exact distance cannot be

determined from the observed shift.

A similar upfield shift for CδH hydrogen atom of 20 is not observed, which might be due to

spatial reasons. The complexes 5 and 10 are enantiomers, differing in the chirality of the

histidine ligand. If these two complexes are superimposed on each other, while forcing the

allyl and carbonyl ligands as well as the Nδ atom to have the same orientation, the carboxylate

and amine NH2 occupy opposite positions relative to each other. In general, every

stereochemical carbon atom can be transformed in the other enantiomer by exchanging two

substituents, while leaving the other two unaffected. The tyrosine residue in both 20 and 21

has the L-conformation, but to make a hydrogen bond of the OH-moiety with the carboxylate

possible, it has to be oriented slightly different in 20 with respect to 21 because the

carboxylate occupies different positions in both compounds. The fact that the resonances of

the Tyr-OH hydrogen atom are different in 20 and 21 shows that the spatial orientation of the

hydrogen bond interaction is not identical. The phenol ring might be closer to the CδH

hydrogen atom in 21 than in 20, or the ring might be rotated to some extent, both of which

would provide an explanation for the observed differences between the resonances of the CδH

hydrogen atoms in 20 and 21.

To derive the exact hydrogen bonding conformations of the tyrosine OH group and the

carboxylate moiety of 20 and 21 an X-ray crystal structure would be very helpful.

Unfortunately, all attempts to crystallise these compounds have been in vain thus far. The

synthesis of Mo(His-Nε-C2H4C(O)-Tyr-OBzl)(η-allyl)(CO)2 (Bzl = benzyl) with both D-

histidine and L-histidine is planned for the near future. Perhaps X-ray quality crystals of these

compounds can be obtained, which would possibly reveal the exact conformation of the

hydrogen bond between the molybdenum complex and the tyrosine-OH moiety.


93

All of the other signals in the 1H NMR spectra of 18-21 were assigned by comparison of the

spectra to those of 14-17, in combination with results from 2D NMR spectroscopic

investigations on 18-21. It could, however, not be determined whether compounds 18-21 have

an extended structure or a bend-conformation, both of which are possible in the case of

normal [Leu]-enkephalin (see Chapter 2). The 1H NMR spectra consists of many overlapping

signals and therefore it is difficult to derive the coupling constants, which is the first step for

elucidating the conformation of a peptide in solution [181].

5.7 Circular dichroism spectroscopy

Circular dichroism spectra were recorded for the compounds described in this chapter, the

results of which are summarised in Table 5.2. DMSO was selected as the solvent for two

reasons. First of all, the ratio between the isomers as well as the structure in solution is in part

known in DMSO, because the 1H NMR spectra of these compounds were also recorded in that

solvent. For example, the tyrosine OH group of 20 and 21 was shown to be involved in a

hydrogen bond with the carboxylate moiety of the histidinate ligand. Secondly, DMSO was

chosen because the compounds dissolve readily in this solvent. The other suitable solvent

from a solubility perspective is MeOH but a spectroscopic advantage of MeOH over DMSO

was not derived from the CD measurements discussed in Chapter 4. The lower limit of optical

spectroscopy on DMSO solutions is around 270-280 nm, whereas this limit is approximately

210 nm in MeOH. Because the compounds show strong UV absorbtion below 280 nm, the

typical concentration of about 5 × 10-4 M used for the determination of the Cotton effects at

around 320 and 410 nm is too high for CD spectroscopic investigations at wavelengths lower

than 280 nm. The use of much more diluted solutions of the bioconjugates in MeOH resulted

in high spectral noise and consequently the dichroic absorption of the amino acids and peptide

chromophores, which occurs below 280 nm [182, 183], could unfortunately not be recorded

properly.

Chapter 5

94

Table 5.2 Circular dichroism spectroscopic data for 7, 11 and their related bioconjugates

in DMSO[a]

Complex θ320[b, c] θ406

[c, d] θ320 Complex θ320[b, c] θ406

[c, d] θ320

× 10-3 × 10-3 θ406 × 10-3 × 10-3 θ406

7 -8.5 -4.7 1.80 11 +9.0 +4.9 1.85

14 -8.7 -4.3 2.01 15 +9.5 +4.7 1.99

16 -9.0 -4.5 2.01 17 +8.4 +4.1 2.03

18 -8.3 -4.2 1.96 19 +8.9 +4.2 2.11

20 -7.9 -4.5 1.75 21 +8.5 +4.7 1.82

[a] Values for θ are not completely reliable because the compounds contain varying amounts of water;[b] at 320 nm; [c] Units of θ are 102 × ° M-1 cm-1; [d] at 406 nm

The CD-spectra of the compounds listed in Table 5.2 are quite similar, each displaying

maxima at 320 nm and 406 nm. The bioconjugates with the D-histidine complex show

positive Cotton effects, whereas those with the L-histidine complex exhibit negative Cotton

effects. Two representative examples of the recorded CD spectra are shown in Figure 5.5 for

the [Leu]-enkephalin conjugates 20 and 21.

Figure 5.5 CD spectra of the [Leu]-enkephalin derivatives 20 and 21 in DMSO


95

The reported molecular ellipticity at the maxima for the compounds in Table 5.2 is not

completely reliable because these amino acid and peptide derivatives are very difficult to

obtain in a completely dry form. Not influenced by the exact concentration on the other hand,

is the ratio between the maximum molecular ellipticity at 320 nm and 406 nm (θ320 / θ406),

which is therefore much more reliable than the values of θ. From Table 5.2, it is observed that

upon substitution of the methyl ester of 7 and 11 for a Phe-OMe or Phe-Leu-OMe rest, the

value θ320 / θ406 increases, independent of the chirality of the histidine. In the previous

chapter, it was found that the maxima in the CD spectrum of 7 exhibit a solvent dependence.

Because the ratio between the isomers a and b of 7 in different solvents is not identical, it

appears therefore that also the ratio between the isomers plays a role. Because the compounds

7, 11 and 14-17 have an identical a / b ratio in DMSO, as revealed by 1H NMR spectroscopy,

the variation of θ320 / θ406 is not caused by a differing a / b ratio. The ratio between the

isomers a / b could not be determined for 18 and 19, whereas this ratio was determined to be

55 / 45 for 20 and 21. It is observed that upon cleavage of the 2-ClTrt protecting group from

the [Leu]-enkephalin compounds, the ratio θ320 / θ406 decreases significantly. From 1H NMR

spectroscopic data, it was derived that the compounds 20 and 21 contain an intramolecular

hydrogen bond between the Tyr-OH and the carboxylate moiety of the histidinate. Hence, the

decrease of the θ320 / θ406 ratio of 20 and 21 with respect to 18 and 19 may be attributed to the

partly ordered structure of compounds 20 and 21 in DMSO.

The pairs of diastereomeric bioconjugates 14 / 15, 16 / 17, 18 / 19 as well as 20 / 21 can be

distinguished by their CD spectra. Only compounds 20 and 21 are also discernible on the

basis of their 1H NMR spectra, as shown in the previous section.

5.8 Electrochemistry and infrared spectroscopy

The observed νCO stretching vibrations in the infrared spectra of 14-21 both as a KBr pellet

and in MeOH solution are listed in Table 5.3. The infrared spectra in MeOH were recorded by

using the OTTLE-cell, which contains CaF2 windows. As can be observed, all the compounds

display similar νCO stretching frequencies, independent of the chirality of the histidine and the

size of the substituent. To illustrate the possible use of the Mo(His)(η-allyl)(CO)2 unit as a

marker, selected regions from the infrared spectra of the [Leu]-enkephalin derivative 20 both

as a KBr pellet and in MeOH solution are depicted in Figure 5.6.

Chapter 5

96

Table 5.3 Oxidation potentials and selected vibrational data for 14-21

Compound νCO (KBr)[a, b] νCO (MeOH)[a, c] E½ [d, e]

14 1830, 1927 1938, 1847 + 0.08[f]

15 1834, 1927 1938, 1847 + 0.08[f]

16 1832, 1931 1938, 1847 + 0.07[f]

17 1831, 1930 1938, 1846 + 0.07[f]

18 1832, 1931 1939, 1846 +0.02[g]

19 1831, 1931 1938, 1846 +0.02[g]

20 1830, 1930 1938, 1846 +0.02[g]

21 1831, 1931 1939, 1848 +0.02[g]

[a] in cm-1; [b] 2 cm-1 resolution; [c] OTTLE-cell with CaF2 windows used; path-length 0.17 mm; 2 cm-

1 resolution; [d] in V vs. Fc/Fc+; [e] NBu4PF6 used as supporting electrolyte; [f] in MeCN; [g] in DMF

Figure 5.6 Selected regions from the infrared spectra of 20 as a KBr pellet (left) and in

MeOH solution (right). Concentration of 20 in MeOH was about 5.8 × 10-4 M.

In the solid state spectrum of 20, it can be observed that the νC=O vibrations owing to the

amide carbonyl groups around 1648 cm-1 have the highest intensity. The νCO vibrations owing

to the metal bound carbonyls are not influenced by any vibration originating from the organic

part of this molecule, which clearly shows that the region between 1800-2200 cm-1 really is a

spectroscopic window and that CMIA is a feasible technique.


97

The compounds 14-21 each show a one-electron oxidation at a potential slightly higher than

the Fc/Fc+ transition, as shown in Table 5.3. The difference between the oxidation potential of

the compounds 14-17 and 18-21 is due to the use of different solvents for the electrochemical

investigations. The electrochemical behaviour of 18-21 had to be investigated in DMF,

because these compounds are insoluble in MeCN. In fact, the oxidation potential of the

complexes 14-17 in DMF is also +0.02 V vs. Fc/Fc+. Thus, like observed for the position of

the νCO stretching frequencies, also the oxidation potential is influenced neither by the

chirality of the histidinate ligand, nor by the size of the substituent.


The complexes Mo(His-Nε-C2H4COOH)(η-allyl)(CO)2 with both L-histidine and D-histidine

can be coupled to peptides, both by peptide synthesis methods in solution and SPPS. The

corresponding bioconjugates exhibit good stability towards dioxygen and allow detection by

infrared spectroscopy and electrochemical methods. The discussion of the possible application

of the [Leu]-enkephalin derivatives in assays as well as the future prospects for this type of

chemistry is postponed until section 7.4. In the next two chapters the properties of a different

class of markers containing the fac-Mo(CO)3 unit are investigated.

98

6Spectroscopic properties and reactivity of the complexes

Mo(bpa)(CO)3 and Mo(benzyl-bpa)(CO)3


In the previous chapter, it has been demonstrated that the compounds Mo(His-Nε-

C2H4COOH)(η-allyl)(CO)2 with both L-histidine and D-histidine are very suitable for the

labelling of peptides. These complexes and their bioconjugates are very resistant towards

dioxygen, they are compatible with solid phase peptide synthesis methods and allow detection

by both infrared spectroscopy and electrochemistry.

Further literature studies revealed that in addition to the Mo(η-allyl)(CO)2 group, also the fac-

Mo(CO)3 moiety is a frequently occurring structural unit in the chemistry of low-valent

molybdenum complexes. A large number of Mo(CO)3L3 compounds (with L3 being a

tridentate N-donor ligand) has been structurally characterised [184-201]. Of this class of

compounds, in particular the complexes with the tridentate ligands 1,4,7-triazacyclononane

and hydrido-tris-pyrazolylborate were reported to be very stable [189, 202-204]. However,

these compounds have a disadvantage for use as a marker because they are not easy to

Spectroscopic properties and reactivity of the complexes Mo(bpa)(CO)3 and Mo(benzyl-bpa)(CO)3

99

functionalise with a group via which they could be coupled to biomolecules. In the case of the

hydrido-tris-pyrazolylborate there is no obvious substitution method, whereas mono-

substitution is difficult to achieve with 1,4,7-triazacyclononane (tacn; [9]aneN3). In principle,

one could alkylate 1,4-dimethyl-tacn instead of tacn in order to avoid formation of di and tri-

substituted derivatives, but its synthesis from tacn and the subsequent purification are not

straightforward [205]. In fact, the synthesis of tacn itself is already time-consuming and

requires several steps [206, 207].

Based on these rationalisations, di(2-picolyl)amine (bpa) appeared to be an attractive

candidate ligand for forming marker complexes with the fac-Mo(CO)3 moiety, because it only

contains one NH group that can be derivatised. In addition, the commercial availability of

di(2-picolyl)amine from Aldrich is also advantageous over the multistep synthesis of tacn. My

interest therefore concerned the properties of the complexes of the type Mo(bpa-R)(CO)3

(with R = H and R = alkyl rest) in comparison to those of the tacn and hydrido-tris-

pyrazolylborate analoga. In this chapter, the synthesis and characterisation of Mo(bpa)(CO)3

and Mo(benzyl-bpa)(CO)3 is described, followed by electrochemical and spectroscopic

investigations on these complexes. The results thereof are compared to those of the hydrido-

tris-pyrazolylborate and tacn derivatives and in Chapter 7, the application of the

Mo(bpa)(CO)3 unit as a spectroscopic and electrochemical marker is investigated.

6.2 Synthesis

The ligand N-benzyl-N,N-di(2-picolyl)amine (b-bpa, benzyl-bpa) was synthesised by

alkylation of benzylamine with 2-chloromethyl-pyridine according to a literature procedure

[224]. The complexes Mo(bpa)(CO)3 (22) and Mo(b-bpa)(CO)3 (23) were synthesised by

heating the ligand with Mo(CO)6 in di-n-butyl ether (for 22) or mesitylene (for 23) at 150°C

for one hour, as shown in Scheme 6.1. Complex 22 is isolated as a deep yellow powder,

whereas 23 is obtained as an orange solid. All the characterisation data for 22 and 23 are

consistent with their constitution. These compounds are stable for days in air in the solid state,

but in solution they exhibit higher sensitivity towards dioxygen. Complex 22 is much more

sensitive in solution towards air than 23, with decomposition of 22 taking place within 1-2

hours.

Chapter 6

100

Scheme 6.1 Synthesis and constitution of the molybdenum tricarbonyl complexes

NN

N

R

NN

N

R 22 R = H

23 R = benzyla

MoOC CO CO

+ Mo(CO)6

Reagents and conditions: a) Bu2O (for 22), mesitylene (for 23), heating for one hour at 150 °C

When 22 is reacted with elemental bromine in CHCl3, the orange-brown complex

[Mo(bpa)(CO)3Br]Br3 (24) is obtained (see Scheme 6.2), as concluded from the results from

elemental analysis, infrared spectroscopy, and ESI-positive mass spectrometry. This

compound is stable for approximately one day in the solid state, but in solution

decarbonylation of complex 24 takes place within 5-10 minutes when only traces of water are

present. Due to the reactivity towards water, NMR data of compound 24 could not be

obtained. When [Mo(bpa)(CO)3Br]Br3 is dissolved in MeOH and an aqueous solution of

KPF6 is added, the dinuclear complex [Mo2(bpa)2(O)2(µ-O)2]Br, PF6 (25) invariably

precipitates, which therefore appears to be the thermodynamic sink of the mixture (see

Scheme 6.2). The reactivity of [Mo(bpa)(CO)3Br]Br3 differs from that of

[Mo([9]aneN3)(CO)3Br]Br3 because the latter complex can be crystallised from aqueous

solutions [202].

Scheme 6.2 Reactivity of the complexes Mo(bpa)(CO)3 (22) and Mo(b-bpa)(CO)3 (23)

NN

N

H

MoOC

OC COBr

N

N

N

H Mo

N

N

N

H

Br3

24 25

Br, PF6MoO

O

OO

a

b

[Mo(b-bpa)(CO)3Br]-Br2 oxidation

electrochemical oxidation

X

Mo(bpa)(CO)3

Mo(b-bpa)(CO)3

Reagents and conditions: a) Br2 in CHCl3, stirring at RT for 30 mins; b) MeOH / H2O, addition of

KPF6, standing overnight.


101

When 23 is reacted with Br2 under identical conditions as for the synthesis of 24, initially an

oil forms, but upon removal of the solvent in vacuo, an orange-brown solid is obtained.

Although results from elemental analysis fit well for [Mo(b-bpa)(CO)3Br]Br3, other

characterisation data could not be obtained, because the isolated solid appears to be very

sensitive towards traces of O2 and H2O. However, [Mo(benzyl-bpa)(CO)3Br]+ can be

generated by electrochemical methods (see Scheme 6.2), as concluded from spectro-

electrochemical investigations (vide infra). Below the X-ray crystal structures of 22, 23 and

25 are described, followed by NMR spectroscopic and electrochemical investigations on 22

and 23.


Crystals of both Mo(bpa)(CO)3 (22) and Mo(b-bpa)(CO)3 (23) were obtained by slow

evaporation of a H2O / CH3CN solution under a stream of argon. The unit cell of each

compound was found to consist of two crystallographically inequivalent molecules, which

display slightly different bond lengths and angles. ORTEP diagrams for one of the

independent molecules of both 22 and 23 are depicted in Figure 6.1. Geometrical information

for both structures is summarised in Table 6.1, arranged in such way that related bond lengths

for both molecules of 22 and 23 are in identical rows. Molecule B of 22 and 23 is labelled

consecutively, i.e. N(1) in A corresponds to N(4) in B, etc.

Figure 6.1 ORTEP projections for one of the crystallographically independent molecules

of 22 (left) and 23 (right). Hydrogen atoms have been omitted for clarity.

Chapter 6

102

Table 6.1 Selected bond lengths (Å) for Mo(bpa)(CO)3 (22) and Mo(b-bpa)(CO)3 (23)

Mo(bpa)(CO)3 (22) Mo(b-bpa)(CO)3 (23)

Molecule A Molecule B Molecule A Molecule B

Mo(1)-N(1) 2.229(10) Mo(2)-N(4) 2.249(11) Mo(1)-N(1) 2.272(3) Mo(2)-N(4) 2.250(3)

Mo(1)-N(2) 2.301(10) Mo(2)-N(5) 2.308(11) Mo(1)-N(2) 2.337(2) Mo(2)-N(5) 2.331(2)

Mo(1)-N(3) 2.234(11) Mo(2)-N(6) 2.249(11) Mo(1)-N(3) 2.255(3) Mo(2)-N(6) 2.252(3)

Mo(1)-C(13) 1.92(2) Mo(2)-C(35) 1.92(2) Mo(1)-C(13) 1.933(3) Mo(2)-C(35) 1.938(3)

Mo(1)-C(14) 1.91(2) Mo(2)-C(36) 1.89(2) Mo(1)-C(14) 1.935(2) Mo(2)-C(36) 1.932(3)

Mo(1)-C(15) 1.91(2) Mo(2)-C(37) 1.92(2) Mo(1)-C(15) 1.945(3) Mo(2)-C(37) 1.953(3)

The Mo atoms in both structures are in a distorted octahedral [N3C3] coordination

environment, with the three carbonyls and the (substituted) di(2-picolyl)amine ligand each

occupying a face of the octahedron. This facial arrangement of the carbonyls is also observed

for analogous complexes with various tridentate N-donor ligands [184-201]. In both

molecules of 22 and 23, the Mo-N(amine) bond distance is considerably longer than the Mo-

N(pyridine) bond lengths. Unfortunately the quality of the structure of 22 is not high enough

to allow a detailed discussion of the other bond lengths of this structure. The Mo-donor atom

bond lengths of the molecules A and B of 22 are all within 3σ, and they show a trend of being

slightly shorter than the corresponding bond lengths in 23. However, it must be noted that as a

consequence of the large esd’s of 22, all of the bond lengths of the two molecules are identical

within 3σ with those of 23. Molecule A and B of 23 differ to a small extent. Whereas

molecule A of 23 shows three Mo-C bond lengths that are within 3σ, molecule B displays a

longer one of 1.953(3) (Mo(2)-C(37)) and two slightly shorter ones of 1.938(3) and 1.932(2)

(Mo(2)-C(35) and Mo(2)-C(36), respectively). The bond lengths Mo(2)-C(37) and Mo(2)-

C(36) are not equal within 3σ, whereas the other Mo(2)-C bond lengths are. In molecule A of

23, the bond distance Mo(1)-N(1) is significantly longer than the other Mo-N(pyridine) bond

length. All of the differences between molecules A and B of 23 might be rationalised on the

basis of packing effects.


103

In both structures the deviation from the idealised octahedral symmetry is evident from the C-

Mo-C angles; values less than 90° are observed for all of those. In fact, all structures reported

thus far of fac-Mo(CO)3 complexes with tridentate N-donor ligands display C-Mo-C angles

smaller than 90° [184-201] and, thus, this deviation appears to be a general feature for this

type of compounds.

There is only one example in the literature of a structurally characterised fac-Mo(CO)3

compound with the molybdenum atom being coordinated by two pyridine nitrogen atoms and

one amine nitrogen atom from the same ligand. The potentially tetradentate ligand N, N’-

dimethyl-2, 11-diaza-[3.3](2,6)pyridinophane (see Scheme 6.3) acts as a tridentate ligand in

that compound [184]. The Mo-N(pyridine) bond lengths for the structure of that complex are

approximately 0.04 Å longer than the Mo-N(amine) bond distances. This is in sharp contrast

with the structures 22 and 23, in which the Mo-N(pyridine) bond distances are considerably

shorter than the Mo-N(amine) bond lengths.

Scheme 6.3 Molecular structure of N, N’-dimethyl-2, 11-diaza-[3.3](2,6)pyridinophane

N

N

N

CH3

N

CH3

X-ray quality single crystals of [Mo2(bpa)2(O)2(µ-O)2]Br, PF6 (25) were obtained by slow

evaporation of a MeOH / H2O solution. An ORTEP representation for the cationic part of 25

is shown in Figure 6.2, with relevant bond-distance and bond-angle information summarised

in Table 6.2.

Chapter 6

104

Figure 6.2 Projection for the cationic part in the X-ray crystal structure of 25. Hydrogen

atoms, except amine hydrogen atoms, have been omitted for clarity.

Table 6.2. Selected bond lengths (Å) and angles (°) for [Mo2(bpa)2(O)2(µ-O)2]Br, PF6

Mo(1)-N(1) 2.219(2) Mo(2)-N(6) 2.227(2).

Mo(1)-N(2) 2.313(2) Mo(2)-N(5) 2.294(2)

Mo(1)-N(3) 2.263(2) Mo(2)-N(4) 2.245(2)

Mo(1)-O(2) 1.687(2) Mo(2)-O(4) 1.692(2)

Mo(1)-O(1) 1.934(2) Mo(2)-O(1) 1.921(2)

Mo(1)-O(3) 1.935(2) Mo(2)-O(3) 1.933(2)

Mo(1)-Mo(2) 2.5432(5)

O(1)-Mo(1)-O(2) 110.10(7) O(1)-Mo(2)-O(4) 110.70(7)

O(3)-Mo(1)-O(2) 111.83(7) O(3)-Mo(2)-O(4) 112.29(7)

O(1)-Mo(1)-O(3) 91.43(6) O(1)-Mo(2)-O(3) 91.88(6)

Mo(1)-O(1)-Mo(2) 82.55(6) Mo(2)-O(3)-Mo(1) 82.22(6)

The cationic part of the structure consists of two Mo(bpa)(O) moieties bridged by two oxo

ligands, with the terminal oxo ligands occupying cis-positions relative to each other. The

four-membered ring comprised of the two Mo atoms and the two bridging oxo ligands shows

a puckered conformation, which is observed in all crystal structures containing a syn-

[Mo2(O)2(µ-O)2] core with additional multidentate ligands [208-211]. The short internuclear


105

Mo-Mo distance of 2.5432(5) Å is indicative for a Mo-Mo single bond, being consistent with

the observed diamagnetism of the complex. Each Mo(bpa) unit displays one shorter Mo-

N(pyridine) bond around 2.22 Å and a longer one (Mo(1)-N(3) = 2.263(2) Å; Mo(2)-N(5) =

2.245(2) Å). The Mo-N(amine) bond distances are considerably longer than the Mo-

N(pyridine) bond lengths, similar to what is observed for the crystal structure of the complex

Mo(bpa)(CO)3 (22). The Mo-donor atom bond lengths of one of the Mo(bpa)(O)(µ-O) units

are virtually identical to those of the other, i.e. they are within 3σ or just outside this region.

The largest difference is observed for Mo(1)-N(2) and Mo(2)-N(5), which are 0.007 Å outside

this 3σ range. Also all of the bond angles of both Mo(bpa)(O)(µ-O) moieties are either

identical or very similar.

The bromide ion is involved in hydrogen-bond interactions with the two NH groups, with

Br⋅⋅⋅N contacts of 3.258(4) Å and 3.358(4) Å (Br⋅⋅⋅N(5) and Br⋅⋅⋅N(2), respectively). Two

very similar structures have been reported with an ethyl or methyl-ethyl linker between both

bpa ligands [209], which display Mo-donor atom bond lengths that are nearly identical to

those of 25. Apparently, the hydrogen bonded bromide ion forces the bpa ligands in the same

conformation as the ethyl and methyl-ethyl linkers do. It is expected that this bromide ion is

bound very tightly because apart from hydrogen bond interactions, it is also involved in an

electrostatic interaction with the dinuclear molybdenum complex of charge 2+.

6.4 NMR and electronic spectroscopy

Parts of the 1H NMR spectra of 22 and 23 in DMSO-d6 are depicted in Figure 6.3. Due to

coordination of the three nitrogen atoms, the hydrogen atoms of the picolyl-CH2 groups

become magnetically inequivalent. Imposed by the Cs symmetry of the complexes in solution,

each diastereotopic picolyl hydrogen atom of both 22 and 23 is equivalent to a hydrogen atom

of the other picolyl-CH2 group. Therefore, in the 1H NMR spectrum of 23 these appear as two

doublets at δ = 4.67 ppm and δ = 3.72 ppm with intensity two and a 2J coupling constant of

15.5 Hz (see Figure 6.3). The singlet in between the doublet at δ = 4.67 ppm is owing to the

benzylic-CH2 group. In the 1H NMR spectrum of 22 on the other hand, the appearance of the

picolyl-CH2 hydrogen atom resonances is influenced by the amine NH group. This NH

hydrogen atom resonates at δ = 6.14 ppm with a 3J coupling constant of 6.6 Hz. One of the

pairs of magnetically inequivalent picolyl-CH2 hydrogen atoms resonates as a doublet at δ =

Chapter 6

106

4.06 ppm with a 2J coupling constant of 16.8 Hz, whereas the other pair appears as a doublet

of doublets at δ = 4.33 ppm with a 2J coupling constant of 16.8 Hz and a 3J coupling constant

of 6.6 Hz.

Figure 6.3 Selected regions from the 1H NMR spectra of 22 and23 (400 MHz; DMSO-d6)

The appearance of the coupling constants in 22 can be explained by the Karplus equation

[179]. One of the CH2 hydrogen atoms of each picolyl-group has a dihedral angle ϕ with the

NH hydrogen atom close to 90°, which makes the 3J coupling constant either zero or too

small to observe. On the basis of the other observed 3J coupling constant of 6.6 Hz, it is

expected that the dihedral angle of the NH hydrogen atom with the other hydrogen atom of

each picolyl-group is around 25° or 155°. It must be noted that the Karplus equation does not

discriminate between postive or negative torsion angles, due to the cosinus dependence of this

curve. IUPAC nomenclature, however, gives torsion angles that increase clockwise a positive

sign and those that increase counterclockwise a negative sign.

Selected parts from both crystallographically independent molecules of the X-ray

crystal structure of 22 are depicted in Figure 6.4, together with the labelling scheme of the NH

and CH2 hydrogen atoms. The torsion angles obtained from the crystal structure of 22 are

summarised in Table 6.3. It must be noted that the hydrogen atoms in 22 were not found

crystallographically but were instead placed at geometrically calculated positions using the

options AFIX 23 (for the CH2 hydrogen atoms) and AFIX 13 (for the NH hydrogen atom) in

ShelXTL. Nevertheless, the position of the hydrogen atoms can be regarded as being reliable.

From Table 6.3, it is observed that the absolute value of the HNCH torsion angles is close to

30° for one set and close to 90° for the other set. This is in good agreement with the observed

coupling constants in the 1H NMR spectrum (with Jobserved and Jcalculated being within 1 Hz),

which indicates that the solid state structure does not differ to a large extent from the structure

of 22 in solution.


107

Figure 6.4 Selected parts from both crystallographically independent molecules of the

X-ray crystal structure of 22, visualising the HNCH torsion angles.

Table 6.3 HNCH torsion angles (in °) for both independent molecules of 22[a]

Molecule A Molecule B

H(2)-N(2)-C(6)-H(6A) -94[b] H(5)-N(5)-C(26)-H(26A) +30[c]

H(2)-N(2)-C(6)-H(6B) +33[c] H(5)-N(5)-C(26)-H(26B) -88[b]

H(2)-N(2)-C(7)-H(7A) +85[b] H(5)-N(5)-C(27)-H(27A) +89[b]

H(2)-N(2)-C(7)-H(7B) -32[c] H(5)-N(5)-C(27)-H(27B) -32[c]

[a] IUPAC nomenclature used. Angles that increase clockwise and counterclockwise have a positiveand negative sign, respectively; [b] The Karplus equation (3JHH = 4.22 – 0.5 cos ϕ + 4.5 cos 2ϕ) yieldsa value of 0.2-0.3 Hz for these angles; [c] The Karplus equation yields a value of 5.8-6.0 Hz for theseangles

The 13C NMR spectra of 22 and 23 each show two signals owing to the CO ligands at around

δ = 230 ppm (see experimental section), with the signal at the higher chemical shift having

approximately two times the intensity of the resonance at lower frequency. This demonstrates

that the π-acceptor ability of the pyridine groups is reflected in the 13C NMR spectra. The CO

ligand trans relative to the amine nitrogen atom is more shielded due to increased

backbonding, whereas the CO ligands in a trans-position to the pyridine nitrogen atom have

to compete for backbonding with the pyridine-rings. Also the electronic spectra of 22 and 23

show that the pyridines have good π-acceptor properties, because MLCT-bands with high

Chapter 6

108

molar extinction coefficients are observed, as shown in Figure 6.5. The electronic spectra of

22 and 23 are very similar, each displaying two intense charge transfer bands, which have

their maxima at 327 nm (ε = 6.1 × 103 M-1 cm-1) and 416 nm (ε = 6.9 × 103 M-1 cm-1) for 22

and at 329 nm (ε = 6.3 × 103 M-1 cm-1) and 429 nm (ε = 7.0 × 103 M-1 cm-1) for 23.

Figure 6.5 Electronic spectra of 22 and 23 in MeCN in the range 250-600 nm

Results from 95Mo NMR measurements are displayed in Table 6.4, together with the chemical

shifts for several other fac-Mo(CO)3 compounds with N-donor ligands for comparison. As

can be observed, the chemical shifts of fac-Mo(N3)(CO)3 complexes span a range of about

350 ppm, with compounds 22 and 23 showing resonances in the high frequency region of that

range. The molybdenum nucleus in 23 is deshielded by 89 ppm compared to complex 22.

This downfield shift upon alkylation of the bis-picolylamine nitrogen atom is the reverse

trend as observed for the 1,4,7-triazacyclononane derivatives: Mo(Me3[9]aneN3)(CO)3

resonates 226 ppm upfield compared to Mo([9]aneN3)(CO)3. Generally, tertiary amine

nitrogen atoms are more basic than their secondary counterparts. Thus, it is expected that the

electron-density at the Mo-nucleus increases upon alkylation of the bpa, but the opposite trend

is observed for the chemical shift. However, 95Mo is a quadrupole nucleus with I = 5/2 and

the chemical shift of complexes containing those nuclei does not merely depend on the

electronegativity of the donor atoms. Also steric effects, such as ligand bulkiness and chelate

ring-size, as well as local symmetry effects play an important role [212, 213]. Therefore, the

downfield shift of 23 with respect to 22 might be rationalised on the basis of steric effects. In

fact, the Mo-donor atom bond lengths in the X-ray crystal structure of 22 show a trend of

being somewhat shorter than the corresponding ones of 23.


109

Table 6.4 95Mo NMR data of 22, 23 and several other fac-Mo(N3)(CO)3 compounds.

Complex Solvent δ (in ppm)[a] ∆ν½ (in Hz) Reference

Mo(pyridine)3(CO)3 Pyridine -800 7 [214]

Mo(b-bpa)(CO)3 (23) MeCN -849 47 This work

Mo([9]aneN3)(CO)3 CH2Cl2 -866 20 [215]

Mo(bpa)(CO)3 (22) MeCN -938 36 This work

Mo([12]aneN3)(CO)3 CH2Cl2 -1001 13 [215]

Mo(dien)(CO)3[b] HCONMe2 -1088 70 [216]

Mo(Me3[9]aneN3)(CO)3 DMSO -1092 160 [215]

Mo(MeCN)3(CO)3 MeCN -1114 10 [217]

[Mo(HB(Me2pz)3)(CO)3]- HCONMe2 -1149 80 [215]

[a] Chemical shifts reported versus 2M Na2MoO4 in D2O at an apparent pH 11;[b] dien = diethylenetriamine

6.5 Electrochemistry

Both 22 and 23 show a reversible one-electron oxidation with NBu4PF6 as supporting

electrolyte at slightly negative potentials compared to the Fc/Fc+ couple, as displayed in Table

6.5. The redox potential appears to be solvent dependent, with the lower oxidation potential

measured in CH2Cl2. From Table 6.5, it is observed that the oxidation potentials of 22 and 23

are very similar to those of the tacn and Me3-tacn complexes. The hydrido-tris-(3,5-

dimethylpyrazolyl)borate anionic analogue [Mo(HB(Me2pz)3)(CO)3]- displays a one-electron

oxidation at a slightly higher potential than the tacn, bpa and benzyl-bpa complexes.

However, the exact redox potential of the [Mo(HB(Me2pz)3)(CO)3]- ion is not clear because

two different studies reported oxidation potentials that differ by more than 0.2 V [218, 219].

Chapter 6

110

Table 6.5 Reversible one-electron oxidations of fac-Mo(CO)3 compounds with tridentate

N-donor ligands[a]

Complex E½ (MeCN)[b] E½ (CH2Cl2)[b]

Mo(bpa)(CO)3 (22) -0.26[c] -0.33[c]

Mo(b-bpa)(CO)3 (23) -0.22[c] -0.28[c]

Mo([9]aneN3)(CO)3 -0.37[c] -0.34[c]

Mo(Me3[9]aneN3)(CO)3 -0.25[d] -0.28[c]

[Mo(HB(Me2pz)3)(CO)3]- -0.14[e, f] -0.24[e, f]

[Mo(HB(Me2pz)3)(CO)3]- 0.08[g, h] 0.00[g, h]

[a] 0.1 M NBu4PF6 as supporting electrolyte; [b] vs. Fc/Fc+; in V; [c] this work; [d] from ref. [203], [e] 0.1

M NBu4ClO4 as supporting electrolyte; [f] from ref. [218]; [g] 0.1 M NBu4BF4 as supporting

electrolyte; [h] from ref. [219]

With NBu4Br as the supporting electrolyte, the electrochemical behaviour of 22 and 23 in

CH2Cl2 is completely different. Irreversible oxidations occur at –0.37 V vs Fc/Fc+ (for 22)

and at –0.28 V vs Fc/Fc+ (for 23) which is attributed to formation of the seven-coordinated

complexes [Mo(bpa)(CO)3Br]+ and [Mo(benzyl-bpa)(CO)3Br]+. Controlled potential

coulometry established the identity of the transition for both compounds to be a two-electron

oxidation. Formation of seven coordinated Mo(II) compounds has also been reported for

Mo(CO)3 complexes with two different tacn derivatives [192, 203].

After generation of the seven coordinated species [Mo(bpa)(CO)3Br]+ and

[Mo(benzyl-bpa)(CO)3Br]+, a reduction peak is observed at a more negative potential, namely

at –1.20 V vs. Fc/Fc+ for [Mo(bpa)(CO)3Br]+ and –0.95 V vs. Fc/Fc+ for Mo(benzyl-

bpa)(CO)3]+, as shown in Figure 6.6 for complex 23. The peak at –0.95 V vs. Fc/Fc+ only

appears during the second cyclic voltammogram cycle, which demonstrates that the species

that forms upon oxidation of the starting material is reduced at this potential. By controlled

potential coulometry, the number of transferred electrons for this reduction process is shown

to be two, yielding again the neutral compounds 22 and 23. The bpa-derivatives differ from

the tacn analogues in this behaviour, because the corresponding reduction peak of the seven

coordinated species for the tacn complexes is not observed in the range from –2.0 V to + 2.0

V [192, 203].


111

0,5 0,0 -0,5 -1,0 -1,5 -2,0

-15

-10

-5

0

5

I (in

µA)

E vs Fc/Fc+ (in V)

Figure 6.6 Cyclic voltammogram of 23 in CH2Cl2 in the presence of NBu4Br; two

consecutive cycles starting from –0.5 V moving first to –2.0 V. Note that the

reduction peak at –0.95 V appears only in the second cycle.

Below more evidence for the assigned electrochemical transitions is provided by UV-Vis and

infrared spectro-electrochemical investigations on 23 and its derivatives.

6.6 Spectro-electrochemistry

During the electrochemical investigations, complex 23 appeared to be much more stable than

22, which may be attributed to steric effects: the benzyl derivative shields the Mo-atom more

efficiently from the chemical environment. UV-Vis spectro-electrochemical investigations

could only be done on 23, because 22+ and [Mo(bpa)(CO)3Br]+ are very reactive in solution

towards traces of water and react in part further during the coulometry (several minutes),

probably to yield oxo-species.

The changes in the electronic spectrum of complex 23 were monitored during its oxidation in

CH2Cl2 with both NBu4PF6 and NBu4Br as the supporting electrolyte, as depicted in Figure

6.7. During the oxidation both in the presence of PF6 and bromide anions, the MLCT-bands at

429 nm (ε = 7.0 × 103 M-1 cm-1) and 329 nm (ε = 6.3 × 103 M-1 cm-1) gradually disappear,

caused by the decreased electron-availability at the Mo-ion in the Mo(I) and Mo(II) forms.

With NBu4PF6 as supporting electrolyte, a new shoulder appears at around 350 nm, with a

Chapter 6

112

molar extinction coefficient around 4.8 × 103 M-1 cm-1 and an isobestic point is present at 320

nm. In the presence of bromide anions, a new shoulder appears at 370 nm which has a molar

extinction coefficient of around 1.4 × 103 M-1 cm-1, and an isobestic point is present at 307

nm. After the oxidations, the formed complexes can be re-reduced in this way yielding the

UV-Vis spectrum of 23, without any significant loss in absorbance.

Figure 6.7 Changes in the electronic spectrum of 23 during coulometry in CH2Cl2 at 0° C.

Left: in the presence of NBu4PF6; right: in the presence of NBu4Br

Infrared spectra of 22 and 23 in THF consist of three νCO vibrations at around 1905, 1780 and

1790 cm-1, which is consistent with Cs symmetry, for which three νCO vibrations are expected

(A' + 2 A''). In the case of Mo(Me3[9]aneN3)(CO)3 the three nitrogen atoms are identical,

resulting in a complex with C3v symmetry, for which only two infrared active vibrations are

expected (A + E). The differences between the infrared spectra in THF of 23 and

Mo(Me3[9]aneN3)(CO)3 are depicted in Figure 6.8.

Figure 6.8 Comparison between the solution infrared spectra in THF of 23 (left)

and Mo(Me3[9]aneN3)(CO)3 (right)

2000 1900 1800 1700

100

90

Tran

smis

sion

(%)

90

100

2000 1900 1800 1700cm-1 cm-1

Tran

smis

sion

(%)

E

A

A

A

'

"A"


113

Infrared spectra of [Mo(bpa)(CO)Br]Br3 (24) were measured between normal NaCl windows

in THF, whereas infrared spectra of 22+, 23+ and [Mo(b-bpa)(CO)3Br]+ were recorded after

generating these complexes electrochemically in an OTTLE-cell. The infrared spectra of 22+,

23+ were obtained in THF, but the infrared spectrum of [Mo(b-bpa)(CO)3Br]+ had to be

recorded in CH3CN due to insolubility of the supporting electrolyte NBu4Br in THF. As a

representative example, the infrared spectral changes during the oxidation of 23 to 23+ are

depicted in Figure 6.9.

Figure 6.9 Changes in the infrared spectrum during the oxidation of 23 to 23+ in THF,

performed in an OTTLE-cell

The carbonyl stretching vibrations in the infrared spectra of the compounds in this study are

summarised in Table 6.6, together with values for analogous tacn and hydrido-tris-

pyrazolylborate compounds. As can be observed, the νCO stretcheing vibrations of 22 and 23,

their 17-electron derivatives 22+ and 23+ and the seven coordinated complexes 24 and [Mo(b-

bpa)(CO)3Br]+ show similar trends as their tacn and hydrido-tris-pyrozolylborate analogues.

The electron-availability of the molybdenum ion in these species decreases upon increase of

the metal ion’s oxidation state, resulting in a shift of the νCO vibrations to higher

wavenumbers.

2100 2000 1900 18000,00

0,02

0,04

0,06

0,08

0,10

0,12

0,14

Abso

rban

ce

cm-1

⇐

⇐

⇐⇐

⇐

Chapter 6

114

Table 6.6 Selected vibrational data for 22, 23, 22+, 23+, 24, [Mo(b-bpa)(CO)3Br]+

and related compounds.

Complex Solvent ν(CO) in cm-1 Reference

Mo(0) compounds

Mo(bpa)(CO)3 (22) THF 1904, 1790, 1779[a] This work

Mo(b-bpa)(CO)3 (23) THF 1906, 1794, 1782[a] This work

Mo([9]aneN3)(CO)3 THF 1901, 1766[b] This work

Mo(Me3[9]aneN3)(CO)3 THF 1907, 1773[b] This work

Mo(benzyl3[9]aneN3)(CO)3 CH2Cl2 1907, 1768[b] [220]

[Mo(HB(pz)3)(CO)3]- CH3CN 1897, 1761[b] [204]

[Mo(HB(Me2pz)3)(CO)3]- CH3CN 1891, 1751[b] [204]

[Mo(HB(Me2pz)3)(CO)3]- CH2Cl2 1884, 1742[b] [218]

Mo(I) compounds

[Mo(bpa)(CO)3]+ (22+) THF 2015, 1910, 1859[c] This work

[Mo(b-bpa)(CO)3]+ (23+) THF 2018, 1922, 1864[c] This work

[Mo(benzyl3[9]aneN3)(CO)3]+ CH2Cl2 2017, 1909, 1873 [220]

Mo(HB(pz)3)(CO)3 CH2Cl2 2010, 1885 (br) [221]

Mo(HB(Me2pz)3)(CO)3 CH2Cl2 1998, 1864 (br) [222]

Mo(HB(Me2pz)3)(CO)3 CH2Cl2 2001, 1860 (br) [218]

Mo(II) compounds

[Mo(bpa)(CO)3Br]+ THF 2055, 1989, 1935 This work

[Mo(benzyl-bpa)(CO)3Br]+ CH3CN 2055, 1980, 1957[c] This work

[Mo([9]aneN3)(CO)3Br]+ THF 2042, 1964, 1928 This work

Mo(HB(pz)3)(CO)3Br CH2Cl2 2053, 1980, 1938 [223]

[a] Complex has Cs symmetry; [b] Complex has C3v symmetry; [c] Complex generated

electrochemically in an OTTLE-cell


115

6.7 EPR spectroscopy

The electrochemical and spectro-electrochemical investigations on 23 showed good stability

of the one-electron oxidised form 23+. Because compound 23+ has an odd number of electrons

(a 17-electron organometallic complex), it was subjected to EPR spectroscopic investigations.

The X-band EPR spectrum of 23+, generated by coulometry, as a frozen CH2Cl2 solution is

depicted in Figure 6.10, together with the resulting spectrum from the simulation-procedure.

Figure 6.10 X-band EPR spectrum of 22+, generated by coulometry, in frozen CH2Cl2

solution (0.1 M NBu4PF6) and its simulated spectrum. T= 50 K; Power: 0.2

mW; Frequency 9.46494 GHz. Modulation: 1.25 mT.

Although the fit is not perfect, it clearly shows the rhombic nature of the S = ½ spectrum,

with simulated g-values: gx = 2.238, gy = 2.131 and gz = 1.954. These g-values are slightly

higher than those reported for [Mo(benzyl3[9]aneN3)(CO)3]+ as a frozen DMF solution, which

are : gx = 2.188, gy = 2.040 and gz = 1.992 [220].

The discrepancy between the simulated and the experimental spectrum of 23+ might be due to

the presence of small amounts of paramagnetic decomposition products. In fact, satisfying

EPR spectra of the related compound 22+ could not be obtained, because the EPR spectra

contained many additional resonances. The spectro-electrochemical investigations on 22

already showed that 22+ reacts in part further during the time required for the coulometry

(several minutes).

Chapter 6

116


The experimental and spectroscopic results described in the previous sections showed that the

properties of the complexes Mo(bpa)(CO)3 (22) and Mo(benzyl-bpa)(CO)3 (23) are similar to

those of the tacn and hydrido-tris-pyrazolylborate analogues. However, small differences in

reactivity and electrochemical behaviour between these three types of compounds were

observed. Compounds 22 and 23 show reversible one-electron oxidations with NBu4PF6 as

supporting electrolyte at potentials similar to those for the tacn derivatives, whereas the

Mo0/Mo1 transition for [Mo(HB(Me2pz)3(CO)3]- occurs at a slightly higher potential. With

NBu4Br as supporting electrolyte, 22 and 23 each show an irreversible two-electron oxidation,

yielding the complexes [Mo(bpa)(CO)3Br]+ and [Mo(benzyl-bpa)(CO)3Br]+, respectively.

These Mo(II) complexes can be re-reduced at a potential around –1.0 V, yielding again

complexes 22 and 23. For the complexes Mo(R3-tacn)(CO)3 (R = Me or benzyl),

electrochemical formation of [Mo(R3-tacn)(CO)3Br]+ was reported, but the corresponding

reduction to yield the Mo0 complexes was not observed between +2.0 V and –2.0 V [192,

203]. A difference in reactivity between [Mo(bpa)(CO)3Br]+ and [Mo(tacn)(CO)3Br]+ is that

the latter compound can be crystallised from aerobic aqueous solutions [202], whereas the

former reacts further to yield the dinuclear Mo(V)-complex [Mo2(bpa)2(O)2(µ-O)2]Br, PF6.

Another difference is observed in the solution infrared spectra of 22 and 23 with respect to

those of the analogous tacn and hydrido-tris-pyrazolylborate complexes because 22 and 23

have point-group symmetry Cs, whereas the tacn and hydrido-tris-pyrazolylborate complexes

are of point-group symmetry C3v. The degeneracy of the vibrations of E symmetry in the tacn

and hydrido-tris-pyrazolylborate complexes is lifted in 22 and 23.

Complexes 22 and 23 show infrared active vibrations and in particular 23 exhibits good

stability during the electrochemical investigations. In the next chapter, the use of the

Mo(bpa)(CO)3 unit as an infrared spectroscopic and electrochemical marker for peptides is

investigated.

117

7The Mo(bpa)(CO)3 unit as a marker


In the previous chapter, it has been shown that the properties of the compounds

Mo(bpa)(CO)3 (22) and Mo(benzyl-bpa)(CO)3 (23) are very similar to those of analogous

complexes with the ligands 1,4,7-triazacyclononane and hydrido-tris-(3,5-

dimethylpyrazolyl)borate. Because complexes 22 and 23 display reversible electrochemical

transitions and contain carbonyl ligands that show infrared active vibrations, the unit Mo(bpa-

R)(CO)3 can serve as an effective electrochemical and infrared spectroscopic marker. In this

chapter, the substitution possibilities of the bpa ligand are investigated and a mild method for

introduction of the fac-Mo(CO)3 moiety to di(2-picolyl)amine derivatives of varying

complexity is presented. In the last section a résumé of the labelling of [Leu]-enkephalin with

the molybdenum carbonyl complexes is presented and the future prospects for this type of

chemistry are discussed.

Chapter 7

118

7.2 Synthesis in solution

My interest concerned substitution of the amine hydrogen atom of di(2-picolyl)amine by an

alkyl group containing an acid functionality in order to make peptide coupling reactions of

either the ligand or the corresponding fac-Mo(CO)3 complex possible. On first sight the

compound N-acetic acid-N,N-di(2-picolyl)amine, which was first synthesised by Que and co-

workers via alkylation of bpa with bromo-acetic acid [225], appears to be an attractive

candidate derivative. However, in a recent paper it was reported that this compound cannot be

coupled to amino acids in high yields via standard peptide coupling reactions [226]. The

reason for this might be that a pyridine nitrogen atom reacts with the activated carboxylic

acid, which is due to steric constraints always in close proximity to the pyridine-rings, leading

to a ring-closure and formation of a six-membered ring [227].

In the case of a different bpa derivative, namely N-4-benzylic acid-N,N-di(2-picolyl)amine,

this type of reaction is expected not to occur because the activated carboxylic acid is sterically

not accessible for the pyridine nitrogen atoms. In addition, it is likely that the presence of a

bulky aromatic ring yields a fac-Mo(CO)3 complex with a similar stability towards dioxygen

as 23. The first thing tried to obtain this compound was the reaction of bpa with 4-

bromomethyl benzoic acid under identical conditions as Que used for the synthesis of his

acetic acid derivative [225], i.e. addition of one equivalent of NEt3 in absolute ethanol and

refluxing overnight. The next day, the reaction-mixture was cooled down to 0°C, and the

triethyl ammonium salts that formed were removed by filtration. In contrast to the synthesis of

Que’s glycine derivative, only an additional amount of triethyl ammonium salts precipitated

upon cooling the mixture further to -20°C. Probably N-4-benzylic acid-di(2-picolyl)amine is

very soluble in EtOH and does not precipitate from that solvent.

Therefore, the idea of direct introduction of the carboxylic acid moiety was omitted, and

instead attempts were made to obtain N-4-carboxymethyl-benzyl-N, N-di(2-picolyl)amine in

pure form and then hydrolyse the ester moiety in a later stage. A convenient synthesis was

found to be the reaction of bpa with one equivalent of 4-bromomethyl benzoic acid methyl

ester in THF in the presence of stoichiometric amounts of NEt3, as shown in Scheme 7.1.

The Mo(bpa)(CO)3 unit as a marker

119

After heating at reflux for one hour, and subsequently allowing the mixture to cool down to

room temperature, the formed HNEt3Br was removed by filtration. Then, the solvent was

removed under reduced pressure and the oily residue treated with Et2O, followed by filtration

to remove a red-brown sticky solid. After removal of the solvent, the compound N-(4-

carboxymethyl)benzyl-N, N-di(2-picolyl)amine (26) was obtained in pure form as a light-

orange oil in 81% yield.

Scheme 7.1 Synthesis of a Mo(bpa)(CO)3 derivative containing a methyl ester and

investigation of its reactivity

NN

N

CO2Me

CO2Me

NN

NMoOC CO CO

CO2Me

CH2Br

NN

N

bpa + ba

MoOC CO CO

CO2Na

cimpure

26 27

Reagents and conditions: a) one equivalent of NEt3 in THF, reflux for 1 hr; b) Mo(CO)6 in mesitylene,

heating at 150° C for 1 hr; c) excess NaOH in MeOH / H2O (95 / 5), reflux for 3 hrs.

Subsequently, compound 26 was reacted with Mo(CO)6 in a similar manner as for the

synthesis of 23, leading to formation of Mo(4-CO2Me-benzyl-bpa)(CO)3 (27), as depicted in

Scheme 7.1. All the characterisation data are consistent with the constitution (see

experimental part) and the electrochemical behaviour as well as the 95Mo NMR spectrum is

identical to that of 23. This shows that introduction of a methyl ester substituent on the para

position of the phenyl ring does not influence the properties of the complex to a large extent,

or in other words: the molybdenum atom does not experience a different chemical

environment.

Chapter 7

120

Next, it was attempted to hydrolyse the ester moiety of 27 in order to obtain either the

corresponding carboxylate or carboxylic acid derivative, in this way making coupling

reactions of this complex with biomolecules possible. However, 27 is insoluble in H2O and

only slightly soluble in MeOH / H2O or 1,4-dioxane / H2O mixtures. Heating compound 27 at

reflux in a 95 / 5 MeOH / H2O mixture containing excess NaOH for three hours results in a

clear orange-red solution. Upon allowing the mixture to cool down to room temperature, an

orange precipitate forms, which was found to be the sodium carboxylate derivative of 27.

However, many additional signals are present in the 1H NMR spectrum as well as in the 13C

NMR spectrum, indicating that the compound is not pure. Probably, the reaction conditions

required for dissolution of the complex and hydrolysis of the methyl ester are too harsh.

Hence, the idea of tethering the complex as such to biomolecules similar to the complexes

discussed in Chapter 3 and Chapter 5 was omitted. Instead, another idea appeared very

attractive, namely coupling of the ligand itself to biomolecules prior to formation of the

tricarbonyl complex.

Because direct alkylation of bpa with 4-bromomethyl benzoic acid did not yield N-4-benzylic

acid-di(2-picolyl)amine, a different strategy was chosen to obtain this compound, which

consists of hydrolysis of the methyl ester moiety of 26. Compound 26 was stirred for two

hours in a MeOH / H2O mixture containing excess NaOH, as shown in Scheme 7.2, leading to

the carboxylate derivative in solution. The only successful work-up to obtain the acid was

adjustment of the pH to 7 with diluted hydrochloric acid, followed by removal of the solvent

in vacuo. The remaining salt / acid mixture was treated with CHCl3, followed by filtration to

remove the salts. Upon removal of the solvent under reduced pressure, a light-orange sticky

oil was obtained, which was found to be the desired compound in impure form. By addition of

MeCN to this oil and stirring vigorously, the compound N-4-benzylic acid-N,N-di(2-

picolyl)amine (28) separated as a white solid.

Subsequently, investigations were made whether compound 28 could be tethered to

biomolecules via peptide synthesis methods in solution. The reaction of 28 with either

phenylalanine methyl ester or the dipeptide H-Ala-Phe-OMe in MeCN for 30 minutes,

employing TBTU as the coupling reagent, yielded the bpa biomolecule derivatives 29 and 30

in 89% and 92% yield, respectively, as shown in Scheme 7.2.


121

Scheme 7.2 Synthesis of the 4-benzylic acid derivative of bpa (compound 28) and

formation of an amino acid and dipeptide conjugate

2

2829 Phe

30 Ala-PheN

N

b

CO OH

2N

N

CO AA-OMe

26 + H-AA-OMeAAa

Reagents and conditions: a) NaOH in MeOH / H2O; 2 hrs at RT; addition of diluted HCl; b) addition

of TBTU in MeCN / NEt3, stirring for 45 mins at RT.

At this stage, my interest concerned transforming compounds 29 and 30 into their

corresponding Mo(CO)3 derivatives. This method should be preferably possible under mild

conditions, in order to make it suitable for the labelling of larger biomolecules. Therefore,

reacting 29 and 30 at 150°C with Mo(CO)6 similarly to the synthesis of the parent complexes

22 and 23 did not appear to be very attractive. Considering the thermodynamic stability of a

tridentate ligand over three monodentate ligands and the kinetic lability of nitrile ligands, it

appeared good to “pre-activate” Mo(CO)6 by transforming it into Mo(CO)3(EtCN)3 according

to a literature procedure [228].

To transform 29 and 30 into their fac-Mo(CO)3 derivatives, a stoichiometric amount of

Mo(CO)3(EtCN)3 in THF was mixed with a THF solution of the biomolecule-bpa derivative

(see Scheme 7.3), and after stirring for 5-10 minutes at room temperature, an orange

precipitate formed. By isolation of the precipitate and drying it in vacuo, the compounds 31

and 32 were obtained in pure form in approximately 80% yield.

Chapter 7

122

Scheme 7.3 Synthesis of phenylalanine and dipeptide conjugates containing the

Mo(bpa)(CO)3 unit

NN

CO AA-OMe

NN

NMoOC CO

CO2

AA-OMe

O

+ Mo(CO)3(NCEt)331 Phe

32 Ala-Phe

AAa

Reagents and conditions: a) THF; stirring for 10 mins

The results from elemental analysis on 31 were invariably low but all the other

characterisation data are consistent with the proposed constitution. First of all, the infrared

spectra as a KBr pellet show that 31 and 32 formed by the presence of the νCO vibrations at

1900, 1778 and 1760 cm-1. Also the 1H NMR spectra clearly indicate that the nitrogen atoms

are coordinated, because the hydrogen atoms of the picolyl CH2 groups have become

magnetically inequivalent, similar to what is observed for complex 23. In addition, resonances

owing to the CO ligands around 230 ppm can be observed in the 13C NMR spectra. All of the

signals in the 1H and 13C NMR spectra are assigned unambiguously and, actually, these NMR

spectra are very similar to those of 23, but differ in the presence of additional resonances

owing to the biomolecule. One interesting feature of the 1H NMR spectra of these

bioconjugates is that the picolyl-CH2 hydrogen atoms show in addition to the 2J coupling

constant of about 15.5 Hz also very small couplings of 1.5 Hz (for 31 in CD3CN) and 3.4 Hz

(for 32 in DMSO-d6). This is likely to be a 4J coupling constant resulting from the picolyl-

hydrogen atom of the other CH2-group, but these couplings were not observed in 23 and 27.

In the 95Mo NMR spectra, 31 and 32 each show a resonance at –849 ppm, identical to the

chemical shift for 23 and 27.


123

7.3 solid phase synthesis

Because compounds 31 and 32 exhibit stability towards dioxygen comparable to 23, my

interest concerned the synthesis the [Leu]-enkephalin-bpa derivative, and subsequent

investigation whether the Mo(CO)3 moiety could be introduced in a similar way. The tyrosine

OH group might disturb the reaction because it also constitutes a potential donor group.

The reaction of [Leu]-enkephalin on the resin, including 2-ClTrt and Fmoc

deprotection, was already discussed in Chapter 2. The resin-bound enkephalin was reacted

with a five-fold excess of the acid 28 overnight, employing TBTU as the coupling reagent.

After washing the resin, it was reacted with a saturated NH3 solution in MeOH for 48 hours at

room temperature in order to cleave the [Leu]-enkephalin derivative from the resin.

Subsequently the resin was removed by filtration and the filtrate concentrated to dryness in

vacuo. Purification by preparative HPLC afforded approximately 130 mg of the pure

enkephalin derivative 33 depicted in Scheme 7.4.

Scheme 7.4 Synthesis and constitution of the [Leu]-enkephalin derivatives 33 and 34

NN

CO Enk-NH2

NN

NMoOC CO

CO

Enk-NH2

O

NNH

OHN

O

NH

O

N

O

OH

2

NH2

O

a

Enk-NH2 =

+ Mo(CO)3(NCEt)3

33 34

Reagents and conditions: a) MeOH, stirring for 10 minutes at RT.

Chapter 7

124

The ESI-positive mass spectrum shows peaks corresponding to [M+H]+, [M+Li]+ and

[M+Na]+, and a high-resolution determination on the [M+Li]+ peak yielded an exact mass for

this fragment of 876.4380, in excellent agreement with the calculated mass (876.4384).

Moreover, the 1H NMR and 13C NMR spectra are also consistent with the proposed

constitution. In the 1H NMR spectrum in DMSO-d6, several resonances are broadened, which

is likely caused by the presence of five different hydrogen-bonding sites for the NH hydrogen

atoms (NH can form hydrogen bonds to amide C=O, pyridine-N, tertiary amine N, DMSO

and H2O), thus resulting in inequivalent chemical environments. In contrast, in the 1H NMR

spectrum in CD3OD, the resonances are narrow and well-defined, but of course the

resonances owing to amide NH hydrogen atoms are not detected because of H / D exchange.

Next the introduction of the Mo(CO)3 moiety via a similar reaction as for 31 and 32 was

attempted. However, MeOH was rather used in this case because the peptide is insoluble in

THF. After reacting 33 in MeOH with stoichiometric amounts of Mo(CO)3(NCEt)3 for 10

minutes at room temperature, as depicted in Scheme 7.4, an orange precipitate formed.

Instead of isolating this precipitate by filtration, the solution was evaporated to dryness in

vacuo to avoid losses, affording compound 34 (see Scheme 7.4) in quantitative yield.

The first evidence that 34 formed was obtained from the KBr infrared spectrum by the

presence of the νCO vibrations at 1761, 1787 and 1896 cm-1. Furthermore, the ESI-positive

mass spectrum is consistent with the proposed constitution, but a high-resolution mass

spectrum could not be obtained due to insufficient intensity of the [M+H]+ and [M+Na]+

peaks. Finally, the 1H and 13C NMR spectra clearly show that 34 formed for similar reasons as

discussed for 31 and 32 and all the resonances are easily assigned (see experimental section).

7.4 Infrared spectroscopic and electrochemical investigations

Selected vibrational and electrochemical data for the compounds 27, 31, 32, and 34 are

summarised in Table 7.1. As can be observed, the size of the biomolecule attached to the

Mo(benzyl-bpa)(CO)3 unit does not influence the electrochemical and infrared spectroscopic

properties of the bioconjugates significantly. This is similar to what has been observed for the


125

bioconjugates presented in Chapter 5. Unfortunately, compound 34 is insoluble in water and

also not soluble in organic solvents other than DMSO and DMF. For the labelling of peptides

which do not contain amino acids with hydrophilic side chains (such as glutamic acid and

lysine), markers derived from the complex Mo(His)(η-allyl)(CO)2 are much more suitable.

Table 7.1 Selected infrared spectroscopic and electrochemical data for 27, 31, 32 and 34

Compound νCO (KBr)[a] νCO (THF)[a] E½[c]

27 1761, 1787, 1896 1787, 1794, 1907 -0.26[d]

31 1759[f], 1779, 1900 1782, 1793, 1906 -0.27[d]

32 1761[f], 1778, 1900 1782, 1793, 1906 -0.27[d]

34 1761, 1787, 1896 insoluble -0.22[e]

[a] in cm-1; [b] in V, vs Fc/Fc+; [c] 0.1 M NBu4PF6 as supporting electrolyte; [d] in CH2Cl2; [e] in DMF; [f] exact position of the νCO vibrations influenced by the νC=O vibration.

7.5 Concluding remarks on the labelling of [Leu]-enkephalin and future outlook

The results presented thus far in this thesis show that spectroscopic labels based on the

complexes Mo(His)(η-allyl)(CO)2 and Mo(benzyl-bpa)(CO)3 can be introduced into a peptide

via SPPS. The derivatives of the complex Mo(His)(η-allyl)(CO)2 discussed in Chapter 5

exhibit slightly better stability towards dioxygen than the compounds presented in this

chapter, but both classes of complexes allow detection by infrared spectroscopy and

electrochemical methods. The markers differ in the way they can be introduced into a peptide.

The Mo(His)(η-allyl)(CO)2 unit is introduced by reacting the compound Mo(His-Nε-

C2H4COOH)(η-allyl)(CO)2 with the resin-bound peptide and in the next step the complex-

peptide conjugate is cleaved from the resin. The introduction of the Mo(benzyl-bpa)(CO)3

moiety is achieved in a different way by first tethering the benzyl-bpa derivative to the

peptide, followed by cleavage the bpa-peptide conjugate from the resin. In a next step, the

fac-Mo(CO)3 moiety can be introduced under mild conditions by reacting the ligand-peptide

conjugate with Mo(CO)3(NCEt)3.

Chapter 7

126

It should be noted that when peptides labelled with the Mo(benzyl-bpa)(CO)3 moiety are

selected for assays, the peptide should have several hydrophilic residues in order to increase

the solubility of the conjugate in water / organic solvent mixtures. The [Leu]-enkephalin

derivatives of the complex Mo(His)(η-allyl)(CO)2 are very soluble in alcohol / water

mixtures, and are even slightly soluble in H2O.

Future research is necessary to investigate whether the [Leu]-enkephalin derivatives

synthesised in this dissertation have affinity for the corresponding receptor. If not, or if the

affinity is decreased significantly, several modifications of the peptide require investigation.

For example, it might be that the free carboxylic acid instead of the acid amide is

indispensable for receptor recognition. Other cleavage methods could be chosen in that case,

for example cleaving the bioconjugate from the resin in the form of a methyl ester, which is

possible with HMBA-AM resins [114]. This methyl ester could be hydrolysed in a next step

under basic conditions, and following acidification to neutrality, the [Leu]-enkephalin

derivative with the C-terminus in the form of a carboxylic acid is obtained. If in addition to

the carboxylic acid also the free amine is required for receptor recognition, one could

substitute the leucine residue for a Nε-Boc-protected lysine. After the peptide synthesis (with

the peptide still bound to the resin, and containing both the 2-ClTrt and the Fmoc groups on

the Tyr-residue), the Boc protecting group and the 2Cl-Trt group can be removed by the

reaction with CF3COOH. Subsequently, the metal complex (or the benzylic acid-bpa

derivative) can be coupled to the lysine-NεH2 moiety, by using TBTU and dipea. At this stage

the Fmoc group can be removed and after cleavage of the peptide-metal complex conjugate

from the resin, a pentapeptide with a similarity to the [Met]-enkephalin is obtained. A

substituted lysine is present instead of the methionine thioether-side chain, but this might not

disturb the binding to the corresponding receptor to a large extent. All the above mentioned

ideas are of course speculative, but they illustrate the kind of chemistry that still might be

required to develop enkephalin labelled peptides that display high affinity for the

corresponding opioid receptor.

Of course, the labelling of peptides with the molybdenum carbonyl complexes is not restricted

to enkephalin derivatives and, in fact, various small peptides are known to have a neurological

action in organisms. The methods developed in this Thesis show that labelled peptides are

accessible in highly pure form via peptide synthesis on a solid support, which opens new

ways for scientific investigation of these peptides and their receptor interaction.

127

8Ferrocene and cobaltocenium conjugates of amino acids

and dipeptides, a hydrogen bonding study

8.1 Introduction

This chapter is not related to the previous ones, in dealing with the introduction of

spectroscopic handles on peptides in the form of organometallic complexes. Instead, amino

acids and dipeptides are bound in a covalent fashion to ferrocene and the cobaltocenium

moiety because the resulting conjugates can be regarded as small mimics for the hydrogen

bond interactions that are present in proteins. In this section, first an overview of the literature

is given, followed by the presentation of the ideas that attracted my attention and the strategy

to investigate these.

Within one decade after the discovery of ferrocene by Kealy and Pauson in 1951 [26], the first

amino acid and dipeptide ferrocene conjugates were reported by Schögl [229]. These results

remained unnoticed to most chemists, because they were published in a rather obscure journal

in german language. About five years ago, the properties of this class of compounds were re-

investigated and several reports of amino acids bound covalently to ferrocene via an amide

linkage appeared more or less simultaneously in the literature, but none of these reports

referenced Schögl’s paper [230-232]. Until now various ferrocene amino acid derivatives

have been characterised by X-ray crystallography [46, 231-235]. In addition, three X-ray

crystal structures of the ferrocene moiety coupled to a dipeptide have been reported [235-

236], and also two tripeptide derivatives and a tetrapeptide conjugate have been structurally

characterised [235].

Chapter 8

128

Because ferrocene contains two cyclopentadienyl rings, another possibility is introduction of a

substituent on each cyclopentadienyl ring. In 1996, it was reported that compounds of the

general type Fe(Cp-C(O)-AA-OMe)2 (with AA = amino acid) have an ordered conformation

in organic solvents such as CH2Cl2 and CHCl3 [230]. This ordered structure, visualised in

Scheme 8.1, is comprised of two symmetrically equivalent hydrogen bonds between the

amide NH and the methyl ester carbonyl moiety of another strand. Because the inter-ring

separation of ferrocene of about 3.3 Å [238] is close to the N⋅⋅⋅O distance in β-pleated sheets,

this ordered conformation can be considered a mimic for a small part of this secondary

structural element.

Scheme 8.1 Visualisation of the ordered conformation of ferrocenes bearing an amino acid

substituent on each cyclopentadienyl ring

N

O

HN

O

O

MeO

O

OMe

H

R

R

Fe

In the past few years, larger substituents in the form of dipeptides have been introduced on

each Cp-ring, and these bioconjugates were shown to exhibit a similar ordered conformation

in solution as well as in the solid state [236, 237, 239, 240]. The only difference between

these dipeptide derivatives and the di-amino acid derivatives is that the hydrogen bond

acceptor of the former constitutes an amide carbonyl moiety and not an ester carbonyl group.

This was the background behind this research project, and several points attracted my

attention. First of all, the fact whether a hydrogen bond between the amide NH and the methyl

ester in the case of Fc(Cp(CO)-Phe-OMe)2 really is present [230]. To me this conformation

appeared somewhat doubtful in view of the relatively poor hydrogen bond acceptor properties

of ester carbonyl oxygen atoms. The plans to investigate this involve the resynthesis of the

above mentioned phenylalanine compound and the synthesis of a chiral diamide analogue that

Ferrocene and cobaltocenium conjugates of amino acids and dipeptides, a hydrogen bonding study

129

lacks ester groups. The infrared and 1H NMR spectroscopic behaviour of these disubstituted

derivatives is compared mutually and with that of the corresponding monosubstituted

derivatives, which are unable to form intramolecular hydrogen bonds.

At the time this project started, the dipeptide derivatives Fc(Cp-C(O)-Ala-Pro-OR)2

were known [236, 239], but these contain only one amide NH group per peptide strand. My

interest concerned replacing the proline residues by other amino acids, such as phenylalanine,

and subsequent investigation whether this newly introduced NH amide moiety would be

involved in any hydrogen bond interactions in solution and in the solid state. This year, the

answer was already provided in the literature, before the project had come to an end [237,

240].

Another interest was the synthesis of the analogous iso-electronic cobaltocenium complexes,

and subsequent investigation of the influence of the positive charge of these compounds in

relation to the ferrocene derivatives. Thus far, cobaltocenium complexes bearing amino acid

or dipeptide substituents have not been reported.

8.2 Synthesis

The route of synthesis and constitution of the selected ferrocene derivatives for this hydrogen

bonding study are depicted in Scheme 8.2. The ferrocene compounds were synthesised

according to a general method, which consists of reacting the commercially available

ferrocene carboxylic acid (mono or 1,1'-di) with stoichiometric amounts of the amine, amino

acid or dipeptide in DMF in the presence of NEt3 and the coupling reagent TBTU. After

stirring at room temperature for 30 minutes, the mixture was evaporated to dryness in vacuo.

The residue was suspended in CH2Cl2, subsequently filtered to remove solid material and

finally subjected to an extractive work-up. This work-up is comprised of three consecutive

washing-steps, first with aqueous 2 M NaHCO3, followed by 1 M HCl and finally water.

After drying of the organic phase over MgSO4 and removal of the solvent under reduced

pressure, the compounds were obtained as orange solids. Only in the case of the S-1-

phenylethylamine (PEA) derivatives 37 and 38, an extra purification step by column

chromatography over silica was necessary because small amounts of unreacted amine were

found to be present in the isolated solids. The yields varied from 52 to 87 %. It has to be noted

that compounds 35 and 36 have been reported before, but these were not fully characterised

[230].

Chapter 8

130

Scheme 8.2 Synthesis and constitution of the mono and disubstituted ferrocene derivatives

Fe OH

O

Fe AA-R

O

Fe OH

O

OH

O

FeAA-R

O

AA-R

O

35 Phe-OMe

37 PEA

39 Ala-Phe-OMe

AA-R

36 Phe-OMe

38 PEA

40 Ala-Phe-OMe

AA-R

+ AA-R

+ 2 AA-R

a

a

PEA = S-1-phenylethylamine

Reagents and conditions: a) TBTU in DMF / NEt3; stirring at RT for 30 mins

The synthesis and constitution of the cobaltocenium compounds for this hydrogen bonding

project are depicted in Scheme 8.3. The cobaltocenium acids (mono or 1,1'-di) are, in contrast

to their ferrocene analogues, not commercially available and were synthesised according to a

published procedure [241].

Scheme 8.3 Synthesis and constitution of the cobaltocenium derivatives

Co OH

O

Co AA-R

O

Co OH

O

OH

O

CoAA-R

O

AA-R

O

41 Phe-OMe

43 Ala-Phe-OMe

AA-R

42 Phe-OMe

44 Ala-Phe-OMe

AA-R

+ AA-R

+ 2 AA-R

a

bPF6

BPh4PF6

PF6

Reagents and conditions: a) TBTU in DMF / NEt3; stirring at RT for 30 mins; addition of NaBPh4 in

MeOH after the work-up (see text); b) TBTU in DMF / NEt3; stirring at RT for 30 mins


131

The first part of the synthesis of the mono cobaltocenium derivatives 41 and 43 was

performed analogously to that of the ferrocene derivatives, but oily residues were obtained

after the extractive work-up and removal of the solvent. After dissolution of these in MeOH

and addition of a NaBPh4 solution in methanol, the compounds precipitated in pure form as

their tetraphenylborate salts. The yield of the cobaltocenium compounds 41 and 43 was lower

than of their ferrocene analogues. This may be attributed to the overall positive charge of the

cobaltocenium complexes, which makes them more water-soluble, resulting in significant

losses during the extractive work-up.

The di-substituted cobaltocenium derivatives 42 and 44 were obtained by reacting

cobaltocenium di-carboxylic acid with either H-Phe-OMe or H-Ala-Phe-OMe in DMF by

employing TBTU as the coupling reagent. After the extractive workup, the oily residue was

dried in vacuo for a few hours, yielding the compounds as hygroscopic powders, with purity

around 90-95 % as concluded from their 1H NMR spectra. Unfortunately the complexes could

not be purified by crystallisation and also did not precipitate from MeOH upon addition of

other anions that are unable to act as hydrogen bond acceptors, such as BF4- and

tetraphenylborate. Instead, analytically pure samples were obtained by preparative HPLC

purification.

My interest also concerned the S-1-phenylethylamine cobaltocenium derivatives. However,

the work-up of the reaction of [Co(Cp-COOH)2]PF6 with two equivalents of S-1-

phenylethylamine (PEA) did not proceed smoothly, because the organic layer was invariably

colourless after the washing steps. Also the mono S-1-phenylethylamine derivative was

difficult to obtain. By using an identical procedure as for the compounds 41 and 43, a small

amount of X-ray quality crystals of [Co(Cp)(Cp-C(O)-PEA)]BPh4 (45) was obtained the first

time the synthesis was performed. However, when tried again, the precipitates were invariably

not pure, because also S-1-phenylethylammonium salts separate from the mixture. Since the

corresponding cobaltocenium di-S-phenylethylamine derivative could not be obtained at all,

the purification as well as the total characterisation of 45 was not further pursued.

S-1-phenylethylamine appears to be much less reactive towards coupling reactions with the

organometallic acids than amino acids or dipeptides. The disubstituted PEA cobaltocenium

derivative could not be obtained at all, and the yield is significantly lower for the other PEA

compounds in comparison to the dipeptide or amino acid derivatives.

Chapter 8

132

The 1H and 13C NMR spectra of the compounds 35-44 are consistent with the proposed

constitution and the resonances were assigned unambiguously. Furthermore, the mass spectra

(EI for 35-40 and ESI-positive for 41-44) show fragments with the appropriate mass.

Satisfying elemental analyses were obtained for all these compounds, except for 44. Because

the absolute yield of complex 44 was only 30 mg, a high-resolution mass spectrum was

recorded for this compound instead of an elemental analysis. Only approximately 2 mg of

substance were required for the determination of the exact mass of the [44-PF6]+ peak in the

ESI-positive mass spectrum, on account of the high intensity of this fragment.

The identity of the compounds 36, 39, 40, 41 and 45 was also confirmed by X-ray

crystallography, which is the subject of the next section.


In this section, the X-ray crystal structures of 36, 39⋅0.25CH2Cl2, 40⋅0.5CHCl3, 41 and

[Co(Cp)(Cp-C(O)-PEA)]BPh4 (45) are presented. The details concerning how the crystals

were obtained can be found in the experimental section. The crystal structures will be

compared mutually, and with related compounds in the literature. No metal-C(Cp) bond

distances will be presented because these show only statistical variations. Several more

interesting structural parameters, such as the tilt and other specific angles will be presented

later in this section.

An ORTEP projection for the cationic part of [Co(Cp)(Cp-C(O)-PEA)]BPh4 (45) is shown in

Figure 8.1. The cation does not display any unusual structural features, and the unit cell

consists of an isolated complex cation and a tetraphenylborate anion. Hydrogen bonds are not

present in the solid state, being consistent with the observation of the amide νNH stretch

vibration at 3384 cm-1 in the infrared spectrum of 45 as a KBr pellet.


133

Figure 8.1 ORTEP plot for the cationic part of [Co(Cp)(Cp-C(O)-PEA)]BPh4 (45)

The unit cell of 41 was found to consist of two crystallographically independent complex

cations and two tetraphenylborate anions. ORTEP diagrams of both cations are depicted in

Figure 8.2. In cation A, depicted on the left of Figure 8.2, the phenyl ring of the phenylalanine

substituent is directed towards the unsubstituted Cp-ring and the methyl ester is pointing

downwards. In cation B on the other hand, the substituents have more or less exchanged

positions relative to cation A, i.e. the phenyl ring is located away from the substituted Cp-ring

and the methyl ester is directed towards the unsubstituted Cp-ring. No hydrogen bonds are

present in the solid state, which is consistent with the observation of the amide νNH stretch

vibration at 3404 cm-1 in the infrared spectrum of 41 as a KBr pellet.

Figure 8.2 ORTEP projection for the independent cations A (left) and B (right) of 41

Chapter 8

134

An ORTEP plot for the asymmetric unit in 36 is shown in Figure 8.3. The compound

crystallises in the hexagonal space group P65, which implies the presence of a left-handed six-

fold screw axis in the unit-cell. The complexes are arranged in a helical fashion around this

six-fold screw axis, linked together via intermolecular hydrogen bond interactions between

N(27) and O(6) of another molecule (N⋅⋅⋅O contact = 2.935 Å). Also an intramolecular

hydrogen bond is present between N(7) and O(26) (N⋅⋅⋅O contact = 2.832 Å). The helix has a

pitch of 10.3690(7) Å, which is equivalent to the length of the crystallographic c-axis, because

the six-fold screw axis is per definition aligned parallel to this axis in the hexagonal space

group P65. Another parameter of any helix is n, which is defined as the number of residues

after which the original orientation of the helix is again obtained or in other words: the

number of units per helical turn. In this case n is equal to six, because a six-fold screw axis is

present.

Figure 8.3 ORTEP representation for the asymmetric unit in the structure of 36

Helices constitute an important class of secondary structural elements observed in proteins [7,

8]. The most common helix is the α helix, which has a pitch of 5.4 Å and 3.6 peptide units per

helical turn. Two other types of helices that occur less frequently in proteins are the 310 helix,

which has a pitch of 6.0 Å and n = 3, and the π helix, which has a pitch of 5.2 Å and a helical

rise of 4.4 units [7, 8]. The helical arrangement of compound 36 in the unit cell differs

considerably from the helices that occur in proteins, which may be attributed to the bulkiness

of the ferrocene residue.


135

The X-ray crystal structure of a different di-amino acid ferrocene derivative, namely Fe(Cp-

C(O)-Val-OMe)2, has been reported [231]. This valine derivative has the ordered

conformation depicted in Scheme 8.1 in the solid state and shows two very weak symmetry-

related hydrogen bond interactions between the amide NH group and the methyl ester

carbonyl oxygen atom with N⋅⋅⋅O contacts of 3.247 Å. In contrast to Fe(Cp-C(O)-Val-OMe)2,

36 does not have intramolecular hydrogen bonds between the ester carbonyl oxygen atom and

the amide NH moiety. Apparently, the presence of bulky phenyl rings in 36 compared to the

smaller iso-propyl side chains in Fe(Cp-C(O)-Val-OMe)2 results in a dramatic packing

difference in the solid state.

An ORTEP representation for the asymmetric unit in 39⋅0.25CH2Cl2 is shown in Figure 8.4.

The compound crystallises in the tetragonal space group P43212, which implies the presence

of a left-handed four-fold screw axis in the unit cell. The molecules are arranged in a helical

fashion around this screw axis, as shown in Figure 8.5. Intermolecular hydrogen bond

interactions are present between N(7) and O(6) of a neighbouring molecule (N⋅⋅⋅O contact =

2.936 Å) and between N(10) and O(9) of the same neighbouring molecule (N⋅⋅⋅O contact =

2.862 Å).

Figure 8.4 ORTEP plot for the ferrocene residue in 39⋅0.25CH2Cl2.

The helix has a pitch of 16.9834(12) Å, which is also in this case equal to the length of the

crystallographic c-axis, and a helical rise of 4.0 units. This type of helix is very interesting,

because two hydrogen bond donors and two acceptors exist per molecule. Three other

ferrocenyl dipeptides have been structurally characterised by X-ray crystallography, which are

Fe(Cp)(Cp-C(O)-Pro-Pro-OBzl) [235], Fe(Cp)(Cp-C(O)-Gly-Pro-OEt) [237] and Fe(Cp)(Cp-

C(O)-Ala-Pro-OEt) [236]. These complexes are significantly different in comparison to 39

Chapter 8

136

because the first one contains no amide NH hydrogen atoms, whereas the second and third

each contain only one amide NH moiety. Consequently, the packing of these complexes is

significantly different compared to 39⋅0.25CH2Cl2, with the compounds Fe(Cp)(Cp-C(O)-

Ala-Pro-OEt) and Fe(Cp)(Cp-C(O)-Gly-Pro-OEt) showing a zigzag hydrogen bonding

arrangement, whereas no hydrogen bonds are present in the case of Fc(Cp)(Cp-C(O)-Pro-Pro-

OBzl due to the absence of amide NH groups.

Figure 8.5 Packing diagram for 39⋅0.25CH2Cl2, showing the helical arrangement of the

ferrocene conjugates and the intermolecular hydrogen bond interactions

X-ray crystallography revealed the presence of two crystallographically independent

molecules in the unit cell of 40⋅0.5CHCl3, which display only slightly different bond lengths

and angles. An ORTEP diagram for one of the crystallographically independent molecules is

depicted in Figure 8.6. There are two intramolecular hydrogen bonds between the alanine NH

group and the alanine-C=O moiety of another strand. In molecule A (depicted in Figure 8.5),

the N···O contacts are 2.921 Å and 2.924 Å (N(1)···O(6) and N(3)···O(2), respectively),

whereas these N···O contacts are slightly shorter in the other independent molecule (molecule

B, not depicted; N···O = 2.901 Å and 2.893 Å).


137

Figure 8.6 ORTEP plot for one of the independent molecules in 40⋅0.5CHCl3, showing

the intramolecular hydrogen bonds. Only specific atoms are labelled for clarity

In addition to the intramolecular hydrogen bonds, there are also intermolecular hydrogen bond

interactions in the lattice, between the phenylalanine NH groups and the ferrocenyl C=O

moieties of two neighbouring molecules that belong to the same crystallographic type. Each

dipeptide strand forms intermolecular hydrogen bonds to a different molecule of the same

type, resulting in infinite chains of molecules A and B throughout the lattice. Between

molecules A, the N···O contacts are 2.834 Å and 2.844 Å (N(4)···O(1) and N(2)···O(5),

respectively), whereas these are 2.830 Å and 2.836 Å for the intermolecular hydrogen bonds

between molecules B. The hydrogen bonds are arranged in such way that a 14-membered ring

forms, as visualised in Figure 8.7. This arrangement is similar to what has been reported very

recently for the packing of Fe(Cp-C(O)-Gly-Phe-OMe)2 [240]. A major difference between

this X-ray crystal structure and that of 40⋅0.5CHCl3 is that the unit cell of 40⋅0.5CHCl3

contains two crystallographically independent molecules.

Figure 8.7 Visualisation of the 14-membered hydrogen bond ring present throughout the

lattice (in this case between molecules A)

Chapter 8

138

Several structural parameters for the five X-ray crystal structures are summarised in Table

8.1, with the corresponding angles being visualised in Scheme 8.4. As can be observed the Cp

rings in those structures are almost coplanar, with the largest deviation observed for 36. The

tilt of 4.7 ° in 36 is likely owing to the rigid hydrogen bonded helical arrangement of the

complex molecules. This deviation is slightly larger than what is usually observed for

disubstituted ferrocenes bearing carboxyl substituents [242, 243], but it is not exceptionally

large because a value for θ of 16.4° was reported for a strained macrocyclic ferrocenyl

derivative [244].

Table 8.1 Summary of specific angles for the X-ray crystal structures[a]

Angle[a] 36 39⋅0.25CH2Cl2 40⋅0.5CHCl3[b] 41[b] 45

θ[c] 4.7 1.1 1.3 1.5 3.3 1.9 0.7

β[d] 31.4, 15.1[e] 6.1 4.1, 5.8[e] 3.7, 10.3[e] 10.5 9.9 5.5

ω[e] 68.9 70.3 68.9

[a] Angles are visualised in Scheme 8.4; [b] Two independent molecules in the unit cell; first value(s)

correspond(s) to molecule A; [c] The dihedral angle between the two Cp rings; [d] The dihedral angle

between the plane of the Cp ring and the C(ipso)-C=O bond; [e] Disubstituted derivatives; first value

for β corresponds to the C=O moiety with the lower atomic label; [f] The torsion angle defined as

C(ipso)-Cp(centroid)-Cp(centroid)-C(ipso)

Scheme 8.4 Schematic representation of the structural angles listed in Table 8.1

Fe θ Fe

O

O

ωFe

Oβ


139

The amide moieties in all structures, except in 36, are almost planar with the Cp-rings, in this

way allowing a maximum π-overlap between the cyclopentadienyl and amide π systems. The

large β angle for 36 of 31.4° is due to the intramolecular hydrogen bond interaction, which

forces one of the amide moieties to rotate significantly out of the Cp plane. However, an angle

of 31.4° still accounts for about 86% of the maximum overlap between the amide π orbitals

and cyclopentadienyl π system, because the overlap of π-orbitals shows a cosinus dependence

on this angle [245]. This deviation from planarity of the amide functionality in 36 is not

exceptional because twists of 45.7° and 50.2° have been observed previously for constrained

disubstituted ferrocene compounds [244, 246].

The observed ω angles for 36 and both independent molecules of 40⋅0.5CHCl3 indicate a 1,2'-

conformation for these structures with the values being close to the idealised value for this

conformation (360°/5 = 72°). The values for ω are in the range reported for Fe(Cp-C(O)-Ala-

Pro-OR)2 (R = Me, Et or Bzl) and for several disubstituted ferrocenes that are part of a

macrocycle [239, 242]. The Cp rings of these derivatives are forced into this 1,2'-

conformation, either imposed by intramolecular hydrogen bonds or by steric constraints of the

rigid macrocycle. In the case of difunctionalised ferrocenes having groups that are unable to

form any hydrogen bonds, the substituents will be arranged in such way that the steric

interaction between these is minimised. Consequently, ω angles much larger than 80° are

observed for ferrocene derivatives of that type.

8.4 Investigation of the conformation in solution

Relevant infrared and 1H NMR spectroscopic data for 35-44 are summarised in Table 8.2.

Amide NH hydrogen atoms that are not hydrogen bonded display νNH stretch vibrations in the

infrared spectrum at wavenumbers higher than 3400 cm-1 and show resonances in between 5.5

and 7.0 ppm in the 1H NMR spectrum, whereas NH stretching vibrations below 3400 cm-1 in

the infrared spectrum are diagnostic of hydrogen bonded amide NH hydrogen atoms. In

addition, amide NH hydrogen atoms involved in hydrogen bond interactions have a chemical

shift higher than 7.5 ppm. From Table 8.2, it is observed that the mono-substituted ferrocenes

and cobaltocenium complexes display νNH stretching vibrations at wavenumbers higher than

3400 cm-1 and the chemical shift of the NH hydrogen atoms is between 5.8 ppm and 6.5 ppm.

This indicates that the NH hydrogen atoms of these compounds do not form hydrogen bonds

in solution, as expected.

Chapter 8

140

Table 8.2 Summary of relevant spectroscopic data for the NH hydrogen atoms in 35-44

Compound νNH[a] δ NH[b] 3JHH

[c] ∆δ[d]

35 3436[e] 6.00[e] n. r.[h, i] -1.4

36 3380[f] 7.75[f] 8.4 -5.2

37 3440[e] 5.82[e] 7.6 -2.2

38 3435[f] 6.93[f] 7.9 -7.2

39 3424[e] 6.53[e]

6.18

7.8

7.5

-5.4

-3.6

40 3322[f]

3414

8.33[f]

6.31

7.1

7.7

-4.5

-2.4

41 3404[e] 6.11[e] 7.9 -0.7

42 3365[f]

3404[k]

7.79[m] 8.3 -8.6

43 3404[g] n. d. [n] n. d. [n] n. d.[n]

44 3296[f]

3405

8.82

6.58

6.8

n. r.[h, i]

-4.7

n. r.[h, o]

[a] in cm-1; in CH2Cl2; [b] in dry CDCl3; T = 293 K; in ppm; [c] in Hz; [d] temperature range: 223-323

K; in ppb K-1; [e] 2 × 10-2 M; [f] 1 × 10-2 M; [g] 5 × 10-3 M; [h] n. r. = not reliable; [i] broad NH resonance

observed; [k] second νNH vibration observed. See text for explanation; [m] value not reliable because

the complex is very hygroscopic;[n] insoluble in CDCl3; [o] resonance becomes very broad at T<293 K.

The ferrocene di-phenylalanine methyl ester derivative 36 shows a νNH stretch vibration at

3380 cm-1 and a chemical shift for the NH hydrogen atoms of 7.75 ppm, similar to what has

been reported previously in the literature for this compound [230]. In the case of the di-S-1-

phenylethylamine derivative 38, the νNH stretch vibration is located at 3435 cm-1 and the NH

hydrogen atom resonates at 6.93 ppm. The differences between 36 and 38 clearly indicate the

presence of two intramolecular hydrogen bonds between the amide NH and the methyl ester

carbonyl moiety in 36. The small differences between the mono PEA derivative 37 and its

disubstituted analogue 38 indicate that strong hydrogen bonds are not present in the solution

conformation of 38. However, if the amide function would be twisted to a small extent out of


141

the Cp-plane, a very weak hydrogen bond interaction is expected to be possible between the

amide NH and the amide C=O of the other Cp-ring. The resonance energy between the planar

Cp-ring and the planar amide is expected to decrease in this case but the weak hydrogen bond

interaction would compensate for that to a certain extent. These rationalisations might provide

an explanation for the observed small infrared and 1H NMR spectroscopic differences

between 37 and 38. For the solid state structure of 36 it has been observed

crystallographically that the amide C=O moiety can rotate out of the Cp-plane.

In the case of the di-substituted Ala-Phe-OMe ferrocene 40 the hydrogen bond

observed in the solid state between the Ala-NH group and the Ala-C=O moiety of another

strand is also present in solution. This can be derived from the position of the νNH stretch

vibration at 3322 cm-1 and the chemical shift of this NH hydrogen atom, which is located at

8.33 ppm. These hydrogen bonds are stronger in 40 than in the di-Phe-OMe derivative 36,

which is consistent with the differences between the hydrogen bond acceptor strengths of

ester and amide carbonyl moieties. The νNH stretch vibration at 3322 cm-1 is in the range

reported for several other disubstituted ferrocene derivatives bearing dipeptide substituents

[236, 237, 239, 240].

The phenylalanine NH hydrogen atom is not involved in any hydrogen bond interactions, as

concluded from the observation of the νNH stretch vibration at 3414 cm-1 and the chemical

shift at 6.31 ppm. It should be noted that the assignment of these alanine and phenylalanine

amide NH hydrogen atoms was made unambiguously via 2D NMR techniques. Thus, it can

be concluded that only the first NH hydrogen atom can form a hydrogen bond in solution,

whereas the second cannot because of spatial reasons, similar to what is observed for the X-

ray crystal structure 40⋅0.5CHCl3. These findings are identical to what has been reported for

similar compounds this year by Hirao and coworkers [237, 240].

The disubstituted cobaltocenium Ala-Phe-OMe derivative 44 has a conformation in

solution identical to that of the ferrocene analogue 40, with the alanine NH moieties of this

compound being involved in two intramolecular hydrogen bonds. The νNH stretch vibration of

the alanine NH group observed for 44 is situated about 25 cm-1 lower in energy compared to

40. Whether this is caused by a stronger hydrogen bond is not clear, because the mono-

substituted derivatives 41 and 43 display νNH stretching vibrations that are located about 20-

30 cm-1 lower than those for the corresponding ferrocene derivatives 36 and 39. The chemical

shift of the alanine NH hydrogen atom in 44 is significantly higher compared to that of the

ferrocene analogue 40 (8.82 ppm in 44 vs. 8.33 ppm in 40), which seems to indicate that the

Chapter 8

142

intramolecular hydrogen bonds in 44 are stronger than those in 40. Because the chemical shift

of the NH hydrogen atoms of 35 and 41 differs significantly less (0.11 ppm), it can be

concluded that the large difference between the resonances of the alanine NH hydrogen atoms

in 44 and 40 is not owing to varying properties of ferrocenyl amides with respect to

cobaltocenium amides. However, the results should be taken with the appropriate care,

because compound 44 is hygroscopic and the presence of small amounts of water would

already influence the measurements to a large extent.

The disubstituted Phe-OMe cobaltocenium derivative 42 is much more hygroscopic than 44,

and in this case water molecules can disrupt the weak hydrogen bond between the amide NH

and the methyl ester C=O moiety. This is derived from the infrared spectrum by the

observation of two NH stretching vibrations. The vibration at 3365 cm-1 is assigned as the NH

vibration owing to the hydrogen bonded NH moiety and the one at 3404 cm-1 is assigned as

NH stretching vibration originating from the non-hydrogen bonded NH. In fact, the value of

the latter is identical to that for the mono phenylalanine derivative 41, which cannot be

involved in any hydrogen bond interactions.

Coupling constants between the CαH hydrogen atom and the neighbouring amide NH

hydrogen atom are important for determination of the conformation of a peptide in solution

[181]. The values of the coupling constants observed for the compounds 35-44 between 6.8

and 8.4 Hz are not very informative, because these indicate a dihedral angle ϕ of around 90°

[179], which is a standard value for a trans-configured amide.

Variable temperature measurements over a range of about 100 K were made to investigate

variation of the chemical shift of the amide NH hydrogen atoms (see Table 8.2). It is known

that amide NH hydrogen atoms that are either locked in a very strong hydrogen bond or those

that are not hydrogen bonded at all exhibit a small temperature dependence of the chemical

shift of about –2 to –4 ppb K-1 [247]. It has to be noted that the concentration should be less

than the concentration at which self-association occurs. When values other than those

mentioned above are observed, nothing can be derived because the values for ∆δ are not

linearly correlated to the strength of the hydrogen bond interaction. From the values for ∆δ of

the NH hydrogen atoms of the compounds summarised in Table 8.2 no trends can be derived.

The presence and the strength of hydrogen bonds can be much better determined from the

position of the νNH stretch vibration in the infrared spectrum and the chemical shift of the


143

amide NH hydrogen atoms. It has to be noted that the self-association concentrations were

not investigated and it might therefore be that a 10 mM solution is already too concentrated.

Furthermore, it must be noted that the presence of small amounts of water is expected to

influence the values for ∆δ to a large extent.

8.5 Mössbauer spectroscopy and electrochemistry

Before the Mössbauer spectroscopic data and the electrochemical properties of the

compounds are presented, it is appropriate to treat the MO diagram of ferrocene and the

cobaltocenium cation, its iso-electronic analogue. This will facilitate the discussion of the

electrochemical and Mössbauer spectroscopic data. The MO diagram for ferrocene in D5d

symmetry has become a textbook example, and is shown in Scheme 8.5 [245].

Scheme 8.5 Construction of the molecular orbitals for ferrocene (adapted from ref [245]).

a1g

a2u

e1u

e1g

___

_

____

__

__

e1ge2ga1g

________

__e1g

a1g

e2g

____

e1u

a1g

a2u

Fe 2+Fe2 _ z

x y

Chapter 8

144

At low energy, the levels a1g and a2u are situated, which are comprised mainly of Cp22- orbitals

but these have mixed to a small extent with metal s and pz orbitals, respectively. The Cp22-

orbital levels e1u and e1g interact with the metal px, py set and dxz, dyz orbitals, respectively.

These six molecular orbitals account for the bonding between the Fe atom and the Cp rings.

At moderate energy, the molecular orbitals labelled a1g and e2u are situated, which are the non-

bonding d(z2) orbital and d(x2-y2), dxy set, respectively. From photo-electron spectroscopic

measurements, it has been derived that the e2g set is slightly higher in energy than the a1g set

[248] and, thus, the HOMO of ferrocene is e2g. The LUMO of ferrocene, labelled e1g, is the

antibonding orbital from the interaction of the Cp22- e1g set with the metal dxz and dyz orbitals.

The results from 57Fe Mössbauer spectroscopic investigations on the ferrocene derivatives are

collected in Table 8.3.

Table 8.3 57Fe Mössbauer spectroscopic data and oxidation potentials for 35-40

Electrochemistry[a] 57Fe Mössbauer spectroscopy[b]

Compound E1/2[c] δ[d] ∆EQ

[d] Γ[d]

Ferrocene 0.00[e] 0.53[f] 2.42[f]

35 +0.19 0.53 2.37 0.34

36 +0.40 0.52 2.27 0.27

37 +0.17 0.53 2.32 0.30

38 +0.35 0.51 2.27 0.30

39 +0.19 0.52 2.32 0.28

40 +0.34 0.52 2.27 0.32

[a] in CH2Cl2; NBu4PF6 as supporting electrolyte; [b] at 80 K; [c] in V; vs Fc / Fc+;[d] in mm s-1; [e] by definition; [f] taken from ref. [249]

The isomer shifts for 35-40 are virtually identical to those for ferrocene [249], whereas these

compounds display lower quadrupole splitting values in comparison to ferrocene. The values

for the quadrupole splitting of the disubstituted derivatives 36, 38 and 40 are significantly

lower than those of the monosubstituted compounds 35, 37 and 41. It has been noted before


145

that the quadrupole splitting of ferrocene decreases when a Cp ring is substituted with an

electron-withdrawing substituent and even more when such a substituent is introduced on

each Cp ring [250]. This effect has been explained by the withdrawal of electron density by

the substituent(s) from the orbital e1g, the highest occupied orbital with Cp character, which

leads to a different field gradient along the molecular axis and, thus, to a different value for

the quadrupole splitting [250].

The redox potentials for the ferrocene derivatives are shown in Table 8.3. The substituted

derivatives display slightly higher oxidation potentials than ferrocene (by definition 0.00 V).

It appears that the substituent effects are additive, because the redox-potential difference

between ferrocene and the mono-substituted derivatives is approximately equal to the

difference between the disubstituted derivatives and the monosubstituted compounds. The

increase of the oxidation potential upon substitution of the Cp ring by amide groups can be

explained by the withdrawal of electron density from the ferrocene moiety by these

substituents, resulting in a derivative that is more difficult to oxidise. A molecular orbital

treatment of the interaction between the orbitals of ferrocene with those of several electron-

withdrawing substituents recently appeared in the literature [243]. The authors of that

publication stated that one of the components of the e2g set (the HOMO) has a higher

interaction with the orbitals from the substituent. This results in stabilisation of one of the

components of the e2g set, with the other component remaining about identical in energy and,

thus, this component becomes the new HOMO of the compound. Although the authors could

correlate trends in bond-lengths observed for various substituted ferrocene derivatives this

way, this MO treatment does not provide an explanation for the shift of the oxidation

potential, because the new HOMO of the system is at nearly the same energy as the previous

e2g set. More satisfying on the other hand are experimental results obtained from photo-

electron spectroscopic investigations, because these clearly demonstrate that the HOMO shifts

to lower energy upon introduction of electron withdrawing substituents on the Cp-ring [248].

From the data summarised in Table 8.3, it is observed that the size of the substituent does not

influence the oxidation potential to a large extent. The monosubstituted derivatives show

about the same redox potential, and a similar behaviour is observed for the disubstituted

derivatives. The redox potentials for the amino acid and dipeptide derivatives are in the range

reported for related compounds in the literature [232, 236, 239].

Chapter 8

146

The main difference between the electrochemical behaviour of ferrocene derivatives and their

cobaltocenium analogues is that the former class can be oxidised, whereas the latter can only

undergo a one-electron reduction. Upon reduction of the cobaltocenium cation, which is iso-

electronic with ferrocene, the LUMO e1g becomes half-filled. The reduction potentials for the

cobaltocenium derivatives 41-44 are shown in Table 8.4. From this table, it is observed that

the reduction shifts to less negative potentials upon substitution of a Cp-ring with an amide

moiety and when both Cp-rings are substituted, the shift of the reduction potential is

significantly larger. This shows that the e1g level is lowered in energy, or in other words,

becomes more accessible when electron-withdrawing substituents are present. The direction

of the observed shift upon substitution of a Cp-ring is the same as observed for the ferrocene

derivatives.

Table 8.4 Reduction potentials for the cobaltocenium derivatives 41-44[a]

Compound E1/2[b] Compound E1/2

[b]

41 -1.13 43 -1.13

42 -0.93 44 -1.00

[Co(Cp)2]PF6 -1.36[c] [Co(Cp)2]PF6 -1.34[d]

[a] in CH2Cl2; NBu4PF6 as supporting electrolyte; [b] in V; vs Fc / Fc+; [c] measured value;[d] from ref. [251]


The goals mentioned in the beginning of this chapter have been established by the

experiments and spectroscopic investigations described in the previous sections. It has been

shown that Fe(Cp-C(O)-Phe-OMe)2 (36) indeed has an ordered conformation in solution, with

the amide NH moieties being hydrogen bonded to the methyl ester carbonyl oxygen atoms.

These hydrogen bonds are significantly weaker than those observed for the dipeptide

derivative Fe(Cp-C(O)-Ala-Phe-OMe)2 (40), which is due to the difference between the

hydrogen bond acceptor strengths of ester and amide carbonyl oxygen atoms. Analogous


147

cobaltocenium derivatives could also be prepared, but these were found to be much more

hygroscopic than their ferrocene analogues. The hydrogen bonds between the amide moieties

and the ester carbonyl oxygen atoms in the case of [Co(Cp-C(O)-Phe-OMe)2]PF6 (42) were

also found to be much weaker than those observed for the dipeptide derivative [Co(Cp-C(O)-

Ala-Phe-OMe)2]PF6 (44). In fact, small amounts of water could easily disrupt the

intramolecular hydrogen bonds of 42. It appears that the intramolecular hydrogen bonds

between the Ala-NH moieties and the Ala-C=O groups in 44 are stronger than those observed

for the ferrocene derivative 40, but the results should be taken with the appropriate care in

view of the hygroscopic nature of the cobaltocenium compounds.

148

9Summary

The title of this Thesis already points out the two important objectives of this graduate work.

The first goal is the synthesis and development of spectroscopic transition metal markers that

exhibit good stability in aerobic aqueous media and the second objective is the subsequent

introduction of these markers onto oligopeptides by solid phase peptide synthesis methods.

Jaouen and co-workers have shown in the last decade that labelling of hormones, drugs and

proteins in solution with organometallic complexes containing carbonyl ligands allows the

conjugates to be detected in picomolar quantities by infrared spectroscopy. However, many

complexes are quite reactive towards dioxygen and none of these markers has been shown to

be compatible with solid phase peptide synthesis methods.

The first class of compounds that are investigated for use as a marker are molybdenum

complexes containing the (η-allyl)(CO)2 moiety. In Chapter 3, the complex Mo(η-Cp)(η-

allyl)(CO)2 is functionalised with a carboxylic acid group by lithiation with n-butyl lithium,

followed by reaction with solid CO2. The resulting complex Mo(η-Cp-COOH)(η-allyl)(CO)2

is coupled to various amino acids and a dipeptide. The phenylalanine methyl ester derivative

Mo(η-Cp-CO-Phe-OMe)(η-allyl)(CO)2 (3a) was structurally characterised and shows an

endo-orientation of the allyl ligand. Thus far, all solid state structures of Mo(η-Cp)(η-

allyl)(CO)2 and its derivatives had an exo-orientation of the allyl ligand. In solution, the

bioconjugates exist as an equilibrium mixture of the endo and exo conformers, as

demonstrated by variable temperature 1H NMR spectroscopic investigations. The presence of

a chiral centre on the amino acid results in inequivalence of the allyl-hydrogen atoms, leading

to observation of ten allyl-hydrogen atom resonances in the low temperature limit. The results

described in Chapter 3 have been published already [252].

Summary

149

The conjugates of Mo(η-Cp-COOH)(η-allyl)(CO)2 do not display reversible electrochemical

transitions and are quite reactive towards dioxygen. Instead of attaching this marker to larger

peptides via solid phase peptide synthesis, the properties of a different class of markers were

investigated. The Cp-ring was replaced by its iso-electronic analogue L-histidinate, which

resulted in complexes with an enhanced stability in aerobic atmospheres. In Chapter 4, the

properties of the complexes Mo(L-His-NεR)(η-allyl)(CO)2 (with R = H or C2H4C(O)OCH3)

are investigated in relation to analogous complexes with the 2-Me-allyl ligand. The idea

behind this was to investigate which of the two types of complexes would be more suitable

for the labelling of peptides. During the investigations, it was found that these complexes are

fluxional in solution, both in their neutral and oxidised form. In the neutral form, the

compounds exist as a mixture of isomers a and b in solution (see Scheme 9.1), as concluded

from the results of NMR spectroscopy and DFT-calculations. Interestingly, the allyl

derivatives crystallised in the a conformer, whereas the crystal structures of the 2-Me-allyl

derivatives revealed the conformation of isomer b in the solid state. In solution, the isomers a

and b are in equilibrium and the netto process for interconversion constitutes a restricted

trigonal twist. The DFT-calculations, however, indicated the interconversion to be a more

complex process, involving some allyl-endo isomers as intermediates.

Scheme 9.1 Conformation of the observed isomers a and b in solution for 5-8. (R = H or

CH3; R' = H or C2H4C(O)OMe). The terminal C-atoms of the allyl or 2-Me-

allyl ligand are directed towards the carbonyl ligands.

N

N

H2N

O O

MoOC

OC

R

R'

R N

N

H2N

O O

MoOC

R'

CO

a b

Chapter 9

150

The complexes display reversible one-electron oxidations, and upon oxidation the compounds

strongly prefer the conformation in which their allyl or 2-Me-allyl ligand is in a trans-position

relative to the carboxylate oxygen atom, as derived from low temperature electrochemical

investigations. The square wave voltammograms of Mo(L-His-Nε-C2H4C(O)OCH3)(η-

allyl)(CO)2 (7) at four different temperatures are shown in Figure 9.1. At low temperatures

two species are discernible, whereas at higher temperatures a rare case of coalescence in the

square wave voltammograms is observed. By a complete analysis of the electrochemical data,

thermodynamic parameters such as ∆G0 and equilibrium constants were determined. In the

oxidised form, the conformation with the allyl or 2-Me-allyl ligand in a trans-position relative

to the carboxylate oxygen atom is more stable than the conformation with the allyl or 2-Me-

allyl ligand trans to the Nδ by 7.5 kJ mol-1 for the allyl complex 7 and by 8.3 kJ mol-1 for the

2-Me-allyl complex 8. These results obtained from low temperature electrochemical

investigations have already been published in part [253].


EtCN; 0.1 M NBu4PF6)

By combining the results from EPR spectroscopic investigations and DFT-calculations on the

oxidised complexes, the orientation of the allyl and 2-Me-allyl ligand was elucidated (See

Scheme 9.2). The oxidised 2-Me-allyl complexes have an endo-orientation of this ligand,

whereas the oxidised allyl complexes exist as a mixture of exo and endo isomers, with the

exo-orientation being the major conformer. The fact that the endo-isomers of the allyl and 2-

Summary

151

Me-allyl complex display identical EPR spectra can be explained by the results from the

DFT-calculations, which showed that the SOMOs of these species are very similar. The exo

isomer of the allyl ligand complex has a different SOMO, which is consistent with the

observed differing EPR spectrum. The allyl complex has been selected for labelling purposes,

because the 1H NMR spectra of the allyl complexes display two distinct sets of resonances for

the isomers a and b at room temperature.

Scheme 9.2 Favoured conformations of the oxidised allyl and 2-Me-allyl complexes

N

N

H2N

O O

MoOC

CO

R

N

N

H2N

O O

MoOC

CO

R +

N

N

H2N

O OMo

+

OCCO

R +

minor major b e e

Complexes with the allyl ligand Complexes with the 2-Me-allyl ligand

In Chapter 6, the electrochemical and spectroscopic properties of a different class of markers

containing the Mo(bpa)(CO)3 unit are investigated (bpa = di(2-picolyl)amine) in relation to

analogous complexes with the ligands 1,4,7-triazacyclononane and hydrido-tris-

pyrazolylborate. Especially the benzyl derivative Mo(benzyl-bpa)(CO)3 displayed reversible

electrochemical transitions, as concluded from spectro-electrochemical investigations, and

showed good air-stability.

In Chapters 5 and 7, markers based on the above mentioned complexes Mo(His)(η-

allyl)(CO)2 and Mo(benzyl-bpa)(CO)3 are developed. The markers Mo(His-

NεC2H4COOH)(η-allyl)(CO)2 (with L-histidine and D-histidine) are first coupled to amino

acids and dipeptides via reactions in solution. Although the resulting bioconjugates with the

L-histidinate and D-histidinate complexes are diastereomers, they can only be distinguished

on the basis of their circular dichroism spectroscopic data. The chirality of the histidinate

ligand is clearly reflected in the CD-spectra, because the L-histidinate complexes show

negative Cotton effects at around 320 and 406 nm, whereas the D-histidinate complexes show

Chapter 9

152

positive Cotton effects. The identity of one of the conjugates, namely the compound Mo(L-

His-Nε-C2H4CO-Phe-OMe)(η-allyl)(CO)2, was confirmed by X-ray crystallography (see

Figure 9.2).

Figure 9.2 ORTEP plot for Mo(L-His-Nε-C2H4CO-Phe-OMe)(η-allyl)(CO)2 (14)

The marker Mo(benzyl-bpa)(CO)3 is introduced into a different way onto amino acids and

dipeptides (Chapter 7). First the ligand N-benzylic acid-N,N-di(2-picolyl)amine is coupled to

the biomolecule and in a next step, the Mo(CO)3 moiety is introduced under mild conditions

by reaction of the bpa-conjugate with Mo(CO)3(NCEt)3 for 10 minutes at room temperature.

In Chapters 5 and 7, [Leu]-enkephalin derivatives with the above mentioned markers have

been synthesised via solid phase peptide synthesis methods. [Leu]-enkephalin is a neuro

peptide with an action similar to morphine, having primary structure H-Tyr-Gly-Gly-Phe-

Leu-OH. The complexes Mo(His-Nε-C2H4COOH)(η-allyl)(CO)2 (with L-histidinate and D-

histidinate) are coupled to the N-terminal tyrosine on the solid support and the resulting

peptide-conjugates withstand even cleavage with concentrated NH3 in MeOH.

The resulting molecular structures are shown in Scheme 9.3 and these compounds constitute

the first carbonyl complexes that are introduced onto a peptide via reactions on a solid

support. These compounds were characterised by 1H and 13C NMR spectroscopy, infrared

spectroscopy and mass spectrometry.

Summary

153

Scheme 9.3 Molecular structures of the [Leu]-enkephalin molybdenum complex conjugates

obtained via solid phase peptide synthesis

N

N

H2N

OO

MoCO

CO

HN

NH

O

O

HN

O

NH

O

N

O

O

R

NH2

O

18 L 19 D 20 L21 D

His R

2-ClTrt2-ClTrt H H

NN

NMoOC CO

CO

O

NH

HN

ONH

OHN

ON

O

OH

34

NH2

O

The presence of the Mo(His)(allyl)(CO)2 moiety on the 2-ClTrt deprotected [Leu]-enkephalin

influences the properties of the pentapeptide to some extent. The phenol OH group of 20 and

21 was found to be involved in a hydrogen bonding interaction with the histidinate

carboxylate moiety in DMSO. From 1H NMR spectroscopic investigations, this hydrogen

bond was found to be stronger in the case of the D-histidinate containing conjugate 21 with

respect to the L-histidinate conjugate 20. In contrast to the diastereomeric amino acid and

dipeptide derivatives, the diastereomeric [Leu]-enkephalin derivatives 20 and 21 can be

distinguished by 1H NMR spectroscopy.

The Mo(bpa)(CO)3 [Leu]-enkephalin derivative 34 was found to be much more hydrophobic

than the derivatives 20 and 21, and is only soluble in DMSO and DMF. It was also found to

be slightly less stable in air than the [Leu]-enkephalin derivatives with the Mo(His)(η-

allyl)(CO)2 markers.

Chapter 9

154

In Chapter 8, the first examples of amino acids and dipeptides attached to the cobaltocenium

moiety are presented. Both mono and 1,1’-disubstituted derivatives are prepared and the

properties of these were investigated in relation to their neutral ferrocene analogues. An X-ray

crystal structure of the bis-dipeptide derivative Fe(Cp-CO-Ala-Phe-OMe)2 (40) was obtained

(see Figure 9.3). This compound was found to have intramolecular hydrogen bonds between

the Ala-NH groups and the Ala-C=O of another peptide strand.

Figure 9.3 ORTEP plot for one of the independent molecules in 40⋅0.5CHCl3, showing

the intramolecular hydrogen bonds. Only specific atoms are labelled for clarity

Unsurprisingly, it was found that the cationic cobaltocenium derivatives are much more

hygroscopic than their neutral ferrocene analogous. In the case of the cobaltocenium

derivative [Co(Cp-CO-Ala-Phe-OMe)2]PF6 (see Scheme 9.4), the intramolecular hydrogen

bonds between the Ala-NH and the Ala-C=O moieties of the other strand appear to be

stronger than those for the neutral ferrocene derivative 40.

Scheme 9.4 Constitution of the cobaltocenium bis-Ala-Phe-OMe derivative 44

CoAla-Phe-OMe

O

Ala-Phe-OMe

O

PF6

155

Experimental section

Physical measurements

Elemental analysis

Elemental analyses were performed by H. Kolbe, Microanalytisches Laboratorium, Mülheim

an der Ruhr, Germany.

Infrared spectra

Infrared spectra were recorded on a Perkin-Elmer FT-IR spectrophotometer 2000 as KBr

disks and/or in solution, with a spectral resolution of either 2.0 cm-1 or 4.0 cm-1. Spectra in

CH2Cl2 and THF were measured between NaCl windows (either a 0.5 mm or 1.0 mm spacer),

whereas spectra in MeOH were acquired by using an OTTLE-cell with CaF2 windows, having

a path-length of 0.17 mm. The concentration was in between 10-2 M and 10-3 M, unless

otherwise noted. For infrared spectro-electrochemical measurements, the above mentioned

OTTLE-cell was used with either CH2Cl2, THF or CH3CN as the solvent and NBu4PF6 as

supporting electrolyte. In two cases, described in Chapter 6, NBu4Br was used as supporting

electrolyte.

NMR spectra1H and 13C NMR spectra were recorded on Bruker ARX 250 (1H at 250.13 MHz), Bruker

DRX (1H at 400.13 MHz) and Bruker DRX 500 (1H at 500.13 MHz; during the experimental

work later adjusted to 500.35 MHz). 1H and 13C NMR spectra were referenced to TMS, using

the 13C or residual protio signals of the deuterated solvents as internal standards. Variable

temperature 1H NMR spectra were recorded on the Bruker DRX 400 NMR spectrometer.95Mo NMR spectra were recorded on the Bruker DRX 500 spectrometer (32.6 MHz for 95Mo),

using a 2M Na2MoO4 solution at an apparent pH 11 as an external standard. For the VT NMR

experiments of the compounds in Chapter 8, CDCl3 was dried over activated molecular sieves

(3 Å) and the NMR tubes were thoroughly dried before use. Resonances of the same

hydrogen or carbon atom of the major and minor isomer are denoted ma / mi.


156

Mass spectrometry

Mass spectra in the Electron Impact mode (EI; 70 eV) were recorded on a Finnigan MAT

8200 mass spectrometer. Only characteristic fragments are given with intensities in

parentheses and possible composition in brackets. The spectra were normalised against the

most intense peak, which therefore has intensity 100. Electron Spray Interface (ESI) mass

spectra were recorded on a either Finnigan Mat 95 instrument or a Hewlett-Packard HP 5989

mass spectrometer. The mode (positive or negative) and the used solvent are given in

parentheses and the possible composition of the peaks is between brackets. Fast atom

bombardment (FAB) mass spectra were recorded on a Finnigan Mat 95 mass spectrometer.

The mode (positive or negative) is given parentheses together with the used matrix. High

Resolution Mass Spectra were recorded in the ESI-positive mode on a Finnigan Mat 95 mass

spectrometer.

UV-Visible spectroscopy

UV-Visible spectra were recorded on either a Perkin-Elmer Lambda 19 spectrophotometer or

on a Hewlett-Packard HP 8452A diode array spectrophotometer in various solvents. For UV-

Vis spectro-electrochemical investigations, the HP 8452A diode array spectrophotometer was

used, by employing a coulometry cuvet, either NBu4PF6 or NBu4Br as supporting electrolyte

and either CH3CN or CH2Cl2 as the solvent, as indicated in the text.

Cicular dichroism spectroscopy

Circular dichroism spectra were recorded on a Jasco J-715 spectropolarimeter, equipped with

a Hamamatsu R376 photomultiplier (185 nm-850 nm) using a band-with of 1 nm.

EPR spectrosocopy

Electron Paramagnetic Resonance spectra were recorded on a Bruker ESP 300E

Spectrometer, equipped with an ER 041 XK-D microwave brigde (X-band EPR spectra) or an

ER 051 QG micowave bridge (Q-band EPR spectra). All EPR spectra were recorded on

frozen solutions generated by controlled potential coulometry.


157

Mössbauer spectroscopy57Fe Mössbauer spectra were recorded on an Oxford Instruments Mössbauer spectrometer in

the constant acceleration mode, by using 57Co/Rh as the radiation source. The measurements

were performed on solid samples at 80 K containing a natural abundance of 57Fe. The

mininum experimental line-width was 0.24 mm s-1. Isomer shifts were determined relative to

α-Iron at 300 K.

Electrochemistry

Cyclic voltammograms and square wave voltammograms were recorded by using an EG&G

Potentiostat / Galvanostat 273A. A three electrode cell was employed with a glassy-carbon

working electrode, a platinum-wire auxiliary electrode and a Ag/AgNO3 reference electrode

(0.01 M AgNO3 in MeCN). For the low temperature measurements described in Chapter 4, a

Ag/AgNO3 reference electrode (0.01 M, AgNO3 in EtCN) was used. Controlled potential

coloumetry measurements (normal as well as for UV-Vis spectro-electrochemical studies)

were performed by employing the same potentiostat, but by making use of Pt-grid as working

electrode, a Pt-brush separated from the working electrode compartment by a vycor frit and an

Ag/AgNO3 (0.01 M AgNO3 in MeCN) working electrode. Infrared spectro-electrochemical

measurements were performed by using an OTTLE-cell consisting of a Pt-grid working

electrode, a glass-carbon counter electrode and a Ag-wire as “quasi-reference electrode”.

X-ray crystallography

X-ray data were collected on either a Siemens SMART CCD diffractometer or an Enraf-

Nonius Kappa CCD diffractometer (with or without rotating anode), by using Mo-Kα

radiation (λ = 0.71073 Å). The data were corrected for absorbtion and Lorentz polarisation

effects. The structures were solved by direct methods and refined by full-matrix least-squares

on F2 by using the ShelXTL software package.

Computational details

Geometries have been fully optimized without symmetry constraints at the BP86/SDD level,

i.e. employing the exchange and correlation functionals of Becke [254] and Perdew [255,

256], respectively, together with a fine integration grid (75 radial shells with 302 angular

points per shell), and the Stuttgart-Dresden relativistic effective core potentials with the

corresponding valence basis sets for Mo (6s5p3d) [257], C, N, and O (valence double-zeta)


158

[258] and Dunning’s double-zeta basis for H [259]. The nature of the stationary points has

been verified by computations of the harmonic frequencies at that level. The nature of

transition states has been verified by the occurrence of a single imaginary frequency and by

visual inspection of the corresponding vibrational modes, in order to ensure that the desired

minima are connected. Subsequently, single-point energy calculations for these BP86/SDD

geometries have been performed at the BPW91/II" level, i.e. using the functionals of Becke

[254], Perdew and Wang [260, 261], a (16s10p9d) all-electron basis for Mo, contracted from

the well-tempered 22s14p12d set of Huzinaga and Klobukowski [262] and augmented with

two d-shells of the well-tempered series, and standard 6-31G* basis set [263, 264] for all

other elements. This level is denoted BPW91/II"//BP86/SDD. Unless otherwise noted,

energies are reported at this level, including the BP86/SDD zero-point corrections. The same

methods, together with more extended basis sets on the ligands, have proven to be well-suited

for a description of the reactivity and 95Mo chemical shifts of a number of inorganic and

organometallic molybdenum compounds [172] and are among those DFT-based methods

which are now routinely used to study many aspects of transition-metal chemistry [265].

Open-shell systems were treated with the unrestricted Kohn-Sham formalism, and no

significant spin contamination was found. All computations employed the Gaussian 98

program package [266].

Synthesis

Synthesis of literature compounds

The complexes Mo(η-allyl)(Br)(CO)2(NCMe)2 and Mo(η-2-Me-allyl)(Cl)(CO)2(NCMe)2

were prepared according to the method by Dieck and Friedel [139]. The complex Mo(η-Cp)(

η-allyl)(CO)2 was prepared via the method by Hayter [138]. The complexes Mo(tacn)(CO)3,

[Mo(tacn)(CO)3Br]Br3 and Mo(Me-tacn)(CO)3 were synthesized as reported by Wieghardt

and coworkers [202, 203]. The ligand N-benzyl-N,N-di-(2-picolyl)amine was prepared from

2-picolylchloride hydrochloride and benzylamine by the method of Dick and Weiss [224].

The compounds [Co(Cp)(Cp-COOH)]PF6 and [Co(Cp-COOH)2]PF6 were prepared via the

method reported by Rausch and Sheats [241].


159

General

For the synthesis of 2, THF was dried over Na / benzophenone and the glassware necessary

for that synthesis was thoroughly dried. MeOH was dried over CaH2 for the synthesis of

compound 34. The synthesis of the Mo-complexes and their bioconjugates were performed

under an atmosphere of argon and normal undried deoxygenated solvents were used, unless

otherwise noted. For coupling reactions with HBTU or TBTU, peptide-grade or HPLC-grade

solvents were used. All chemicals as well as all other solvents were obtained from

commercial sources.

Solid phase peptide synthesis

All sovents were of HPLC or peptide synthesis grade, whereas all chemicals were obtained

from NovaBiochem and used as received. The synthesis were performed batchwise on an

Eppendorf-Biotronik Ecosyn P. peptide synthesiser using highly crosslinked polystyrene

beads functionalised with a 4-hydroxymethylbezoic acid linker (NovaSyn TGA resin) as the

solid support (loading 0.2-0.3 mmol g-1). The first amino acid, L-leucine, was coupled to the

solid support via a literature procedure [114], employing MSNT and 1-methyl imidazole as

coupling reagents, yielding a resin with a loading of 0.04-0.06 mmol g-1. Assembly of the

peptide part was performed by reaction of the resin-bound amino acid (L-leucine for the first

coupling step) or resin bound peptide with a three-fold excess of the Fmoc protected amino

acids, coupling reagents (TBTU and HOBT) in 4 mL DMF / dipea (7/1 v/v) for 35 mins. The

resin was washed thoroughly with DMF after each coupling step. Removal of the Fmoc group

was performed by reaction of the resin-bound amino acid (L-leucine for the first deprotection

step) or resin-bound peptide with 15 mL of a DMF / piperidine solution (3/ 1 v/v) for 20 mins.

After the Fmoc-deprotected [Leu]-enkephalin was obtained, the next steps were performed

manually. First the 2-ClTrt group was removed (except for the synthesis of 18 and 19, in

which cases this step was omitted) by reaction of the resin-bound Fmoc-2-ClTrt-[Leu]-

enkephalin with 5% triisopropyl silane and 5% CF3COOH in CH2Cl2 (3 × 10 mL, 2 minutes

reaction time). Subsequently the Fmoc group was removed by reacting the Fmoc-[Leu]-

enkephalin with 15 mL of a DMF / piperidine solution (3/1 v/v) for 30 minutes. Coupling

reactions of the metal complexes or the functionalised di(2-picolyl)amine ligand were


160

performed with a five-fold excess of the acid, six equivalents of TBTU in 15 mL of a DMF /

dipea solution (3/1 v/v) for 16 hours. Cleavage of the resin was performed by reaction of the

resin-bound [Leu]-enkephalin conjugate with a saturated NH3 solution in MeOH for 48 hours

at room temperature. The [Leu]-enkephalin conjugates were purified by preparative HPLC.

HPLC purification

High Performance Liquid Chromatography (HPLC) purifications were performed by using a

Macherey & Nagel Nucleosil 7-C-18 Column (250 mm × 21 mm). As eluent for the

purification of 42 and 44 a mixture of MeCN and H2O, the latter containing 0.05 mol L-1

NaPF6 (2/1 v/v), was used, whereas for the purification of all other compounds MeOH/H2O

mixtures with the ratio varying between 1/1 and 3/1 were used as eluent. All purifications

were performed without an applied gradient.

Synthesis of Boc-Phe-Leu-OMe (1a)

To a stirred solution of Boc-phenylalanine (1.33 g; 5.0 mmol) in THF (50 mL) was added N-

methylmorpholine (0.56 mL; 0.51 g; 5.0 mmol) and isobutyl chloroformiate (0.66 mL; 0.68 g;

5.0 mmol), resulting in formation of a white precipitate. In another flask, leucine methyl ester

hydrochloride (0.91 g; 5.0 mmol) was suspended in THF (50 mL), followed by addition of

NEt3 (0.69 mL; 0.51 g; 5.0 mmol). The suspensions were mixed, and the resulting suspension

stirred for one hour at ambient temperature, followed by filtration to remove a white

precipitate. The solvent was removed under reduced pressure and the residue dissolved in

CH2Cl2 (75 mL) and the CH2Cl2 solution was washed with H2O (50 mL) and the phases were

separated. The aqueous phase was extracted with CH2Cl2 (3 × 50 mL), and the combined

CH2Cl2 solutions were dried over MgSO4. Evaporation of the solvent under reduced pressure

yielded Boc-Phe-Leu-OMe as a white solid. Yield: 1.7 g (87%).

C21H32N2O5 = 392.5 g mol-1.

IR (KBr): ν = 3292 (w, br) νNH, 1750 (s) νC=O, ester, 1683 (s), 1648 (s) νC=O, amide + Boc. MS (EI):

m/z = 392 (3) [M]+, 164 (47), 120 (75), 57 (100). 1H NMR (500.35 MHz; DMSO-d6): δ = 8.21

(1H, d, 3JHH = 7.7 Hz, NH), 7.26 (4H, m, HAr), 7.17 (1H, m, HAr), 6.84 (1H, d, 3JHH = 8.7 Hz,

NH), 4.34 (1H, m, Cα-PheH), 4.21 (1H, m, Cα-LeuH), 3.61 (3H, s, OCH3), 2.95 (1H, m, Cβ-


161

PheH), 2.74 (1H, m, Cβ-PheH), 1.67 (1H, m, Cγ-LeuH), 1.58 (1H, m, Cβ-LeuH2), 1.50 (1H, m, Cβ-

LeuH2), 1.29 (9H, s, Boc-CH3), 0.89 (3H, d, 3JHH = 6.5 Hz, Leu-CH3), 0.84 (3H, d, 3JHH = 6.5

Hz, Leu-CH3). 13C NMR (125.8 MHz; DMSO-d6): δ = 172.8 (C=Oester), 171.8 (C=OPhe),

155.2 (C=OBoc), 138.0 (CAr), 129.2 (CAr), 127.9 (CAr), 126.1 (CAr), 77.9 (Cq, Boc), 55.4, 51.7,

50.1 (Cα-Leu, Cα-Phe, OCH3), Cβ-Leu is obscured by the solvent signal, 37.2 (Cβ-Phe), 28.0 (Boc-

CH3), 24.0 (Cγ-Leu), 22.8, 21.2 (Both Leu-CH3).

Synthesis of Boc-Ala-Phe-OMe (1b)

This compounds was prepared via a similar procedure as described for 1a, by using Boc-Ala-

OH (0.95; 5.0 mmol) and phenylalanine methyl ester hydrochloride (1.08 g; 5.0 mmol).

Yield: 1.5 g (86%).

C18H26N2O5 = 350.4 g mol-1.

IR (KBr): ν = 3330 (m) νNH, 1744 (s) νC=O, ester, 1698 (m), 1665 (s), 1655 (s) νC=O, amide + Boc.

MS (EI): m/z = 350 (2) [M]+, 162 (65), 88 (58), 57 (85), 44 (100) 1H NMR (500.35 MHz;

DMSO-d6): δ = 8.09 (d, 3JHH = 7.6 Hz, NH), 7.26 (2H, m, HAr), 7.20 (3H, m, HAr), 6.86 (1H,

d, 3JHH = 7.4 Hz, NH), 4.49 (1H, m, Cα-PheH), 3.98 (1H, m, Cα-AlaH), 3.57 (3H, s, OCH3), 3.01

(1H, m, Cβ-PheH), 2.95 (1H, m, Cβ-PheH), 1.36 (9H, s, Boc-CH3), 1.12 (3H, d, 3JHH = 6.9 Hz).13C NMR (DMSO-d6; 125.8 MHz): δ = 173.3 (C=Oester), 172.3 (C=OAla), 155.3 (C=OBoc),

137.5 (CAr), 129.5 (CAr), 128.6 (CAr), 127.0 (CAr), 78.5 (Cq, Boc), 53.8, 52.2, 50.0 (Cα-Ala, Cα-

Phe, OCH3), 37.2 (Cβ-Phe), 28.6 (Boc-CH3), 18.6 (Ala-CH3).

Synthesis of Mo(η-Cp-COOH)(η-allyl)(CO)2 (2)

To a solution of Mo(η-Cp)(η-allyl)(CO)2 (1.02 g, 3.95 mmol) in THF (40 mL) was added

3.70 mL of 1.6 M n-BuLi in hexane (5.92 mmol) at –78 ºC. After 45 min of stirring about 7 g

of solid CO2 was added and the reaction mixture was stirred for another 10 min at –78 ºC.

After allowing the mixture to warm up to room temperature, about 40 mL of water was added,

followed by reduction of the volume to about 25 mL in vacuo. The pH was adjusted to 1 with

6 M HCl, resulting in precipitation of a yellow solid. The precipitate was collected by

filtration, washed with water (10 mL) and diethyl ether (5mL) and dried in vacuo. Yield: 1.06

g (89 %).


162

C11O4H10Mo = 302.1 g mol-1.

Anal. Calcd.: C 43.73, H 3.34. Found: C 43.62, H 3.27. IR (KBr): ν = 1930 (vs), 1869 (vs),

νCO, 1681 (s), νCOO; (CH2Cl2): ν = 1957 (vs), 1875 (vs), νCO, 1732 (m), 1713 (s), 1683 (m)

νCOO. MS (EI): m/z = 304 (59) [M]+, 276 (41) [M-CO]+, 248 (97) [M-2CO]+. 1H NMR

(400.13 MHz, Acetone-d6): (all signals were broadened and signals from the minor isomer

were difficult to detect with certainty), δ = 5.87 (2H, s, Cp-H), 5.54 (2H, pseudo-t, Cp-H),

4.16 (1H, br, Hc-allyl), 2.85 (2H, br, Hs/a-allyl), 0.97 (2H, Hs/a-allyl). 13C NMR (62.9 MHz,

Acetone-d6): δ = 236.8 (CO), 166.1 (C=O), 95.1, 94.6 (CCp), 71.2, 42.4 (Callyl).

General synthesis of Mo(η-Cp-CO-AA-R)(η3-allyl)(CO)2

(AA-R= L-Phe-OMe (3a), L-Leu-NH2 (3b) and R = Gly-OMe (3c))

Compound 2 (300 mg, 0.99 mmol) and 0.99 mmol of the amino acid hydrochloride (L-

phenylalanine methyl ester hydrochloride (214 mg) for 3a, L-leucinamide hydrochloride (165

mg) for 3b or glycine methyl ester hydrochloride (125 mg) for 3c were dissolved in DMF (3

mL). Triethylamine (1.5 mL) and HBTU (377 mg, 0.99 mmol) were added and the reaction

mixture was stirred for 45 min. at room temperature. Upon addition of 40 mL of an aqueous

0.5 M NaHCO3 solution, a yellow precipitate formed, which was collected by filtration,

washed with water (10 mL) and dried in vacuo. Yield: 420 mg (91%) for 3a, 310 mg (75 %)

for 3b and 247 mg (67%) for 3c. The obtained precipitates were of sufficient purity for

characterisation, but analytically pure samples as well as single crystals of 3a suitable for X-

ray analysis were obtained by slow pentane diffusion in a THF solution at 4 °C.

3a: C21NO5H21Mo = 463.3 g mol-1.

Anal. Calcd.: C 54.44, H 4.57, N 3.02. Found: C 54.31, H 4.63, N 2.98. IR (KBr): ν = 3307

(m, br) νNH, 1966 (vs), 1945 (vs), 1867 (vs), νCO, 1742 (s), νC=0-ester, 1623 (s), νC=0-amide;

(CH2Cl2): ν = 3429 (m), νNH, 1950 (vs), 1869 (vs) νCO, 1742 (s) νC=0-ester, 1664 (s) νC=0-

amide. MS (EI): m/z = 465 (23) [M]+, 437 (3) [M-CO]+, 409 (4) [M-2CO]+, 367 (83), 254

(100). 1H NMR (400.13 MHz, Acetone-d6): δ = 7.50 (1H, br, NH), 7.30 (4H, m, Ar-H), 7.22

(1H, m, Ar-H), 5.86 (1H, s, Cp-H), 5.84 (1H, s, Cp-H), 5.53 (2H, pseudo-t, Cp-H), 4.76 (1H,


163

br, CαH), 3.87 / 3.63 (1H, m, Hc-allyl), 3.68 (3H, s, OCH3), 3.19 (1H, m, CβH), 2.97 (1H, m,

CβH), 2.79 / 2.80 and 2.59 / 2.80 (2H, overlapping m, Hs/a-allyl), 0.86 / 1.66 and 0.86 / 1.5

(2H, overlapping m, Hs/a-allyl). 13C NMR (125.8 MHz, Acetone-d6): δ = 237.4, 237.3 (CO),

172.6 (C=O-ester), 163.6 (Cp-C=O), 138.3 (CAr, q), 129.9, 129.3, 127.5 (CAr), 104.7 (CCp, q),

94.1, 91.3, 91.2 (CCp), 72.4 (Callyl), 54.7 (Cα), 52.4 (CH3), 43.0, 42.9 (Callyl), 37.9 (Cβ).

3b: C17O4N2H22Mo = 414.3 g mol-1.

Anal. Calcd.: C 49.28, H 5.35. N 6.76; Found: C 49.35, H 5.42, N 6.68. IR (KBr): ν = 1947

(vs), 1869 (vs) νCO, 1684 (s), 1668 s), 1631 (s) νC=O; (CH2Cl2): 3515 (w), ν(NH2), 3402 (w)

νNH, 1953 (vs), 1870 (vs) νCO, 1698 (s), 1675 (vs) νC=O. MS (EI): m/z = 416 (16) [M]+, 360

(21) [M-2CO]+, 316 (100). 1H NMR (400.13 MHz, Acetone-d6): δ = 7.43 (1H, br, NH), 7.04

(1H, br, O=CNH2), 6.47 (1H, br, O=CNH2), 5.98 (1H, pseudo-t, Cp-H), 5.94 (1H, br, Cp-H),

5.52 (2H, m, Cp-H), 4.56 (1H, m, CαH), 4.1 / 3.68 (1H, m, Hc-allyl), 2.88 (2H, Hs/a-allyll),

1.75 (1H, m, CγH), 1.61 (2H, m, CβH2), 0.92 (8H, d, 3JHH = 6.5 Hz, both CH3 and 2 Hs/a-allyl).13C NMR (100.6 MHz, Acetone-d6): CO not observed, δ = 175.2 (C=Oester), 163.8 (Cp-C=O),

104.4 (CCp, q) 94.2, 93.6, 92.2, 91.9 (CCp), 72.1 (Callyl), 52.1 (Cα), 42.8, 42.7 (Callyl), 41.9 (Cβ),

25.4 (Cγ), 23.6 (CH3), 21.8 (CH3).

3c: C14O5NH15Mo = 373.2 g mol-1.

Anal. Calcd.: C 45.06, H 4.05, N 3.75; Found: C 45.20, H 3.98, N 3.72. IR (KBr): ν = 3279

(m) νNH, 1946 (vs), 1866 (vs) νCO, 1753 (s) νC=O-ester, 1629 (s) νC=O-amide; (CH2Cl2): ν =

3452 (w) νNH, 1952 (vs), 1869 (vs) νCO, 1748 (s) νC=O-ester, 1667 (s) νC=O-amide. MS (EI):

m/z = 375 (43) [M]+, 347 (8) [M – CO]+, 319 (65) [M – 2CO]+, 277 (100). 1H NMR (400.13

MHz, Acetone-d6) 7.70 (1H, br, NH), 5.89 (2H, pseudo-t, Cp-H), 5.58 (2H, s, Cp-H), 4.04 /

3.61 (1H, m, Hc-allyl), 3.99 (2H, d, 3JHH = 5.3 Hz, CH2), 3.68 (3H, s, OCH3), 2.91 (2H, m,

Hs/a-allyl), 0.95 / 1.77 (2H, m, Hs/a-allyl). 13C NMR (100.6 MHz, Acetone-d6): δ = 237.1

(CO), 170.8 (C=Oester), 164.1 (Cp-C=O), 104.7 (CCp, q), 94.0, 91.8 (CCp), 72.5 (Callyl), 52.1

(OCH3), 43.0 (Callyl), 41.8 (CH2).


164

Synthesis of Mo(η-Cp-CO-Phe-Leu-OMe)(η-allyl)(CO)2 (4)

Boc-Phe-Leu-OMe (1a) (190 mg, 0.5 mmol) was dissolved in a mixture of CH2Cl2 (3 mL)

and CF3COOH (3 mL) and stirred for 1 hr at room temperature. After removal of the solvent,

diethyl ether (10 mL) was added and the mixture was taken to dryness again. This step was

performed three times in total in order to remove any traces of trifluoro acetic acid. A solution

of 2 (151 mg, 0.5 mmol) in DMF (2 mL) and NEt3 (1 mL) was added, followed by addition of

HBTU (190 mg, 0.5 mmol). After stirring at room temperature for 45 min., 30 mL of an

aqueous 0.5 M NaHCO3 solution was added, resulting in formation of a yellow precipitate.

The solid was collected by filtration, washed with water (10 mL) and dried in vacuo. Yield:

180 mg (62%).

C27O6N2H32Mo = 576.51 g mol-1.

Anal. Calcd.: C 56.25, H 5.59, N 4.86; Found: C 54.56, H 6.03, N 4.49. IR (KBr): δ = 3296

(m) νNH, 1951 (vs), 1868 (vs) νCO, 1746 (s), νC=O-ester, 1655 (s), 1629 (s), νC=O-amide;

(CH2Cl2): 3420 (m) νNH, 1953 (vs), 1870 (vs) νCO, 1743 (s) νC=O-ester, 1712 (s), 1682 (s),

1660 (s) νC=O-amide. MS (EI): m/z = 578 (330 [M]+, 550 (2) [M-CO]+, 522 (2) [M-2CO]+,

480 (100) [M - 2CO - allyl + H]+. 1H NMR (500.35 MHz, Acetone-d6): δ = 7.61 (1H, br, NH),

7.46 (1H, br, NH), 7.28 (4H, m, Ar-H), 7.20 (1H, t, 3JHH = 7.1 Hz, Ar-H), 5.84 (2H, br, Cp-

H), 5.27 (2H, m, Cp-H), 4.79 (1H, br, Cα-PheH), 4.52 (1H, br, Cα-LeuH), 3.78 / 3.47 (1H, m, Hc-

allyl), 3.66 (3H, s, OCH3), 3.16 (1H, m, Cβ-PheH), 2.93 (1H, m, Cβ-PheH), 2.76 / 2.72 (1H, m,

Hs/a-allyl), 2.48 / 2.80 (1H, m, Hs/a-allyl), 1.71 (1H, br, Cγ,-LeuH), 1.60 (2H, t, 3JHH = 6.3 Hz,

Cβ-LeuH2) 0.90 (6H, overlapping d, both CH3), 0.89 / 1.35 (1H, d, J = 11 Hz, Hs/a-allyl), 0.82 /

1.17 (1H, d, J = 10.5 Hz, Hs/a-allyl). 13C NMR (100.6 MHz, Acetone-d6): δ = 237.4, 237.3

(CO), 173.5, 173.0 (C=O), 163.8 (Cp-C=O), 138.7 (CAr, q), 130.1, 129.2, 127.3 (CAr), 105.2

(CCp, q), 95.0, 93.6, 91.0, 90.4 (CCp), 72.7 (Callyl), 55.5 (Cα-Phe), 52.3 (OCH3), 51.6 (Cα-Leu),

43.1, 42.9 (Callyl), 41.3 (Cβ-Leu), 38.4 (Cβ-Phe), 25.4 (Cγ-Leu), 23.2, 21.9 (both Leu-CH3).

Synthesis of Mo(L-His)(η-allyl)(CO)2 (5)

To a solution of Mo(η-allyl)(Br)(CO)2(MeCN)2 (3.95 g, 11.0 mmol) in MeOH (35 mL) was

added a solution of L-histidine (1.73 g, 11.0 mmol) and KOH (0.62 g, 11.0 mmol) in H2O (30


165

mL). After 45 mins of stirring, the mixture was concentrated in vacuo to about 20 mL. The

yellow precipitate was isolated by filtration, washed with H2O (15 mL) and diethyl ether (15

mL) and dried in vacuo. Yield: 3.70 g (97 %). X-ray quality crystals of 5⋅MeOH were

obtained by evaporation of a H2O/MeOH (1/3 v/v) solution under a stream of argon.

C11H13MoN3O4 = 347.2 g mol-1.

IR (KBr): ν = 3332 (m) νNH, 1932 (vs), 1835 (s), 1805 (s) νCO, 1615 (m) νC=O; (MeOH): ν =

1939, 1847 νCO; MS (ESI-pos., MeOH): m/z = 350 [M+H]+, 372 [M+Na]+, 721 [2M+Na]+. 1H

NMR (400.13 MHz, DMSO-d6): δ = 8.56 / 7.94 (1H, s, CεH), 7.03 / 6.84 (1H, s, CδH), 4.71 /

3.93 (1H, br, NH2), 4.05 / 3.90 (1H, br, NH2), 3.50 (1H, m, CαH), 2.85 / 2.76 (2H, m, CβH2),

3.51 / 3.51 (1H, m, Hc-allyl), 3.28 / 3.17 (1H, m, Ha/s-allyl), 3.11 / 2.78 (1H, m, Ha/s-allyl),

1.22 / 0.80 (1H, m, Ha/s-allyl), 0.99 / 0.80 (1H, m, Ha/s-allyl). 13C NMR (100.6 MHz, DMSO-

d6): δ = 230.0 / 228.8, 228.2 / 227.9 (CO), 178.2 / 179.1 (C=O-His), 138.7 / 140.5 (Cε), 133.5

/ 135.0 (Cγ), 114.7 / 114.6 (Cδ), 74.2 / 68.1 (Cc-allyl), 60.2 / 57.8 (C-allyl), 51.7 / 52.2 (C-

allyl), 52.2 / 52.0 (Cα), 28.0 / 28.0 (Cβ).

Synthesis of Mo(L-His)(η-2-Me-allyl)(CO)2 (6)

A suspension of Mo(η-2-Me-allyl)(Cl)(CO)2(MeCN)2 (1.64 g, 5.1 mmol) in EtOH (20 mL)

was heated on a water bath (50° C) for 5-10 mins, during which time a clear orange-reddish

solution formed. A solution of L-histidine (0.79 g, 5.1 mmol) and KOH (0.29 g, 5.1 mmol) in

H2O (5 mL) was added, resulting in formation of a small amount of precipitate, which

dissolved again upon stirring. After the mixture was stirred at room temperature for 30 min.,

the solvent was removed in vacuo. The sticky residue was re-dissolved in EtOH (15 mL) and

stirred for 15 mins, resulting in formation of a yellow precipitate. The solid was isolated by

filtration and washed with H2O (5 mL) and diethyl ether (15 mL) and dried in vacuo. Yield:

1.50 g (82 %). Crystals suitable for X-ray analysis were obtained by evaporation of a

H2O/MeOH (1/3 v/v) solution under a stream of argon.

C12O4MoN3H15 = 361.2 g mol-1.

IR (KBr): ν = 3323 (m), 3248 (m) νNH, 1928 (vs), 1842 (s), 1805 (vs) νCO, 1645 (s), 1616 (m)

νC=O; (MeOH): ν = 1939, 1848 νCO. MS (FAB-neg.; dimethoxybenzylalcohol): m/z = 362 (M


166

– H). 1H NMR (400.13 MHz, DMSO-d6): δ = 12.62 (1H, br, NH), 8.41 / 7.99 (1H, br, CεH),

6.91 (1H, br, CδH), 5.04 / 4.19 (1H, br, NH2), 3.32 (1H, br, NH2), 3.28 (2H, br, Cα + HAllyl),

2.91 (2H, m, CβH2), 2.66 (1H, br, HAllyl), 2.66 (1H, br, HAllyl), 1.75 (3H, br, CH3), 0.80 (2H,

br, HAllyl). 13C NMR (62.9 MHz, DMSO-d6): δ = 228.2 (br, CO), 178.7 (C=O), 138.9 (br, Cε),

134.2 (Cγ), 114.7 (Cδ), 79.9 (br, Callyl), 55.9 (br, Callyl), 52.2 (Cα), 28.0 (Cβ), 19.3 (CH3).

Synthesis of Mo(L-His-Nε-C2H4CO2Me)(η-allyl)(CO)2 (7)

To a solution of 5 (5.90 g; 17.0 mmol) in DMF (35 mL) was added 3-bromo-propionic acid

methyl ester (1.90 mL; 2.91 g; 17.4 mmol) and Cs2CO3 (5.55 g; 17.0 mmol). After heating the

mixture at 80° C for 2 hrs, it was filtered, followed by evaporation to dryness. The residue

was re-suspended in THF (70 mL), subsequently filtered to remove the Cs salts, followed by

removal of THF in vacuo. Purification by preparative HPLC yielded 4.81 g (65 %) of 7. X-ray

quality crystals of 7⋅2MeOH were obtained by slow diffusion of a pentane solution into a

solution of 7 in MeOH at +4° C.

C15H19MoN3O6 = 433.3 g mol-1.

Anal. Calcd.: C 41.37, H 4.37, N 9.66; Found C 41.51, H 4.49, 9.56. IR (KBr): 1927 (vs),

1824 (vs) νCO, 1730 (s) νC=O-ester, 1614 (s) νC=O- carboxylate; (THF): ν = 1928 (vs), 1834

(vs) νCO, 1743 (s) νC=O-ester, 1652 νC=O- carboxylate. MS (ESI-pos; MeOH): m/z = 436

[M+H]+, 458 [M+Na]+, 474 [M+K]+, 568 [M+Cs]+, 871 [2M+H]+. 1H NMR (400.13 MHz,

DMSO-d6): 8.62 / 8.02 (1H, s, CεH), 7.12 / 6.92 (1H, s, CδH), 4.69 / 3.95 (1H, br, NH2), 4.38 /

4.19 (2H, m, Nε-CH2), 4.09 / 3.92 (1H, br, NH2), 3.50 (1H, m, CαH), 2.94 / 2.80 (2H, m, Nε-

CH2-CH2), 2.85 / 2.76 (2H, m, CβH2), 3.51 / 3.51 (1H, m, Hc-allyl), 3.28 / 3.17 (1H, m, Ha/s-

allyl), 3.11 / 2.78 (1H, m, Ha/s-allyl), 1.22 / 0.80 (1H, m, Ha/s-allyl), 0.99 / 0.80 (1H, m, Ha/s-

allyl). 13C NMR (100.6 MHz, DMSO-d6): δ = 229.7 / 228.6, 227.9 / 227.1 (CO), 178.0 / 178.9

C=OHis, 171.0 C=Oester, 140.3 / 141.9 Cε, 134.1 / 135.5 (Cγ), 117.5 (Cδ), 68.2 / 74.1 (Cc-allyl),

57.7 / 60.1 (C-allyl), 51.6 / 54.7 (C-allyl), 52.0 / 51.8 (Cα), 51.6 (OCH3), 42.6 / 42.3 (Nε-

CH2), 34.5 / 34.3 (Nε-CH2-CH2), 27.9 / 27.9 (Cβ).


167

Synthesis of Mo(L-His-Nε-C2H4CO2Me)(η-2-Me-allyl)(CO)2 (8)

To a solution of 6 (1.24 g, 3.4 mmol) in DMF (20 mL) was added 3-bromo-propionic acid

methyl ester (0.8 mL, 7.32 mmol) and Cs2CO3 (1.12 g, 3.4 mmol), followed by heating the

mixture at 80° C for 2 hrs. After the mixture was allowed to cool down to RT, it was filtered.

The filtrate was concentrated to dryness in vacuo and re-suspended in THF (35 mL). The

suspension was filtered to remove the Cs-salts, followed by removal of the solvent in vacuo.

Purification by preparative HPLC afforded 1.27 g (83%) of 8. Crystals suitable for X-ray

structure determination were grown by slow evaporation of a MeOH/H2O (1/1 v/v) mixture

under a stream of argon.

C16H21MoN3O6 = 447.3 g mol-1.


(m), 3248 (m) νNH, 1924 (vs), 1838 (vs), 1816 (vs) νCO, 1724 (s) νC=O-ester, 1633 (s) νC=O-

carboxylate; (CH2Cl2): ν = 1932, 1835 νCO. MS (FAB-pos.; dimethoxybenzylalcohol): m/z =

450 [M+H]+, 899 [2M+H]+, 422 [M-CO+H]+, 394 [M-2CO+H]+; (FAB-neg.;

dimethoxybenzylalcohol) 448 [M-H]-, 897 [2M-H]-. 1H NMR (250.13 MHz, DMSO-d6): δ =

8.07 (1H, br, CεH), 6.99 (1H, br, CδH), 5.04 / 4.26 (1H, br, NH2), 4.26 (2H, br, NεCH2), 3.60

(3H, s, OCH3), 3.34 (1H, br, NH2), 3.37 (2H, br, CαH + HAllyl), 2.80 (2H, br, CβH2), 2.87 (2H,

br, Nε-CH2CH2), 2.48 (1H, br, HAllyl), 1.71 (3H, br, Allyl-CH3), 0.80 (2H, br, HAllyl). 13C NMR

(62.9 MHz, DMSO-d6): 228.5 (br, CO), 178.7 (br, C=OHis), 171.0 (C=O-ester), 140.0 (br, Cε),

134.9 (Cγ), 117.5 (Cδ), 79.2 (br, Callyl), 55.8 (br, Callyl), 52.0 (Cα), 51.6 (OCH3), 42.5 (Nε-

CH2), 34.5 (Nε-CH2-CH2), 27.9 (Cβ), 19.2 (CH3).

Synthesis of AsPh4[Mo(L-His)(CO)3] (9)

This compound was prepared as reported in the literature [156]. Crystals of 9⋅H2O suitable for

X-ray diffraction were grown by slow evaporation of a MeOH/H2O (9/1 v/v) mixture under a

stream of argon.


168

AsC33H28MoN3O5 = 717.5 g mol-1.1H NMR (400.13 MHz, DMSO-d6): δ = 12.13 (1H, br, NεH) 7.89 (4H, m, Ar-H), 7.78 (24H,

m, Ar-H), 7.62 (1H, s, CεH), 6.80 (1H, s, CδH), 3.92 (1H, m, CαH), 3.39 (1H, m, NH2), 3.12

(1H, d, 2JHH = 11.0 Hz, NH2), 2.91 (1H, pseudo-dd, CβH), 2.79 (pseudo-dd, 1H, CβH). 13C

NMR (100.6 MHz, DMSO-d6): δ = 233.2 (CO), 229.7 (CO), 178.5 (C=O), 136.5 (Cε), 135.8

(Cγ), 134.3 (CAr), 133.2 (CAr), 130.9 (CAr), 121.0 (CAr, q), 113.5 (Cδ), 52.4 (Cα), 28.8 (Cβ).

Synthesis of Mo(D-His)(η-allyl)(CO)2 (10)

This compound was prepared in a similar manner as its L-histidine analogue 5 by using D-

histidine hydrochloride instead of L-histidine and two equivalents of KOH instead of one.

Yield: 87%. The characterisation data for 10 are identical to those for 5 because the

complexes are enantiomers. X-ray quality crystals of 10⋅MeOH were obtained by evaporation

of a H2O/MeOH (1/3 v/v) solution under a stream of argon.

Synthesis of Mo(D-His-Nε-C2H4CO2Me)(η-allyl)(CO)2 (11)

This compound was prepared similarly to the synthesis of 7, but by using 10 instead of 6. The

characterisation data for 11 are identical to those for 7 because the complexes are

enantiomers.

Synthesis of Mo(L-His-Nε-C2H4COOH)(η-allyl)(CO)2 (12)

To a solution of 7 (0.43 g; 1.0 mmol) in MeOH (20 mL) was added a solution of 0.28 g (5.0

mmol) KOH in H2O (5 mL) and the resulting mixture was stirred overnight. The pH was

adjusted to 7 by addition of 1 M HCl and the solvent was subsequently removed in vacuo,

yielding compound 12 in quantitative yield as a mixture with KCl.

C14H17MoN3O6 = 421.3 g mol-1.

IR (KBr): ν = 1928 (s), 1826(s) νCO, 1600 (s, br) νC=O; (MeOH): ν = 1940 (s), 1849(s) νCO.

MS (FAB-pos., glycerine): m/z = 444 [M+H]+; (FAB-neg.; glycerine): m/z = 420 [M-H]-, 443


169

[M+Na-H]-. 13C NMR (125.8 MHz; CD3OD): δ = 230.0, 228.1 (CO), 182.7, 177.9 (C=O),

143.8 (Cε), 135.2 (Cγ), 119.2 (Cδ), 75.8 (Cc-allyl), 60.6 (C-allyl), 55.6 (C-allyl), 53.7 (Cα),

46.3 (Nε-CH2), 39.8 (Nε-CH2-CH2), 29.4 (Cβ-His).

Synthesis of Mo(D-His-Nε-C2H4COOH)(η-allyl)(CO)2 (13)

This complex was prepared similarly as 12, but by using complex 11 instead of 7. The

characterisation data for 13 and 12 are identical, because these compounds are enantiomers.

Synthesis of Mo(His-Nε-C2H4CO-Phe-OMe)(η-allyl)(CO)2 (L-His 14; D-His 15)

A 1.0 mmol sample of either 12 or 13 (as a mixture with KCl) was suspended in 20 mL of a

DMF / dipea mixture (3/1 v/v) and subsequently filtered to remove most of the salts. To the

resulting clear solution was added phenylalanine methyl ester hydrochloride (216 mg; 1.0

mmol) and TBTU (323 mg; 1.0 mmol) and the mixture was stirred for 30 minutes at room

temperature. Next, the solvent was removed in vacuo and the residue suspended in EtOH (25

mL), followed by filtration to remove insoluble material. The mixture was evaporated to

dryness and the resulting solid purified by preparative HPLC (MeOH / H2O as the eluent).

The pure compound was obtained by evaporation of the MeOH / H2O mixture in vacuo.

Yield: 0.30 g (51%). Crystals suitable for X-ray diffraction of 14 were obtained by slow

evaporation of the MeOH / H2O solution from the HPLC purification under a stream of argon.

14: C24H28MoN4O7 = 580.5 g mol-1.


(vs), 1927 (vs) νCO, 1748 (s), νC=O, ester, 1669 (s), νC=O, amide, 1627 (s) νC=O, His; (MeOH): ν =

1847, 1938 νCO. MS (FAB-pos.; dimethoxybenzylalcohol): m/z = 583 [M+H]+; (FAB-neg.;

dimethoxybenzylalcohol): m/z = 581 [M-H]-. 1H NMR (500.13 MHz; DMSO-d6): δ = 8.57 /

7.93 (1H, s, CεH), 8.55 / 8.49 (1H, d, 3JHH = 7.5 Hz, NH), 7.27 (2H, pseudo-t, Ar-H), 7.20

(3H, m, Ar-H), 6.95 / 6.78 (1H, s, CδH), 4.69 / 3.90 (1H, br, NH2), 4.48 (1H, m, Cα-PheH), 4.28

/ 4.10 (2H, m, Nε-CH2), 4.14 / 3.86 (1H, br, NH2), 3.57 (3H, s, OCH3), 3.37 (1H, m, Cα-HisH),

2.99 (1H, m, Cβ-PheH), 2.89 (1H, m, Cβ-PheH), 2.85 (1H, m, Cβ-HisH), 2.79 (1H, m, Cβ-HisH),


170

2.63 / 2.56 (2H, m, Nε-CH2CH2), 3.52 / 3.48 (1H, m, Hc-allyl), 3.30 / 3.14 (1H, m, Hs/a-allyl),

3.11 / 2.75 (1H, m, Hs/a-allyl), 1.19 (d, 3JHH = 9.3 Hz) / 0.78 (1H, m, Hs/a-allyl), 0.99 (d, 3JHH =

9.3 Hz) / 0.78 (1H, m, Hs/a-allyl). 13C NMR (125.8 MHz; DMSO-d6): δ = 229.7 / 228.6, 227.8

/ 227.8 (CO), 177.9 / 178.9 (C=OHis), 171.9 (C=O-ester), 169.5 (C=Oamide), 141.7 (br) / 140.1

(Cε), 137.0 (CAr-Phe, q), 135.4 / 134.0 (Cγ), 129.0 (CAr-Phe), 128.3 (CAr-Phe), 126.6 (CAr,-Phe),

117.5 (Cδ), 68.1 / 74.1 (Cc-allyl), 60.3 (br) / 57.7 (C-allyl), 54.9 (br) / 51.8 (C-allyl), 53.7 (Cα-

Phe), 52.0 / 51.7 (Cα-His), 51.8 (OCH3), 43.0 / 42.7 (Nε-CH2), 36.7 (Cβ-Phe), 35.8 / 35.6 (Nε-

CH2CH2), 27.9 (Cβ-His).

15: C24H28MoN4O7 = 580.5 g mol-1.


(vs), 1927 (vs) νCO, 1740 (s), νC=O, ester, 1669 (s), νC=O, amide, 1651 (s), 1627 (s) νC=O, His;

(MeOH): ν = 1847, 1938 νCO. MS (FAB-pos.; dimethoxybenzylalcohol): m/z = 583 [M+H]+;

(FAB-neg.; dimethoxybenztlalcohol): m/z = 581 [M-H]-. 1H NMR (500.13 MHz; DMSO-d6):

δ = 8.59 / 7.96 (1H, s, CεH), 8.54 / 8.49 (1H, d, 3JHH = 7.5 Hz, NH), 7.28 (2H, pseudo-t, Ar-

H), 7.20 (3H, m, Ar-H), 6.93 / 6.76 (1H, s, CδH), 4.70 / 3.91 (1H, br, NH2), 4.51 (1H, m, Cα-

Phe), 4.27 / 4.10 (2H, m, Nε-CH2), 4.12 / 3.87 (1H, br, NH2), 3.59 (3H, s, OCH3), 3.40 (1H, m,

Cα-HisH), 2.92 (1H, m, Cβ-PheH), 2.79 (1H, m, Cβ-PheH), 2.89 (1H, m, Cβ-HisH), 2.82 (1H, m, Cβ-

HisH), 2.60 / 2.55 (2H, m, Nε-CH2CH2), 3.52 / 3.48 (1H, m, Hc-allyl), 3.29 / 3.16 (1H, m, Hs/a-

allyl), 3.13 / 2.77 (1H, m, Hs/a-allyl), 1.20 (d, 3JHH = 9.3 Hz) / 0.78 (1H, m, Hs/a-allyl), 1.00 (d,3JHH = 9.3 Hz) / 0.78 (1H, m, Hs/a-allyl). 13C NMR (125.8 MHz; DMSO-d6): δ = 229.7 /

228.6, 227.9 / 227.8 (CO), 178.0 / 178.9 (C=OHis), 171.8 (C=Oester), 169.5 (C=Oamide), 141.7

(br) / 140.0 (Cε), 137.0 (CAr-Phe, q), 135.4 / 134.0 (Cγ), 129.0 (CAr-Phe), 128.3 (CAr-Phe), 126.6

(CAr-Phe), 117.5 (Cδ), 68.1 / 74.1 (Cc-allyl), 60.2 (br) / 57.7 (C-allyl), 54.8 (br) / 51.9 (C-allyl),

53.7 (Cα-Phe), 52.1 / 51.7 (Cα-His), 51.9 (OCH3), 43.0 / 42.8 (Nε-CH2), 36.7 (Cβ-Phe), 35.8 / 35.6

(Nε-CH2CH2), 28.0 (Cβ-His).

Synthesis of Mo(His-Nε-C2H4CO-Phe-Leu-OMe)(η-allyl)(CO)2 (L-His 16; D-His 17)

These compounds were prepared by an almost identical procedure as described for 14 and 15.

Instead of addition of phenylalanine methyl ester, a solution of 1 mmol of [H2-Phe-Leu-


171

OMe]CF3CO2 (obtained via Boc-deprotection of Boc-Phe-Leu-OMe as described for 4) in

DMF (10 mL) was added. The remaining part of the synthesis as well as the work-up is

identical to that described for 14 and 15. These compounds were also purified by preparative

HPLC. Yield: 0.32 g (46%).

16: C30H39MoN5O8 = 693.6 g mol-1.


(vs), 1931 (vs) νCO, 1738 (s) νC=O, ester, 1651 (br, s) νC=O; (MeOH): ν = 1938, 1847 νCO. MS

(FAB-pos.; dimethoxybenzylalcohol): m/z = 696 [M+H]+. 1H NMR (400.13 MHz; DMSO-

d6): δ = 8.57 / 7.89 (1H, s, CεH), 8.46 (1H, d, 3JHH = 7.5 Hz, NHLeu), 8.28 / 8.22 (1H, d, 3JHH =

8.3 Hz, NHPhe), 7.25 (4H, m, Ar-H), 7.18 (1H, m, Ar-H), 6.87 / 6.73 (1H, s, Cδ-HisH), 4.68 /

3.90 (1H, br, NH2), 4.61 (1H, m, Cα-PheH), 4.29 (1H, m, Cα-LeuH), 4.22 / 4.03 (2H, m, Nε-

CH2), 4.03 / 3.87 (1H, br, NH2), 3.60 (3H, s, OCH3), 3.37 (1H, m, Cα-HisH), 3.00 (1H, m, Cβ-

PheH), 2.78 (1H, m, Cβ-PheH), 2.77 (1H, m, Cβ-HisH), 2.72 (1H, m, Cβ-HisH), 2.59 / 2.49 (2H, m,

Nε-CH2CH2), 1.56 (3H, m, Cβ-LeuH2 + Cγ-LeuH), 0.89 (3H, d, 3JHH = 6.4 Hz, Leu-CH3), 0.84

(3H, d, 3JHH = 6.4 Hz, Leu-CH3), 3.52 / 3.48 (1H, m, Hc-allyl), 3.28 / 3.12 (1H, m, Hs/a-allyl),

3.08 / 2.74 (1H, m, Hs/a-allyl), 1.20 (d, 3JHH = 9.3 Hz) / 0.78 (1H, m, Hs/a-allyl), 0.97 (d, , 3JHH

= 9.3 Hz) / 0.78 (1H, m, Hs/a-allyl). 13C NMR (100.6 MHz; DMSO-d6): δ = 229.7 / 228.6,

227.9 / 227.8 (CO), 177.9 / 178.8 (C=OHis), 172.7 (C=Oester), 171.3, 169.1 (C=Oamide), 141.6

(br) / 140.0 (Cε), 137.7 (CAr-Phe, q), 135.4 / 134.0 (Cγ-His), 129.1 (CAr-Phe), 128.0 (CAr-Phe), 126.3

(CAr, Phe), 117.4 (Cδ-His), 68.1 / 74.0 (Cc-allyl), 60.1 (br) / 57.8 (C-allyl), 54.8 (br) / 51.8 (C-

allyl), 53.6 (Cα-Phe), 52.0 / 51.7 (Cα-His), 51.8 (OCH3), 50.3 (Cα-Leu) 43.1 / 42.9 (Nε-CH2), 39.8

(Cβ-Leu; obscured by the solvent signal), 37.7 (Cβ-Phe), 36.0 / 35.8 (Nε-CH2CH2), 27.9 (Cβ-His),

24.2 (Cγ-Leu), 22.7, 21.3 (both Leu-CH3).

17: C30H39MoN5O8 = 693.6 g mol-1.


(vs), 1901 (vs) νCO, 1737 (s) νC=O, ester, 1648 (br, s) νC=O; (MeOH): ν = 1846, 1938 νCO. MS

(FAB-pos.; dimethoxybenzylalcohol): m/z = 696 [M+H]+. 1H NMR (400.13 MHz; DMSO-

d6): δ = 8.58 / 7.94 (1H, s, CεH), 8.45 (1H, pseudo-s, NHLeu), 8.26 / 8.21 (1H, d, 3JHH = 8.1

Hz, NHPhe), 7.24 (4H, m, Ar-H), 7.18 (1H, m, Ar-H), 6.85 / 6.69 (1H, s, Cδ-HisH), 4.70 / 3.90

(1H, br, NH2), 4.61 (1H, m, Cα-PheH), 4.30 (1H, m, Cα-LeuH), 4.23 / 4.06 (2H, m, Nε-CH2),


172

4.02 / 3.87 (1H, br, NH2), 3.60 (3H, s, OCH3), 3.34 (1H, m, Cα-HisH), 3.02 (1H, m, Cβ-PheH),

2.76 (1H, m, Cβ-PheH), 2.74 (1H, m, Cβ-HisH), 2.68 (1H, m, Cβ-HisH), 2.60 / 2.49 (2H, m, Nε-

CH2CH2), 1.55 (3H, m, Cβ-LeuH2 + Cγ-LeuH), 0.89 (3H, d, 3JHH = 6.4 Hz, Leu-CH3), 0.84 (3H,

d, 3JHH = 6.4 Hz, Leu-CH3), 3.52 / 3.48 (1H, m, Hc-allyl), 3.28 / 3.12 (1H, m, Hs/a-allyl), 3.09 /

2.73 (1H, m, Hs/a-allyl), 1.20 (d, 3JHH = 9.3 Hz) / 0.79 (1H, m, Hs/a-allyl), 0.99 (d, 3JHH = 9.3

Hz) / 0.79 (1H, m, Hs/a-allyl). 13C NMR (100.6 MHz; DMSO-d6): δ = 229.7 / 228.6, 227.9 /

227.8 (CO), 178.0 / 178.9 (C=OHis), 172.7 (C=Oester), 171.3, 169.1 (C=Oamide), 141.7 (br) /

140.1 (Cε), 137.6 (CAr-Phe, q), 135.4 / 134.0 (Cγ-His), 129.2 (CAr-Phe), 128.0 (CAr-Phe), 126.3 (CAr,-

Phe), 117.4 (Cδ-His), 68.1 / 74.1 (Cc-allyl), 60.1 (br) / 57.7 (C-allyl), 54.9 (br) / 51.9 (C-allyl),

53.6 (Cα-Phe), 52.0 / 51.7 (Cα-His), 51.8 (OCH3), 50.3 (Cα-Leu) 43.1 / 42.9 (Nε-CH2), 39.5 (Cβ-

Leu; obscured by the solvent signal), 37.8 (Cβ-Phe), 36.1 / 35.8 (Nε-CH2CH2), 28.0 (Cβ-His), 24.2

(Cγ-Leu), 22.7, 21.4 (both Leu-CH3).

General synthesis of the [Leu]-enkephalin conjugates with Mo(His-C2H4-CO-R)(η-

allyl)(CO)2

(L-His, 2-ClTrt-Tyr 18; D-His, 2-ClTrt-Tyr 19; L-His, Tyr 20; D-His, Tyr 21)

These compounds were prepared by solid phase peptide synthesis methods. After the

automated coupling steps (vide supra), the 2-ClTrt group was first cleaved from the resin-

bound [Leu]-enkephalin with CF3COOH and tri-isopropylsilane in CH2Cl2 (5/5/90 v/v/v) in

the case of the compounds 20 and 21, as described above. For the synthesis of 18 and 19 this

step was omitted. Subsequently the Fmoc protecting group was removed by reacting the resin-

bound peptide with piperidine in DMF (1/3 v/v), as described previously (vide supra).

Thereafter, the resin-bound peptide was reacted with five equivalents of either 12 or 13 and

six equivalents of TBTU in 15 mL of a DMF/dipea mixture (4/1 v/v) for 16 hours. Cleavage

of the resin bound conjugates was affected by treatment of the resin with a saturated NH3

solution in MeOH for 48 hours. The resin was removed by filtration, the solvent removed in

vacuo and the resulting crude product purified by preparative HPLC (MeOH / H2O) as the

eluent. The pure product was obtained by evaporation of the MeOH / H2O solution in vacuo.

Yield: approximately 100 mg for 18 and 19 and approximately 65 mg for 20 and 21.


173

18: C61ClH66N9O11Mo = 1232.6 g mol-1.

IR (KBr): ν = 1931 (s), 1831 (s) νCO, 1654 (vs) νC=O; (MeOH): ν = 1846, 1939 νCO. MS (ESI-

pos.; MeOH): m/z = 1256.5 [M+Na]+. 1H NMR (500.35 MHz; DMSO-d6): δ = 8.59 / 7.95

(CεH), 8.24 (1H, m, NHGly), 8.21 / 8.18 (1H, br, NHTyr), 8.07 (1H, d, 3JHH = 8,0 Hz, NHPhe),

7.96 (2H, m, NHGly + NHLeu), 7.76 (1H, d, 3JHH = 7.5 Hz, Ar-H), 7.35 (5H, m, Ar-H), 7.27

(6H, m, Ar-H), 7.23 (6H, m, Ar-H), 7.16 (1H, m, Ar-H-Phe), 7.08 (1H, br, O=CNH2), 6.97 /

6.80 (1H, s, Cδ-HisH), 6.95 (1H, br, O=CNH2), 6.88 (2H, pseudo-t, Ar-H-Tyr), 6.54 (2H,

pseudo-d, Ar-H-Tyr), 4.72 / 3.95 (1H, br, His-NH2), 4.50 (1H, m, Cα-PheH), 4.44 (1H, m, Cα-

TyrH), 4.20 (1H, m, Cα-LeuH), 4.16 / 4.01 (2H, m, NεCH2), 4.14 / 3.92 (1H, br, His-NH2), 3.73

(2H, m, Gly-CH2), 3.60 (2H, m, Gly-CH2), 3.30 (1H, m, Cα-His H), 3.02 (1H, m, Cβ-PheH),

2.83-2.76 (4H, overlapping m, Cβ-PheH + Cβ-TyrH + Cβ-HisH2), 2.62-2.47 (3H, overlapping m,

Cβ-TyrH + Nε-CH2CH2), 1.56 (1H, m, Cγ-LeuH), 1.46 (2H, m, Cβ-LeuH2), 0.87 (3H, d, 3JHH = 6.5

Hz, Leu-CH3), 0.82 (3H, d, 3JHH = 6.5 Hz, Leu-CH3), 3.53 / 3.49 (1H, m, Hc-allyl), 3.34 / 3.15

(Hs/a-allyl), 3.12 / 2.76 (1H, m, Hc-allyl), 1.20 (d, 3JHH = 9.1 Hz) / 0.79 (1H, m, Hs/a-allyl),

1.01 (d, 3JHH = 9.1 Hz) / 0.79 (1H, m, Hs/a-allyl). 13C NMR (125.8 MHz; DMSO-d6): δ =

229.7 / 228.6, 228.6 / 227.8 (CO), 178.0 / 178.9 (C=OHis), 173.9, 171.5, 170.6, 169.2, 168.9,

168.6 (C=O), 153.7, 143.0, 142.9, 138.5, 137.7, 133.6, 132.1, 132.0, 130.5, 130.0, 129.2,

129.0, 128.0, 127.8, 127.7, 127.7, 127.0, 126.3, 126.2, 119.4 (CAr), 140.0 / 141.5 (br) (Cε),

135.5 / 134.1 (Cγ-His), 117.5 (br) (Cδ-His), 89.2 (Cq-2-ClTrt), 74.1 / 68.2 (Cc-allyl), 60.2 /

57.7 (br) (C-allyl), 54.8 (br) / 51.9 (C-allyl), 54.0, 53.9 (Cα-Phe + Cα-Tyr), 52.1 / 51.7 (Cα-

His), 51.0 (Cα-Leu), 43.1 / 42.9 (Nε-CH2), 42.0, 41.9 (CH2-Gly), 40.8 (Cβ-Leu), 37.3, 36.6

(Cβ-Phe + Cβ-Tyr), 36.1 / 35.9 (Nε-CH2-CH2), 28.0 (Cβ-His), 24.2 (Cγ-Leu), 23.0, 21.6 (Both

Leu-CH3).

19: C61ClH66N9O11Mo = 1232.6 g mol-1.

IR (KBr): ν = 1931 (s), 1831 (s) νCO, 1654 (vs) νC=O; (MeOH): ν = 1846, 1938 νCO. MS (ESI-

pos.; MeOH): m/z = 1256.5 [M+Na]+. 1H NMR (500.35 MHz; DMSO-d6): δ = 8.58 / 7.95

(CεH), 8.24 (1H, m, NHGly), 8.21 / 8.17 (1H, br, NHTyr), 8.06 (1H, d, 3JHH = 8,1 Hz, NHPhe),

7.96 (2H, m, NHGly + NHLeu), 7.75 (1H, d, 3JHH = 7.6 Hz, Ar-H), 7.34 (5H, m, Ar-H), 7.28

(6H, m, Ar-H), 7.22 (6H, m, Ar-H), 7.16 (1H, m, Ar-H-Phe), 7.07 (1H, br, O=CNH2), 6.95 /

6.77 (1H, s, Cδ-HisH), 6.94 (1H, br, O=CNH2), 6.87 (2H, pseudo-t, Ar-H-Tyr), 6.54 (2H,

pseudo-d, Ar-H-Tyr), 4.70 / 3.91 (1H, br, His-NH2), 4.49 (1H, m, Cα-PheH), 4.43 (1H, m, Cα-


174

TyrH), 4.19 (1H, m, Cα-LeuH), 4.16 / 3.98 (2H, m, NεCH2), 4.14 / 3.88 (1H, br, His-NH2), 3.72

(2H, m, Gly-CH2), 3.60 (2H, m, Gly-CH2), 3.30 (1H, m, Cα-His H), 3.02 (1H, m, Cβ-PheH),

2.83-2.75 (4H, overlapping m, Cβ-PheH + Cβ-TyrH + Cβ-HisH2), 2.60-2.45 (3H, overlapping m,

Cβ-TyrH + Nε-CH2CH2), 1.55 (1H, m, Cγ-LeuH), 1.46 (2H, m, Cβ-LeuH2), 0.86 (3H, d, 3JHH = 6.5

Hz, Leu-CH3), 0.82 (3H, d, 3JHH = 6.5 Hz, Leu-CH3), 3.52 / 3.48 (1H, m, Hc-allyl), 3.34 / 3.15

(Hs/a-allyl), 3.12 / 2.76 (1H, m, Hc-allyl), 1.21 (d, 3JHH = 9.1 Hz) / 0.79 (1H, m, Hs/a-allyl),

1.01 (d, 3JHH = 9.1 Hz) / 0.79 (1H, m, Hs/a-allyl). 13C NMR (125.8 MHz; DMSO-d6): δ =

229.7 / 228.6, 228.6 / 227.8 (CO), 177.9 / 178.9 (C=OHis), 173.9, 171.5, 170.7, 169.2, 169.0,

168.6 (C=O), 153.7, 143.0, 142.9, 138.5, 137.7, 133.6, 132.1, 131.9, 130.4, 130.0, 129.2,

128.9, 128.0, 127.8, 127.7, 127.7, 127.0, 126.3, 126.2, 119.4 (CAr), 140.0 / 141.6 (br) (Cε-

His), 135.5 / 134.0 (Cγ-His), 117.5 (br) (Cδ-His), 89.2 (Cq-2-ClTrt), 74.1 / 68.1 (Cc-allyl), 60.2

/ 57.7 (br) (C-allyl), 54.9 (br) / 51.9 (C-allyl), 54.0, 53.9 (Cα-Phe + Cα-Tyr), 52.1 / 51.7 (Cα-

His), 51.0 (Cα-Leu), 43.1 / 42.9 (Nε-CH2), 42.0, 41.8 (CH2-Gly), 40.8 (Cβ-Leu), 37.3, 36.5

(Cβ-Phe + Cβ-Tyr), 36.1 / 35.9 (Nε-CH2-CH2), 28.0 (Cβ-His), 24.2 (Cγ-Leu), 23.0, 21.6 (Both

Leu-CH3).

20: C42H53N9O11Mo = 955.9 g mol-1.

IR (KBr): ν = 1930 (s), 1830 (s) νCO, 1654 (br, vs) νC=O; (MeOH): ν = 1938, 1846 νCO. MS

(ESI-pos.; MeOH): m/z = 958 [M+H]+, 980 [M+Na]+; exact mass for the [M+Na]+ fragment:

980.2826; C42H53N9NaO11Mo requires 980.2817. 1H NMR (500.35 MHz; DMSO-d6): δ =

9.19 / 9.15 (1H, s, Tyr-OH), 8.58 / 7.94 (CεH), 8.26 (1H, m, NHGly), 8.24 / 8.19 (1H, br,

NHTyr), 8.06 (1H, d, 3JHH = 8.0 Hz, NHPhe), 7.95 (2H, m, NHGly + NHLeu), 7.23 (4H, m, Ar-

HPhe), 7.16 (1H, m, Ar-HPhe), 7.07 (1H, s, O=CNH2), 6.98 (2H, pseudo-d, Ar-HTyr), 6.93 (1H,

br, O=CNH2), 6.91 / 6.76 (1H, s, Cδ-HisH), 6.60 (2H, pseudo-t, Ar-HTyr), 4.70 / 3.90 (1H, br,

His-NH2), 4.50 (1H, m, Cα-PheH), 4.47 (1H, m, Cα-TyrH), 4.23 / 4.06 (2H, m, NεCH2), 4.19

(1H, m, Cα-LeuH), 4.09 / 3.93 (1H, br, His-NH2), 3.75-3.58 (4H, m, both Gly-CH2), 3.32 (1H,

m, Cα-His H), 3.02 (1H, m, Cβ-PheH), 2.91 (1H, m, Cβ-TyrH), 2.80-2.73 (3H, overlapping m, Cβ-

PheH + Cβ-HisH2), 2.67-2.50 (3H, overlapping m, Cβ-TyrH + Nε-CH2CH2), 1.56 (1H, m, Cγ-LeuH),

1.46 (2H, m, 2 Cβ-LeuH), 0.87 (3H, d, 3JHH = 6.5 Hz, Leu-CH3), 0.81 (3H, d, 3JHH = 6.5 Hz,

Leu-CH3), 3.52 / 3.48 (1H, m, Hc-allyl), 3.34 / 3.15 (Hs/a-allyl), 3.12 / 2.76 (1H, m, Hs/a-allyl),

1.20 (d, 3JHH = 9.2 Hz) / 0.79 (1H, Hs/a-allyl), 1.00 (d, 3JHH = 9.1 Hz) / 0.79 (1H, m, Hs/a-allyl).


175

13C NMR (125.8 MHz; MeOH-d4): δ = 230.2, 228.3 (CO), 182.6 (C=OHis), 177.3, 174.6,

173.4, 172.5, 172.3, 171.9 (C=O), 157.3, 138.3, 131.3, 130.4, 129.6, 128.8, 127.9, 116.4

(CAr), 143.9 (Cε), 135.4 (Cγ-His), 119.3 (Cδ-His), 76.0 (Cc-allyl), 60.8 (C-allyl), 57.2, 56.6

(Cα-Phe + Cα-Tyr), 55.6 (C-allyl), 54.0 (Cα-His), 53.2 (Cα-Leu), 44.6 (Nε-CH2), 44.0, 43.7

(CH2-Gly), 41.7 (Cβ-Leu), 38.4, 37.2 (Cβ-Phe + Cβ-Tyr), 37.8 (Nε-CH2-CH2), 29.5 (Cβ-His),

25.8 (Cγ-Leu), 23.6, 21.8 (Both Leu-CH3). Because the a / b ratio in MeOH-d4 is

approximately 4:1, only resonances owing to isomer a were detected with certainty.

21: C42H53N9O11Mo = 955.9 g mol-1.

IR (KBr): ν = 1931 (s), 1831 (s) νCO, 1653 (br, vs) νC=O; (MeOH): ν = 1848, 1939 νCO. MS

(ESI-pos.; MeOH): m/z = 958 [M+H]+, 980 [M+Na]+. 1H NMR (500.35 MHz; DMSO-d6): δ =

9.25 / 9.18 (1H, s, Tyr-OH), 8.59 / 7.96 (CεH), 8.26 (1H, m, NHGly), 8.25 / 8.18 (1H, br,

NHTyr), 8.06 (1H, d, 3JHH = 8.1 Hz, NHPhe), 7.95 (2H, m, NHGly + NHLeu), 7.24 (4H, m, Ar-

HPhe), 7.16 (1H, m, Ar-HPhe), 7.07 (1H, s, O=CNH2), 6.98 (2H, pseudo-d, Ar-HTyr), 6.93 (1H,

br, O=CNH2), 6.81 / 6.68 (1H, s, Cδ-HisH), 6.60 (2H, pseudo-t, Ar-HTyr), 4.70 / 3.90 (1H, br,

His-NH2), 4.49 (1H, m, Cα-PheH), 4.46 (1H, m, Cα-TyrH), 4.23 / 4.05 (2H, m, NεCH2), 4.19

(1H, m, Cα-LeuH), 4.09 / 3.93 (1H, br, His-NH2), 3.73-3.58 (4H, m, both Gly-CH2), 3.30 (1H,

m, Cα-His H), 3.02 (1H, m, Cβ-PheH), 2.90 (1H, m, Cβ-TyrH), 2.78-2.72 (3H, overlapping m, Cβ-

PheH + Cβ-HisH2), 2.65-2.50 (3H, overlapping m, Cβ-TyrH + Nε-CH2CH2), 1.56 (1H, m, Cγ-LeuH),

1.46 (2H, m, Cβ-LeuH2), 0.87 (3H, d, 3JHH = 6.5 Hz, Leu-CH3), 0.81 (3H, d, 3JHH = 6.5 Hz,

Leu-CH3), 3.52 / 3.48 (1H, m, Hc-allyl), 3.34 / 3.15 (Hs/a-allyl), 3.12 / 2.76 (1H, m, Hs/a-allyl),

1.20 (d, 3JHH = 9.2 Hz) / 0.79 (1H, Hs/a-allyl), 1.00 (d, 3JHH = 9.1 Hz) / 0.79 (1H, m, Hs/a-allyl).13C NMR (500.35 MHz; MeOH-d4): δ = 230.2, 228.3 (CO), 182.6 (C=OHis), 177.3, 174.6,

173.5, 172.5, 172.3, 171.8 (C=O), 157.3, 138.3, 131.3, 130.4, 129.6, 128.8, 127.9, 116.4 (CAr-

Phe + Tyr), 144.0 (Cε-His), 135.5 (Cγ-His), 119.1 (Cδ-His), 75.9 (Cc-allyl), 60.8 (C-allyl),

57.1, 56.6 (Cα-Phe or Cα-Tyr), 55.6 (C-allyl), 53.8 (Cα-His), 53,2 (Cα-Leu) 44.6 (Nε-CH2),

44.0, 43.6 (CH2-Gly), 41.7 (Cβ-Leu), 38.4, 37.2 (Cβ-Phe or Cβ-Tyr), 37.8 (Nε-CH2-CH2), 29.4

(Cβ-His), 25.8 (Cγ-Leu), 23.6, 21.8 (Both Leu-CH3). Because the a / b ratio in MeOH-d4 is

approximately 4:1, only resonances originating from isomer a were detected with certainty.


176

Synthesis of Mo(bpa)(CO)3 (22)

A suspension of di(2-picolyl)amine (1.0 g; 5 mmol) and Mo(CO)6 (1.32 g; 5 mmol) in

degassed di-n-butyl ether (75 mL) was heated at 150°C for 1 hr, during which time a deep-

yellow precipitate formed. After allowing the mixture to reach room temperature, the solid

material was collected by filtration and dried in vacuo. Yield: 1.8 g (95%). Crystals suitable

for X-ray analysis were grown by slow evaporation of a CH3CN / H2O mixture (5/1 v/v)

under a stream of argon.

C15H13MoN3O3 = 379.2 g mol-1.


(w) νNH, 1901 (s), 1891 (vs), 1784 (s), 1769 (s), 1742 (vs) νCO; (THF): ν = 1904 (vs), 1790

(s), 1779 (s) νCO; MS (EI): m/z = 381 (14) [M]+, 353 (2) [M-CO]+, 325 (1) [M-2 CO]+, 297

(31) [M-3 CO]+, 93 (100). 1H NMR (400.13 MHz, DMSO-d6): δ = 8.73 (2H, apparent-d,

HPyr), 7.65 (2H, m, HPyr), 7.27 (2H, apparent-d, HPyr), 7.14 (2H, pseudo-t, HPyr), 6.14 (1H, t,3JHH = 6.6 Hz, NH), 4.33 (2H, dd, 2JHH = 16.8 Hz, 3JHH = 6.6 Hz, CH2), 4.06 (2H, d, 2JHH =

16.8 Hz, CH2). 13C NMR (DMSO-d6, 62.9 MHz): δ = 231.4, 229.3 (CO), 159.5, 150.1, 137.5,

123.0, 122.0 (CPyr), 59.6 (CH2). 95Mo NMR (CH3CN, 32.6 MHz): δ = -938 (∆1/2 = 36 Hz).

Synthesis of Mo(benzyl-bpa)(CO)3 (23)

A suspension of N-benzyl-N,N-di(2-picolyl)amine (0.96 g; 3.3 mmol) and Mo(CO)6 (0.88 g;

3.3 mmol) in deoxygenated mesitylene (40 mL) was heated at 150° C for 1 hr, during which

time an orange solid precipitated. After allowing the mixture to cool down to ambient

temperature, the precipitate was isolated by filtration, washed with n-hexane (10 mL) and

dried in vacuo. Yield: 1.28 g (83%). Crystals suitable for X-ray structure determination were

grown by slow evaporation of a CH3CN / H2O (10 / 1 v / v) solution under a stream of argon.

C22H19MoN3O3 = 469.4 g mol-1.


(vs), 1776 (s), 1752 (vs) νCO; (THF): ν = 1906 (vs), 1794 (s), 1782 (s) νCO. MS (EI): m/z =

471 (6) [M]+, 443 (4) [M-CO]+, 387 (13) [M-3 CO]+, 197 (100). 1H NMR (400.13 MHz,

DMSO-d6) 8.68 (2H, apparent-d, HPyr), 7.68 (2H, m, HAr), 7.59 (2H, m, HAr), 7.46 (3H, m,


177

HAr), 7.16 (2H, pseudo-d, HAr), 7.08 (2H, m, HAr), 4.67 (2H, d, 2JHH = 15.5 Hz, CH2-picolyl),

4.66 (2H, s, CH2-benzyl), 3.72 (2H, d, 2JHH = 15.5 Hz, CH2-picolyl). 13C NMR (125.8 MHz,

CD3CN): δ = 232.3, 231.3 (CO), 159.9, 151.6, 138.4, 133.6, 132.5, 129.7, 129.4, 124.1, 123.3

(CAr), 71.0 (benzyl-CH2), 65.4 (picolyl-CH2). 95Mo (CH3CN, 32.6 MHz): δ = -849 (∆1/2 = 47

Hz).

Synthesis of [Mo(bpa)(CO)3Br]Br3 (24)

To a suspension of Mo(bpa)(CO)3 (22) (0.5 g; 1.3 mmol) in degassed CHCl3 (50 mL) was

added dropwise Br2 (1.0 g; 0.3 mL; 6.3 mmol), resulting in a clear orange-red solution. The

mixture was stirred for 45 mins, during which time an orange-brown powder precipitated. The

solid was collected by filtration, washed with CHCl3 (10 mL) and pentane (10 mL) and air-

dried. Yield: 0.66 g (73%).

Br4C15H13MoN3O3 = 698.8 g mol-1.


(w) νNH, 2037 (s), 1950 (br, vs) νCO; (THF): ν = 2055 (s), 1989 (s), 1935 (s) νCO. MS (ESI-

pos., CH2Cl2): m/z = 460 [Mo(bpa)(CO)3Br]+, 432 [Mo(bpa)(CO)2Br]+, 404

[Mo(bpa)(CO)Br]+, 376 [Mo(bpa)Br]+.

Synthesis of [(Mo(bpa)2(O)2(µ-O)2]Br, PF6 (25)

[Mo(bpa)(CO)3Br]Br3 (24) (0.5 g; 0.7 mmol) was dissolved in MeOH (30 mL), followed by

addition of a KPF6 (0.66 g; 3.6 mmol) solution in H2O (10 mL). Upon standing overnight an

orange powder formed, which was isolated by filtration, washed with Et2O (15 mL) and air-

dried. Yield: 0.21 g (68%). Crystals suitable for X-ray diffraction were obtained by slow

evaporation of a MeOH solution.

BrC24F6H26Mo2N6O4P = 879.3 g mol-1.

Anal. Calcd.: C 32.78, H 2.98, N 9.56; Found: 32.67, H 2.91, N 9.47. IR(KBr): ν = 962 (m)

νMo=O, 747 (m) νasM-O. MS (ESI-pos., MeOH): m/z = 737 [M –PF6]+, 657 [M – PF6 – HBr]+,


178

327 [M – PF6 – Br]2+. 13C NMR (DMSO-d6; 62.9 MHz): δ = 158.6, 151.7, 142.4, 125.7, 124.2

(CPyr), 55.3 (CH2). 1H NMR (DMSO-d6; 250.13 MHz) shows broadened signals at room

temperature. Could be due to paramagnetic impurities or coalescence as found for a related

compound [209].

Synthesis of N-(4-carboxymethyl)benzyl-N, N-di(2-picolyl)amine (26)

To a solution of 1.00 g di(2-picolyl)amine (5 mmol) and 1.15 α-bromo-toluic acid methyl

ester (5 mmol) in THF (35 mL) was added 0.69 mL (5 mmol) of NEt3 and the mixture was

refluxed for 1.5 hrs. The mixture was allowed to reach room temperature and subsequently

filtered to remove a white precipitate. After removal of the solvent under reduced pressure,

the oily residue was re-dissolved in 40 mL of diethylether and filtered to remove a red gooey

solid. Evaporation of the solvent yielded 1.4 g (81 %) of a light-orange oil, which was used as

such in the next step.

C21H21N3O2 = 347.4 g mol-1.

MS (EI): m/z = 347 (2) [M+], 316 (2) [M – OCH3]+, 255 (100) [M – C6H6N]+. 1H NMR (250.1

MHz, CDCl3): δ = 8.49 (2H , apparent-d, Hpyr), 7.95 (2H, d, 3JHH = 8.3 Hz, HAr), 7.62 (2H,

pseudo-t, Hpyr), 7.52 (2H, pseudo-d, Hpyr) 7.46 (2H, d, 3J = 8.3 Hz, HAr), 7.12 (2H, pseudo-t,

Hpyr), 3.87 (s, 3H, OCH3), 3.78 (s, 4H, picolyl-CH2), 3.71 (s, 2H, CH2). 13C (100.6 MHz,

CDCl3): δ = 166.9 (C=O), 159.3 (Cq, pyr), 148.9 (CPyr), 144.5 (CAr, q), 136.3 (CPyr), 129.5

(CAr), 128.9 (CAr, q), 128.6 (CAr), 122.7 (CPyr), 122.0 (CPyr), 60.0 (CH2, picolyl), 58.1 (CH2),

51.9 (OCH3).

Synthesis of Mo(4-CO2Me-benzyl-bpa)(CO)3 (27)

This compound was synthesised analogously to 23, by using 1.15 g (3.3 mmol) of N-(4-

carboxymethyl)benzyl-N, N-di(2-picolyl)amine (26) and 0.88 g Mo(CO)6 (3.3 mmol). Yield:

1.50 g (86%).


179

C24H21MoN3O5 = 527.4 g mol-1.


(vs), 1787 (vs), 1761 (vs) νCO, 1719 (m) νC=O; (THF): ν = 1907 (vs), 1794 (s), 1782 (vs) νCO,

1727 (m) νC=O. MS (EI): m/z = 529 (3) [M]+, 501 (2) [M- CO]+, 445 (3) [M – 3 CO]+, 255

(100). 1H NMR (250.13 MHz, DMSO-d6) δ = 8.68 (2H, apparent-d, Hpyr), 8.02 (2H, d, 3JHH =

8.1 Hz, HAr), 7.85 (2H, d, 3JHH = 8.1 Hz, HAr), 7.59 (2H, pseudo-t, HPyr), 7.15 (2H, pseudo-d,

HPyr), 7.09 (2H, pseudo-t, HPyr), 4.73 (2H, s, CH2-benzyl), 4.68 (2H, d, 2JHH = 15.6 Hz, CH-

picolyl), 3.89 (3H, s, CH3), 3.75 (2H, d, 2JHH = 15.6 Hz, CH-picolyl). 13C NMR (62.9 MHz,

DMSO-d6) δ = 231.3, 230.3 (CO), 166.0 (C=O), 158.7, 150.0, 137.6, 132.0, 129.7, 129.1,

123.1, 122.4 (CAr) only 8 aromatic signals instead of the expected 9 were detected; two

resonances probably coincidentally overlap, 68.8 (CH2-benzyl), 64.2 (CH2-picolyl), 52.2

OCH3. 95Mo (32.6 MHz, CH3CN): δ = -847 (∆1/2 = 49 Hz).

Synthesis of N-4-benzylic acid-di(2-picolyl)amine (28)

To a solution of 26 (1.4 g; 4.0 mmol) in MeOH (20 mL) was added a solution of 0.8 g (20

mmol) of NaOH in 5 mL of H2O and the mixture was stirred for two hours at ambient

temperature. The pH was adjusted to 7 by dropwise addition of 2M HCl, followed by removal

of the solvent under reduced pressure. The sticky white residue was triturated with CHCl3

(200 mL), followed by filtration to remove NaCl and the CHCl3 solution was dried over

MgSO4. Removal of the CHCl3 solution under reduced pressure afforded a yellow sticky oil,

to which was added CH3CN (30 mL), followed by vigorous stirring. After approximately 15-

30 mins, a white precipitate formed. The solution was stored at 0° C for 2 hrs, to further affect

precipitation and thereafter the white solid was isolated by filtration and dried in vacuo. Yield

0.7 g (53%).

C20H19N3O2 = 333.4 g mol-1.

IR (KBr): ν = 1698 (m) νC=O. MS (EI): m/z = 333(1) [M]+, 241 (100) [M-C6H6N]+. 1H NMR

(250.1 MHz, CDCl3): δ = 11.35 (1H, br, CO2H), 8.59 (2H, apparent-d, HPyr), 8.01 (2H, d, 3JHH

= 8.0 Hz, HAr), 7.66 (2H, pseudo-t, HPyr), 7.59 (2H, apparent-d, HPyr), 7.44 (2H, d, 3JHH = 7.7

Hz, HAr), 7.80 (2H, pseudo-t, HPyr), 3.85 (4H, s, CH2-picolyl), 3.73 (2H, s, CH2-benzyl). 13C

NMR (62.9 MHz, CDCl3): δ = 169.3 (C=O), 159.8, 148.4, 143.1, 137.2, 130.0, 128.8, 123.4,


180

122.4, 122.0 (CAr), only 8 aromatic C-atoms detected whereas 9 signals are expected, 59.3

(CH2-picolyl), 58.2 (CH2-benzyl).

Synthesis of N-(4-C(O)Phe-OMe)benzyl-N,N-di(2-picolyl)amine (29)

To a suspension of 28 (333 mg; 1.0 mmol) in 10 mL of CH3CN was added NEt3 (1 mL; 0.73

g; 7.2 mmol), phenylalanine methyl ester hydrochloride (216 mg; 1.0 mmol) and TBTU (323

mg; 1.0 mmol) and the mixture was stirred for 30 min at ambient temperature. After

evaporation of the solvent in vacuo, CH2Cl2 (100 mL) was added to the sticky residue,

followed by filtration to remove triethyl ammonium salts. The CH2Cl2 solution was washed

with an aqueous 2M NaHCO3 solution (75 mL) and the phases were separated. The aqueous

phase was extracted with CH2Cl2 (2 x 75 mL) and the combined organic extracts were dried

over MgSO4. Most of the solvent was removed under reduced pressure, followed by further

evaporation of the solvent in vacuo, affording a light-yellow glassy solid. Yield: 440 mg

(89%).

C30H30N4O3 = 494.6 g mol-1.

MS (EI): m/z = 494 (3) [M]+, 463 (1) [M-OCH3]+, 435 (2) [M-CO2CH3]+, 402 (100) [M-

C6H6N]+. 1H NMR (250.13 MHz, CDCl3): δ = 8.49 (2H, m, Hpyr), 7.63 (4H, m, HAr), 7.50

(2H, d, 3JHH = 7.9 Hz, HAr), 7.43 (2H, d, 3JHH = 7.9 Hz, HAr), 7.24 (3H, m, HAr), 7.01 (4H, m,

HAr), 6.57 (1H, d, 3JHH = 8.2 Hz), 5.06 (1H, m, Cα), 3.76 (4H, s, CH2-picolyl), 3.72 (3H, s,

OCH3), 3.68 (2H, s, CH2-benzyl), 3.21 (2H, m, CβH2). 13C NMR (62.9 MHz, CDCl3): δ =

172.1 (C=Oester), 166.9 (C=Oamide), 159.0, 148.7, 142.9, 136.7, 136.0, 132.6, 129.1, 128.8,

128.5, 127.1, 127.0, 122.9, 122.2 (CAr), 59.7 (CH2-picolyl), 58.0 (CH2-benzyl), 53.6, 52.3 (Cα

+ OCH3), 37.6 (Cβ).

Synthesis of N-(4-C(O)-Ala-Phe-OMe)benzyl-N,N-di(2-picolyl)amine (30)

The dipeptide Boc-Ala-Phe-OMe (350 mg; 1 mmol) was dissolved in a mixture of DCM (5

mL) and CF3CO2H (10 mL) and the mixture was stirred for one hour at room temperature.

After removal of the solvent in vacuo, Et2O (15 mL) was added to the residue, followed by


181

removal of the solvent in vacuo. This last step was performed three times in total in order to

remove any traces of trifluoro acetic acid. The white solid residue was suspended in CH3CN

(10 mL), followed by addition of NEt3 (1.5 mL; 1.09 g; 10.8 mmol), bpa-acid 28 (333 mg; 1.0

mmol) and TBTU (323 mg; 1 mmol). After stirring of the mixture at ambient temperature for

30 mins, the solvent was removed in vacuo. To the sticky residue was added CH2Cl2 (100

mL), followed by filtration to remove triethylammonium salts. The CH2Cl2 solution was

washed with an aqueous 2 M NaHCO3 solution (75 mL), followed by separation of the

phases. The aqueous phase was extracted with DCM (2 x 75 mL) and the combined organic

extracts were dried over MgSO4. Most of the CH2Cl2 was removed under reduced pressure,

followed by further evaporation of the solvent in vacuo, affording a white glassy solid. Yield:

520 mg (92 %).

C33H35N5O4 = 565.7 g mol-1.

MS (EI): m/z = 565 (6) [M]+, 534 (3) [M-OCH3]+, 506 (1) [M-CO2CH3]+, 473 (100) [M-

C6H6N]+. 1H NMR (400.13 MHz; CDCl3): δ = 8.51 (2H, apparent-d, HPyr), 7.69 (4H, m, HAr),

7.53 (2H, d, 3JHH = 7.9 Hz, HAr), 7.46 (2H, d, 3JHH = 7.9 Hz, HAr), 7.18 (2H, m, HAr), 7.06

(5H, m, HAr), 6.92 (2H, m, both NH), 4.80 (1H, m, CαH), 4.66 (1H, m, CαH), 3.91 (4H, s,

CH2-picolyl), 3.82 (2H, s, CH2-benzyl), 3.67 (3H, s, OCH3), 3.09 (1H, m, Cβ-PheH), 3.00 (1H,

m, Cβ-PheH), 1.40 (3H, d, 3JHH = 7.0 Hz). 13C NMR (100.6 MHz; CDCl3): δ = 172.0, 171.7

(C=Oester + C=OAla-Phe), 166.7 (C=O), 157.7, 148.4, 137.3, 135.7, 132.9, 129.2, 129.1, 128.4,

127.3, 127.0, 123.5, 122.6 (CAr), only 12 aromatic signals detected instead of the expected 13,

59.2 (CH2-picolyl), 58.1 (CH2-benzyl), 53.4, 52.3 (OCH3 + Cα-Phe), 49.0 (Cα-Ala), 37.7 (Cβ-Phe),

18.2 (Ala-CH3).

Synthesis of (4-CO-Phe-OMe-benzyl-bpa)Mo(CO)3 (31)

To a stirred solution of Mo(CO)3(NCEt)3 (0.28 g ; 0.8 mmol) in 15 mL of THF was added

dropwise a solution of 29 (0.40 g; 0.8 mmol) in 10 mL of THF. The mixture was stirred for an

extra period of 10 mins at room temperature, during which time an orange precipitate formed.

The mixture was concentrated to about 15 mL in vacuo, and the orange solid collected by

filtration, washed with Et2O (10 mL) and dried in vacuo. Yield: 0.44 g (82 %).


182

C33H30N4O6 = 674.6 g mol-1.

Anal. Calcd.: C 58.76, H 4.48, N 8.31; Found: C 53.13, H 4.67, N 8.92; Repeated attempts did

not improve the C-values. IR (KBr): ν = 1900 (vs) νCO, 1779 (s) νCO, 1759 (s) νCO + νC=O;

(THF): ν = 1906 (vs), 1793 (s), 1782 (s) νCO. MS (FAB-pos.; dimethoxybenzylalcohol): m/z =

676 [M]+, 648 [M-CO]+. 1H NMR (CD3CN; 500.35 MHz): δ = 8.77 (2H, m, HPyr), 7.81 (2H,

d, 3JHH = 8.1 Hz, HAr), 7.62 (2H, d, 3JHH = 8.1 Hz, HAr), 7.49 (2H, m, HAr), 7.31 (5H, m, NH +

HAr-Phe), 7.24 (1H, m, HAr-Phe), 7.00 (4H, m, HPyr), 4.92 (1H, m, CαH), 4.76 (2H, s, CH2-

benzyl), 4.51 (2H, d, 2JHH = 15.4 Hz, CH2-picolyl), 3.73 (2H, dd, , 2JHH = 15.4 Hz, 4JHH = 1.6

Hz, CH2-picolyl), 3.69 (3H, s, OCH3), 3.32 (1H, m, CβH), 3.17 (1H, m, CβH). 13C NMR

(CD3CN; 125.8 MHz): δ = 232.3, 231.3 (CO), 173.1 (C=Oester), 167.4 (C=Oamide), 159.8,

151.7, 138.4, 138.3, 137.1, 135.0, 132.8, 130.2, 129.5, 128.2, 127.8, 124.1, 123.3 (CAr), 70.4

(CH2-benzyl), 65.5 (CH2-picolyl), 55.2, 52.9 (OCH3 + Cα), 38.0 (Cβ).

Synthesis of (4-CO-Ala-Phe-OMe-benzyl-bpa)Mo(CO)3 (32)

To a stirred solution of Mo(CO)3(NCEt)3 (0.28 g ; 0.8 mmol) in 15 mL of THF was added

dropwise a solution of 30 (0.40 g; 0.8 mmol) in 10 mL of THF. Stirring was continued for 10

mins at room temperature, during which time an orange precipitate formed. The mixture was

concentrated to about 15 mL in vacuo, and the orange solid collected by filtration, washed

with Et2O (10mL) and dried in vacuo. Yield: 0.47 g (79 %).

C36H35N5O7Mo = 745.6 g mol-1.


(vs) νCO, 1778 (s) νCO, 1761 (s) νCO + νC=O; (THF): ν = 1906 (vs), 1793 (s), 1782 (s) νCO. MS

(FAB-pos.; dimethoxybenzylalcohol): m/z = 747 [M]+, 719 [M-CO]+. 1H NMR (500.35 MHz;

DMSO-d6): δ = 8.69 (2H, apparent-d, HPyr), 8.49 (1H, d, 3JHH = 7.6 Hz, NH), 8.30 (1H, d,3JHH = 7.6 Hz, NH), 7.97 (2H, d, 3JHH = 8.1 Hz, HAr), 7.80 (2H, d, 3JHH = 8.1 Hz, HAr), 7.59

(2H, pseudo-t, HPyr), 7.22 (5H, m, HAr-Phe), 7.15 (2H, apparent-d, HPyr), 7.09 (2H, pseudo-t,

HAr), 4.72 (3H, s, CH2-benzyl + CH2-picolyl), 4.69 (1H, d, 4JHH = 3.4 Hz, CH2-picolyl), 3.71

(2H, dd, 2JHH = 15.7 Hz, 4JHH = 3.4 Hz, CH2-picolyl), 4.56 (1H, m, CαH), 4.50 (1H, m, CαH),

3.59 (3H, s, OCH3), 3.03 (1H, m, Cβ-PheH2), 2.99 (1H, m, Cβ-PheH2), 1.32 (3H, d, 3JHH = 7.2

Hz, Ala-CH3). 13C NMR (125.8 MHz; DMSO-d6): δ = 231.2, 230.3 (CO), 172.4, 171.8


183

(C=Oester + C=OAla-Phe), 165.6 (C=O), 158.8, 150.0, 137.5, 137.0, 135.4, 134.0, 131.4, 129.1,

128.2, 127.5, 126.5, 123.0, 122.3 (CAr), 68.9 (CH2-benzyl), 64.2 (CH2-picolyl), 53.6, 51.8

(Cα-Phe + OCH3), 48.6 (Cα-Ala), 36.6 (Cβ-Phe), 17.7 (Ala-CH3).

Synthesis of the [Leu]-enkephalin-N(4-CO-benzyl)-N,N-di(2-picolyl)amine conjugate (33)

This compound was prepared by solid phase peptide synthesis methods. After the automated

coupling steps, the 2-Cl-Trt was first cleaved from the resin-bound peptide with CF3COOH

and tri-isopropylsilane in CH2Cl2 (5/5/90 v/v/v) and subsequently the Fmoc protecting group

was removed by reacting the resin-bound peptide with piperidine in DMF (1/3 v/v), as

described previously (vide supra). Thereafter, the resin-bound peptide was reacted with five

equivalents of 28 and six equivalents of TBTU in 15 mL of a DMF/dipea mixture (4/1 v/v) for

16 hours. Cleavage of the resin bound di(2-picolyl)amine conjugate was affected by treatment

of the resin with a saturated NH3 solution in MeOH for 48 hours. The resin was removed by

filtration, the solvent removed in vacuo and the resulting crude product purified by

preparative HPLC (MeOH / H2O) as the eluent. The pure product was obtained by

evaporation of the MeOH / H2O solution in vacuo. Yield: 130 mg.

C48H55N9O7 = 870.0 g mol-1.

MS (ESI-pos.; MeOH): m/z = 870 [M+H]+, 876 [M+Li]+, 892 [M+Na]+; exact mass of the

[M+Li]+ fragment: 876.4380; C48H55LiN9O7 requires 876.4384. 1H NMR (500.35 MHz;

CD3OD): δ = 8.40 (2H, m, HPyr), 7.76 (2H, m, HPyr), 7.69 (2H, d, 3JHH = 8.2 Hz, HAr), 7.61

(2H, apparent-d, HPyr), 7.43 (2H, d, 3JHH = 8.2 Hz, HAr), 7.24 (2H, m, HPyr), 7.21 (4H, m, HAr-

Phe), 7.14 (1H, m, HAr-Phe), 7.07 (2H, d, 3JHH = 8.4 Hz, HAr-Tyr), 6.79 (2H, d, 3JHH = 8.4 Hz, HAr-

Tyr), 4.68 (1H, m, Cα-PheH or Cα-TyrH), 4.55 (1H, m, Cα-PheH or Cα-TyrH), 4.29 (1H, m, Cα-

LeuH), 3.83 (2H, m, CH2-Gly), 3.78 (2H, m, CH2-Gly), 3.74 (4H, s, CH2-picolyl), 3.67 (2H, s,

CH2-benzyl), 3.14 (2H, m, Cβ-PheH + Cβ-TyrH), 2.98 (2H, m, Cβ-PheH + Cβ-TyrH), 1.57 (3H, m,

Cβ-Leu + Cγ-Leu), 0.85 (3H, d, 3JHH = 6.0 Hz, Leu-CH3), 0.81 (3H, d, 3JHH = 6.0 Hz, Leu-CH3).

(No resonances owing to NH and Tyr-OH detected due to H / D exchange); (500.35 MHz;

DMSO-d6): δ = 8.48 (3H, m, NH + 2 HPyr), 8.27 (1H, br, NH), 8.08 (1H, d, 3JHH = 5.1 Hz,

NH), 8.02 (1H, br, NH), 7.95 (1H, d, 3JHH = 7.6 Hz, NH), 7.78 (2H, m, HPyr), 7.75 (2H,

apparent-d, HPyr), 7.56 (2H, d, 3JHH = 8.0 Hz, HAr), 7.46 (2H, d, 3JHH = 8.0 Hz, HAr), 7.25 (2H,


184

m, HPyr), 7.22 (4H, m, HAr-Phe), 7.15 (1H, m, HAr-Phe), 7.09 (2H, d, 3JHH = 8.3 Hz, HAr-Tyr), 7.06

(1H, br, NH), 695 (1H, br, NH), 6.61 (2H, d, 3JHH = 8.3 Hz, HAr-Tyr), 4.60 (1H, m, Cα-PheH or

Cα-TyrH), 4.49 (1H, m, Cα-PheH or Cα-TyrH), 4.19 (1H, m, Cα-LeuH), 3.73 (2H, m, CH2-Gly),

3.70 (4H, s, CH2-picolyl), 3.67 (2H, s, CH2-benzyl), 3.64 (2H, m, CH2-Gly), 3.02 (2H, m, Cβ-

PheH2 or Cβ-TyrH2), 2.87 (1H, m, Cβ-PheH or Cβ-TyrH), 2.79 (1H, m, Cβ-PheH or Cβ-TyrH), 1.54

(1H, m, Cγ-Leu), 0.85 (3H, d, 3JHH = 6.4 Hz, Leu-CH3), 0.80 (3H, d, 3JHH = 6.4 Hz, Leu-CH3);

Tyr-OH not detected. 13C NMR (100.6 MHz; DMSO-d6): δ = 177.2, 174.7, 173.3, 172.3,

171.9, 170.2 (C=O), 160.2, 157.4, 149.4, 144.2, 138.7, 138.2, 134.0, 131.4, 130.3, 130.0,

129.6, 129.0, 128.7, 127.9, 124.8, 123.9, 116.3 (CAr), 60.9 (CH2-picolyl), 59.4 (CH2-benzyl),

57.4, 56.6 (Cα-Phe + Cα-Tyr), 53.0 (Cα-Leu), 44.0, 43.7 (CH2-Gly), 41.6 (Cβ-Leu), 38.3, 37.6 (Cβ-

Phe + Cβ-Tyr), 25.7 (Cγ-Leu), 23.5, 21.8 (both Leu-CH3)

Mo([Leu]-enkephalin-bpa)(CO)3 (34)

To a stirred solution of Mo(CO)3(NCEt)3 (40 mg; 0.11 mmol) in 10 mL of MeOH was added

dropwise a solution of 33 (100 mg; 0.11 mmol) in 10 mL of MeOH. The mixture was stirred

for an additional period of 10 mins, during which time a yellow-orange precipitate formed.

The mixture was evaporated to dryness in vacuo, affording complex 34 in quantitative yield.

C51H55MoN9O10 = 1050.0 g mol-1.

IR (KBr): ν = 3394 (m), 3311 (m) νNH, 1897 (vs), 1776 (vs), 1746 (s) νCO, 1637 (br, s) νC=O.

MS (ESI-pos.; DMF): m/z = 1052 [M+H]+, 1074 [M+Na]+. 1H NMR (500.35 MHz; DMSO-

d6): δ = 9.13 (1H, s, Tyr-OH), 8.67 (2H, pseudo-d, HPyr), 8.60 (1H, d, 3JHH = 8.2 Hz, NHTyr or

NHPhe), 8.30 (1H, pseudo-t, NHGly), 8.59 (1H, d, 3JHH = 8.1 Hz, NHTyr or NHPhe), 8.05 (1H,

pseudo-t, NHGly), 7.94 (1H, d, 3JHH = 8.3 Hz, NHLeu), 7.89 (2H, d, 3JHH = 8.2 Hz, HAr), 7.64

(2H, d, 3JHH = 8.2 Hz, HAr), 7.57 (2H, m, HPyr), 7.24 (4H, m, HAr-Phe), 7.17-7.09 (8H,

overlapping m, 1HAr-Phe + 2HAr-Tyr + 4HPyr + 1H NH2), 6.94 (1H, br, NH2), 6.64 (2H, d, 3JHH =

8.5 Hz, HAr-Tyr), 4.70 (2H, s, CH2-benzyl), 4.68 (1H, m, Cα.-PheH or Cα.-TyrH), 4.66 (2H, d, 2JHH

= 15.5 Hz, CH2-picolyl), 4.51 (1H, m, Cα.-PheH or Cα.-TyrH), 4.17 (1H, m, Cα.-LeuH), 3.76 (2H,

m, CH2-Gly), 3.74 (2H, d, 2JHH = 15.5 Hz, CH2-picolyl), 3.67 (2H, m, CH2-Gly), 3.05 (2H, m,

Cβ-PheH2 or Cβ-TyrH2), 2.91 (1H, m, Cβ-PheH or Cβ-TyrH), 2.79 (1H, m, Cβ-PheH or Cβ-TyrH), 1.55

(1H, m, Cγ-LeuH), 1.46 (2H, m, Cβ-LeuH2), 0.87 (3H, d, 3JHH = 6.5 Hz, Leu-CH3), 0.82 (3H, d,


185

3JHH = 6.5 Hz, Leu-CH3). 13C NMR (100.6 MHz; DMSO-d6): δ = 231.2, 230.3 (CO), 173.8,

171.9, 170.6, 169.0, 168.6, 165.9 (C=O), 158.8 (2×), 155.6, 150.0, 137.7, 137.5, 135.4, 134.0,

131.3, 130.0, 129.1, 128.4, 128.0, 127.4, 126.2, 123.0, 122.3, 114.9 (CAr), 68.9 (CH2-benzyl),

64.1 (CH2-picolyl), 55.2, 54.0 (Cα-Phe + Cα-Tyr), 51.0 (Cα-Leu), 42.1, 41.9 (CH2-Gly), 40.8 (Cβ-

Leu), 37.3, 36.2 (Cβ-Phe + Cβ-Tyr), 24.1 (Cγ-Leu), 22.9, 21.6 (both Leu-CH3).

General synthesis of the ferrocene derivatives

(Fe(η-Cp)(η-C5H4-C(O)-Phe-OMe) (35), Fe(η-C5H4-C(O)-Phe-OMe)2 (36), Fe(η-Cp)(η-

C5H4-C(O)-PEA) (37), Fe(η-C5H4-C(O)-PEA)2 (38), Fe(η-Cp)(η-C5H4-C(O)-Ala-Phe-OMe)

(39) and Fe(η-C5H4-C(O)-Ala-Phe-OMe)2 (40))

Ferrocene carboxylic acid (230 mg; 1 mmol) or 1,1’-ferrocene dicarboxylic acid (274 mg; 1

mmol) were dissolved in DMF (15 mL) and NEt3 (1.5 mL; 1.09 g; 10.8 mmol) was added. To

this mixture was added stoichiometric amounts (i.e. 1 mmol for the synthesis of 35, 37 and 39

and 2 mmol for the synthesis of 36, 38 and 40) of phenylalanine methyl ester hydrochloride

(for 35 and 36), the dipeptide [H2-Ala-Phe-OMe]CF3CO2 (prepared from the dipetide Boc-

Ala-Phe-OMe as described for the synthesis of 4) or S-1-phenylethylamine (for 37 and 38).

After addition of stoichiomteric amounts of HBTU (1 mmol for 35, 37 and 39; two

equivalents for 36, 38 and 40), the mixture was stirred for 30 minutes at room temperature

and subsequently evaporated to dryness in vacuo. The residue was suspended in CH2Cl2 (150

mL) and filtered to remove insoluble material, followed by washing of the CH2Cl2 solution

with 2M aqueous NaHCO3 (100 mL), 1 M HCl (100 mL) and water (100 mL). The organic

phase was dried over MgSO4, filtered and subsequently evaporated to dryness under reduced

pressure to yield an orange solid. Only in the case of 37 and 38 it was necessary to purify this

solid by column chromatography over silica by using ethyl acetate / hexane (3/1 v/v) as the

eluent. The first coloured band that eluted was the desired compound. The yields were 0.34 g

(87%) for 35, 0.46 g (77%) for 36, 0.21 g (63%) for 37, 0.25 g (52%) for 38, 0.34 g (74%) for

39 and 0.57 g (77%) for 40. Crystals suitable for X-ray analysis of 36 were grown by pentane

diffusion into an ethyl acetate solution at room temperature. X-ray quality crystals of

39⋅0.25CH2Cl2 were grown by slow evaporation of a CH2Cl2 / hexane solution, whereas


186

crystals suitable for X-ray structure determination of 40⋅0.5CHCl3 were obtained by slow

evaporation of a CHCl3 / heptane solution.

35: C21H21NO3Fe = 391.3 g mol-1.

Anal. Calcd.: C 64.47, H 5.41, N 3.58; Found: C 64.41, H 5.53, N 3.64. IR (KBr):ν = 3299

(w, sh), 3266 (m, br) νNH, 1756, 1736 (s) νC=O, ester, 1631, 1620 (s) νC=O, amide; (CH2Cl2): ν =

3426 (w) νNH, 1741 (s) νC=O, ester, 1659 (s) νC=O, amide. MS (EI): m/z = 391 (100) [M]+, 331 (2)

[M-Cp]+, 213 (45). 1H NMR (400.13 MHz; CDCl3; 293 K; 2 × 10-2 M): δ = 7.32 (m, 2H, HAr),

7.26 (m, 1H, HAr), 7.18 (m, 2H, HAr), 6.00 (br, pseudo-s, 1H, NH), 5.01 (m, 1H, CαH), 4.62

(s, 1H, CpsubH), 4.59 (s, 1H, CpsubH), 4.33 (s, 2H, CpsubH), 4.12 (s, 5H, CpunsubH), 3.76 (s, 3H,

OCH3), 3.20 (m, 1H, CβH), 3.14 (m, 1H, CβH). 13C NMR (100.6 MHz; CDCl3): δ = 172.3

(C=Oester), 170.0 (C=Oamide), 136.1 (CAr-Phe), 129.2 (CAr-Phe), 128.7 (CAr-Phe), 127.2 (CAr-Phe),

75.3 (CCp, q), 70.5 (CCp-sub), 69.7 (CCp-unsub), 68.3, 68.0 (CCp-sub). 52.8 (Cα), 52.3 (O-CH3), 38.0

(Cβ).

36: C32FeH32N2O6 = 596.5 g mol-1.


(w), 3219 (w), νNH, 1759 (m), 1740 (s), νC=O, ester, 1630 (s) νC=O, amide; (CH2Cl2): ν = 3380 (w)

νNH, 1729 (s) νC=O, ester, 1652 (s) νC=O, amide. MS (EI): m/z = 596 (100) [M]+, 326 (15) [M –

CpC(O)NPhe-OMe]+. 1H NMR (400.13 MHz; CDCl3; 293 K; 1 × 10-2 M): δ = 7.75 (2H, d,3JHH = 8.4 Hz, NH), 7.29 (6H, m, HAr), 7.21 (m, 4H, HAr), 5.06 (2H, m, CαH) 4.78 (2H,

pseudo-t, CpH), 4.66 (2H, pseudo-t, CpH), 4.49 (2H, pseudo-t, CpH), 4.28 (2H, pseudo-t,

CpH), 3.84 (s, 6H, OCH3), 3.18 (m, 2H, CβH), 2.95 (m, 2H, CβH). 13C NMR (CDCl3; 100.6

MHz): δ = 175.4 (C=Oester), 170.3 (C=Oamide), 136.7 (CAr-Phe), 128.9 (CAr-Phe), 128.6 (CAr-Phe),

127.0 (CAr-Phe), 75.9 (CCp, q) 71.9, 71.3, 70.4, 70.0 (CCp), 54.0 (Cα), 52.8 (O-CH3), 37.0 (Cβ).

37: C19FeH19NO = 333.2 g mol-1.


(br, w) νNH, 1621 (s) νC=O; (CH2Cl2): ν = 3440 (w) νNH, 1654 (s) νC=O. MS (EI): m/z = 333

(100) [M] +. 1H NMR (400.13 MHz; CDCl3; 293 K; 2 × 10-2 M): δ = 7.37 (4H, m, HAr), 7.27

(1H, m, HAr), 5.82 (1H, d, 3JHH = 7.6 Hz, NH), 5.28 (1H, m, N-CH), 4.67 (1H, pseudo-d,

CpsubH), 4.61 (1H, pseudo-d, CpsubH), 4.31 (2H, pseudo-t, CpsubH), 4.11 (5H, s, CpunsubH),


187

1.56 (3H, d, 3JHH = 6.9 Hz, CH3). 13C (62.9 MHz; CDCl3): δ = 169.3 (C=O), 143.6 (Car, q),

128.7, 126.2 (CAr,), 127.4 (CAr), 76.0 (CCp, q), 70.4, 70.4, 68.4, 67.8 (CCp-sub), 69.2 (CCp-unsub)

48.5 (CH), 21.7 (CH3).

38: C28FeH28N2O2 = 480.4 g mol-1.


(w), 3382 (m), νNH, 1637 (s) νC=O; (CH2Cl2): ν = 3435 (w) νNH, 1638 (s) νC=O. MS (EI): m/z =

480 (100) [M] +, 268 (10) [M – CpC(O)-PEA]. 1H NMR (400.13 MHz; CDCl3; 293 K; 1 × 10-

2 M): δ = 7.42 (4H, m, HAr), 7.34 (4H, m, HAr), 7.25 (2H, m, HAr), 6.93 (2H, d, 3JHH = 7.9 Hz,

NH), 5.25 (2H, m, N-CH), 4.45 (2H, pseudo-t, CpH), 4.32 (2H, pseudo-t, CpH), 4.24 (2H,

pseudo-t, CpH), 4.21 (2H, m, CpH), 1.61 (6H, d, 3JHH = 7.0 Hz; both CH3). 13C NMR: δ =

(62.9 MHz; CDCl3): δ = 169.3 (C=O), 143.9 (CAr, q), 128.5, 126.4 (CAr), 127.2 (CAr), 78.2

(CCp, q), 72.3, 71.1, 70.7, 69.2 (CCp), 49.2 (CH), 21.9 (CH3).

39: C24FeH26N2O4 = 462.3 g mol-1.

Anal. Calcd.: C 62.35, N 5.67, H 6.06; Found: C 62.19, H 5.61, N 6.05. IR (KBr): ν = 3301

(br, m) νNH, 1746 (s) νC=O, ester, 1654 (s), 1625 (vs) νC=O, amide; (CH2Cl2): ν = 3424 (m) νNH-Phe

+ νNH-Ala, 1744 (s) νC=O, ester, 1683, 1651 (s), νC=O, amide. MS (EI): m/z = 462 (100) [M]+. 1H

NMR (400.13 MHz; CDCl3; 293 K; 2 × 10-2 M): δ = 7.21 (3H, m, HAr), 7.07 (2H, m, HAr),

6.53 (1H, d, 3JHH = 7.8 Hz, NHAla), 6.18 (1H, d, 3JHH = 7.5 Hz, NHPhe), 4.83 (1H, m, Cα-PheH),

4.68 (1H, s, CpsubH), 4.63 (1H, s, CpsubH), 4.59 (1H, m, Cα-AlaH), 4.35 (2H, s, CpsubH), 4.17

(5H, s, CpH), 3.71 (3H, s, OCH3), 3.13 (1H, m, Cβ-PheH), 3.07 (1H, m, Cβ-PheH), 1.41 (3H, d,3JHH = 7.0 Hz, Ala-CH3). 13C NMR (100.6 MHz; CDCl3): δ = 172.3 (br), 171.5 (C=O, only

two resoannces owing to C=O carbon atoms observed, whereas three are expected), 135.6

(CAr, q), 129.4 (CAr), 128.6 (CAr), 126.9 (CAr), 74.9 (CCp, q), 70.8, 70.7, 69,2, 68.8 (CCp-sub),

70.1 (CCp-unsub), 53.4, 52.5 (OCH3 + Cα--Phe), 49.2 (br, Cα-Ala), 38.1 (Cβ-Phe), 19.6 (br, Ala-

CH3).

40: C38FeH42N4O8 = 738.6 gmol-1.

Anal. Calcd.: 61.79, H 5.73, N 7.59; Found: C 61.57, H 5.78, N 7.65. IR (KBr): ν = 3277 (br,

m) νNH, 1750 (s) νC=O, ester, 1669 (s), 1625 (s) νC=O, amide; (CH2Cl2): ν = 3414 (m) νNH-Phe, 3322

(m) νNH-Ala, 1744 (s) νC=O, ester, 1677, 1644 (s), νC=O, amide. MS (EI): m/z = 738 (100) [M] +. 1H


188

NMR (400.13 MHz; CDCl3; 293 K; 1 × 10-2 M): δ = 8.33 (2H, d, 3JHH = 7.1 Hz, NHAla), 7.32

(4H, m, HAr), 7.27 (2H, m, HAr), 7.20 (4H, m, HAr), 6.31 (2H, d, 3JHH = 7.7 Hz, NHPhe), 4.88

(2H, s, CpH), 4.84 (2H, m, Cα-PheH), 4.81 (2H, s, CpH), 4.61 (2H, m, Cα-AlaH) 4.46 (2H, s,

CpH), 4.29 (2H, s, CpH), 3.66 (6H, s, OCH3), 3.19 (2H, m, Cβ-PheH), 3.16 (2H, m, Cβ-PheH),

1.33 (6H, d, 3JHH = 7.1 Hz, Ala-CH3). 13C NMR (100.6 MHz; CDCl3): δ = 174.5 (C=Oester),

171.6, 170.4 (br, C=O), 135.5 (CAr, q), 129.5 (CAr), 128.6 (CAr), 127.1 (CAr), 75.6 (CCp, q), 72.0,

71.3, 70.5, 70.4 (CCp), 53.9, 52.3 (OCH3 + Cα-Phe), 49.5 (br, Cα-Ala), 37.9 (Cβ-Phe), 17.5 (br,

Ala-CH3).

General synthesis of the monsubstituted cobaltocenium complexes

([Co(Cp)(Cp-C(O)-Phe-OMe)]BPh4 (41) and [Co(Cp)(Cp-C(O)-Ala-Phe-OMe)]BPh4 (42))

To a solution of [Co(Cp)(Cp-COOH)]PF6 (379 mg; 1.0 mmol) in DMF (10 mL) was added

NEt3 (1.0 mL; 0.73g; 7 mmol) and TBTU (323 mg; 1.0 mmol) and either phenylalanine

methyl ester hydrochloride (216 mg; 1.0 mmol; for 41) or [H2-Ala-Phe-OMe]CF3COO (1

mmol; for 43; prepared by Boc-deprotection of Boc-Ala-Phe-OMe via an identical procedure

as described for Boc-Phe-Leu-OMe for the synthesis of 4). The mixture was stirrred for 30

mins and subsequently evaporated to dryness in vacuo. The residue was suspended in CH2Cl2

(100 mL) and filtered to remove insoluble material. The resulting clear yellow solution was

washed with 2 M NaHCO3 (50 mL), 1 M HCl (50 mL) and water (50 mL). The CH2Cl2 was

dried over MgSO4 and subsequently removed under reduced pressure, yielding a sticky

yellow residue. This was dissolved in MeOH (10 mL), followed by addition of a NaBPh4

(0.25 g; 0.73 mmol) solution in MeOH (10 mL). Upon standing overnight, a yellow

precipitate formed, which was isolated by filtration and washed with MeOH (5 mL) and air

dried. Yields were 0.26 g (36%) for 41 and 0.10 g (13%) for 43. X-ray quality crystals of 41

were grown by slow evaporation of a MeOH solution.

41: BC45CoH41NO3 = 713.6 g mol-1.


(w) νNH, 1735 (s) νC=O, ester, 1674 (s) νC=O, amide; (CH2Cl2): 3403 (w) νNH, 1744 (s) νC=O, ester,

1654 (s) νC=O, amide. MS (ESI-pos.; CH3OH): m/z = 394 [41-BPh4]+. 1H NMR (400.13 MHz;


189

CDCl3; 293 K; 2 × 10-2 M): δ = 7.48 (10H, m, HAr), 7.30 (2H, m, HAr), 7.15 (2H, m, HAr),

7.05 (10H, m, HAr), 6.90 (5H, m, HAr), 6.11 (1H, d, 3JHH = 7.9 Hz, NH), 4.99 (2H, pseudo-s,

CpsubH), 4.85 (2H, pseudo-s, CpsubH), 4.80 (5H, s, CpunsubH), 4.54 (2H, pseudo-s, CpsubH),

4.50 (2H, pseudo-s, CpsubH), 3.77 (3H, s, OCH3), 3.28 (1H, m, CβH), 3.06 (1H, m, CβH). 13C

NMR (125.8 MHz; CD3OD): δ = 173.1 (C=Oester), 165.3 (q, 2JBC = 49 Hz, Cq-BPh4), 163.8

(C=Oamide), 138.5 (CAr-Phe, q), 137.3 (CAr-BPh4), 130.1 (CAr-Phe), 129.8 (CAr-Phe), 128.2 (CAr-Phe),

126.4 (CAr-BPh4), 122.7 (CAr-BPh4), 94.1 (CCp, q), 87.4 (CCp-sub), 87.3 (CCp-unsub), 85.3 (CCp-

sub), 85.0 (CCp-sub), 55.5 (Cα), 53.1 (O-CH3), 37.7 (Cβ).

43: BC48H46CoN2O4 = 784.7 g mol-1.


(m), 3381 (w) νNH, 1724 (s), νC=O, ester, 1679 (vs) νC=O, amide. MS (ESI-pos; MeOH): m/z = 465

[43 -BPh4]+. 1H NMR (400.13 MHz; DMSO-d6): δ = 8.69 (1H, d, 3JHH = 6.9 Hz, NH), 8.53

(1H, d, 3JHH = 7.1 Hz, NH), 7.26 (13H, m, HAr), 6.93 (8H, m, HAr), 6.79 (4H, m, HAr), 6.40

(1H, s, CpHsub), 6.24 (1H, s, CpHsub), 5.86 (2H, pseudo-d, CpHsub), 5.79 (5H, s, CpHunsub),

3.58 (3H, s, OCH3), Cα-AlaH obscured by the H2O signal, 3.02 (2H, m, Cβ-PheH2), 1.35 (3H, d,3JHH = 7.0 Hz, Ala-CH3). 13C NMR (100.6 MHz; DMSO-d6): δ = 172.21, 171.9 (C=O), 163.3

(q, 2JBC = 49 Hz, Cq-BPh4), 137.1 (Cq-Phe), 135.5 (CAr-BPh4), 129.1, 128.3, 126.6 (CAr-Phe),

125.3 (m, CAr-BPh4), 121.5 (CAr-BPh4), 92.6 (CCp, q), 86.0, 85.7, 84.5, 83.6 (CCp,-sub), 86.0

(CCp-unsub), 53.9, 51.4 (OCH3 + Cα-Phe), 36.5 (Cβ-Phe), 17.5 (Ala-CH3).

General synthesis of the disubstituted cobaltocenium complexes

([Co(Cp-Phe-OMe)2]PF6 (42) and ([Co(Cp-Ala-Phe-OMe)2]PF6 (44))

To a solution of [Co(Cp-COOH)2]PF6 (217 mg; 0.5 mmol) in DMF (6 mL) was added NEt3

(1.0 mL; 0.73g; 7 mmol) and HBTU (182 mg; 0.5 mmol) and either phenylalanine methyl

ester hydrochloride (216 mg; 1.0 mmol; for 42) or [H2-Ala-Phe-OMe]CF3COO (1 mmol; for

44; prepared by Boc-deprotection of Boc-Ala-Phe-OMe via an identical procedure as

described for Boc-Phe-Leu-OMe for the synthesis of 4). The mixture was stirrred for 30

minutes and subsequently evaporated to dryness in vacuo. The residue was suspended in


190

CH2Cl2 (100 mL) and filtered to remove insoluble material. The resulting clear yellow

solution was washed with 2 M NaHCO3 (50 mL), 1 M HCl (50 mL) and water (50 mL). The

CH2Cl2 was dried over MgSO4 and subsequently removed under reduced pressure. The

resulting sticky residue was dried for several hours in vacuo, yielding a hygroscopic foam-like

solid. This solid was purified by preparative HPLC, employing a 3/1 mixture of MeCN and

H2O (the latter containing 0.05 M NaPF6) as the eluent. The work-up consisted of

concentration of the eluent to about a quarter of its volume under reduced pressure. The

remaining solution was extracted with CH2Cl (3 × 50 mL) and the CH2Cl2 was dried over

MgSO4. Removal of the CH2Cl2 in vacuo afforded the desired compounds in highly pure

form. Yield: 110 mg (32%) for 42 and 50 mg (11%) for 44.

42: C32CoF6H32N2OP = 685.6 g mol-1.


(m, br), 1742 (s) νC=O, ester, 1670 (s) νC=O, amide; (CH2Cl2): ν = 3403 (w), 3362 (w) νNH, 1743

(s), 1721 (s) νC=O, ester. MS (ESI-pos.; MeOH): m/z = 599 [42 – PF6] +. 1H NMR (400.13 MHz;

CDCl3; 293 K; 1 × 10-2 M): δ = 7.79 (2H, d, 3JHH = 8.3 Hz, NH), 7.33 (6H, m, HAr), 7.25 (2H,

m, HAr), 6.12 (2H, pseudo-t, CpH), 5.82 (6H, m, CpH), 3.83 (6H, s, OCH3), 3.35 (2H, m,

CβH), 3.05 (2H, m, CβH). 13C NMR (CDCl3; 100.6 MHz): δ = 171.8 (C=Oester), 162.2

(C=Oamide), 137.2 (CAr, q), 129.2 (CAr), 128.5 (CAr), 127.4 (CAr), 92.9 (CCp, q) 88.2, 87.3, 86.6,

84.9 (CCp), 54.9 (Cα), 52.6 (OCH3), 36.3 (Cβ).

44: C38CoF6H42N4O8P = 886.7 gmol-1.

IR (CH2Cl2): ν = 3405 (w) νNH-Phe, 3296 (w) νNH-Ala, 1744 (s) νC=O, ester, 1660 (s) νC=O, amide.

MS (ESI-pos.; MeOH): m/z = 741 [44-PF6]+; exact mass: 741.2332; C38CoH42N4O8 requires

741.2335. 1H NMR (400.13 MHz; CDCl3; 5 × 10-3 M; 293 K): δ = 8.79 (2H, d, 3JHH = 6.4 Hz,

NHAla), 7.32 (4H, m, HAr), 7.28 (2H, m, HAr), 7.22 (4H, m, HAr), 6.57 (2H, br, NH), 6.32 (2H,

br, CpH), 6.26 (2H, br, CpH), 6.02 (2H, br, CpH), 5.87 (2H, br, CpH), 4.83 (2H, m, Cα-PheH),

4.57 (2H, m, Cα-AlaH), 3.70 (s, 6H, OCH3), 3.16 (4H, m, Cβ-PheH2), 1.36 (6H, d, 3JHH = 7.4 Hz,

Ala-CH3). 13C NMR (62.9 MHz; CDCl3): δ = 173.7 (C=Oester), 171.6 (C=OAla), 161.2

(C=OCp), 135.9 (CAr, q), 129.3 (CAr), 128.6 (CAr), 127.1 (CAr), 92.1 (CCp, q), 87.9, 87.5, 86.0,

85.4 (CCp), 54.0, 52.4 (OCH3 + Cα-Phe), 49.9 (Cα-Ala), 37.6 (Cβ-Phe), 16.9 (Ala-CH3).


191

Synthesis of [Co(Cp)(Cp-CO-S-PEA)2] (45)

This compound was prepared analogously as described for the other mono-substituted

cobaltocenium compounds 41 and 43. After addition of a methanolic NaBPh4 solution, about

10 mg of X-ray quality crystals separated the first time the synthesis was performed. The X-

ray crystal structure confirms the identity of this compound. However, it could not be

reproduced because when tried again, the precipitate that formed upon addition of NaBPh4

was found to be contaminated with S-1-phenylethylammonium salts. Probably this compound

can be purified by preparative HPLC but this was not pursued.

BC43CoH39NO = 655.5 g mol-1.


(s) νNH, 1667 (s), νC=O. MS (ESI-pos., MeOH): m/z = 336 [45 – BPh4]+.

192

X-ray crystallographic data

Compound 3a 5⋅MeOH

Empirical formula C21H21MoN3O5 C12H17MoN3O5

Formula weight 463.33 379.23

Crystal size [mm] 0.80 × 0.04 × 0.02 0.35 × 0.32 × 0.32

Crystal system Monoclinic Trigonal

Space group C2 P3121

a [Å] 27.088(4) 11.7007(6)

b [Å] 9.909(1) 11.7007(6)

c [Å] 14.865(2) 18.9244(11)

α [°] 90 90

β [°] 102.82(2) 90

γ [°] 90 120

V [Å3] 3890.5(9) 2243.8(2)

Z 8 6

ρ (calcd.) [g/cm3] 1.582 1.684

Absorption coefficient [mm-1] 0.707 0.901

F(000) 1888 1152

T [K] 100(2) 100(2)

θ Range for data collection [°] 2.89 to 25.00 2.01 to 27.40

Reflections collected 14173 24508

Independent refections 6081 [R(int) = 0.1297] 3402 [R(int) = 0.0639]

Observed reflections [I>2σ(I)] 4451 3268

Data / restraints / parameters 6072 / 7 / 507 3402 / 6 / 238

Goodness-of-fit on F2 0.999 1.049

Final R indices (obs. data)[a] R1 = 0.0606

wR2 = 0.1413

R1 = 0.0223

wR2 = 0.0522

R indices (all data) R1 = 0.0874

wR2 = 0.1592

R1 = 0.0243

wR2 = 0.0527

Flack parameter -0.03(5) -0.05(3)

Largest diff. peak and hole [e/Å3] 0.928 and –0.817 0.496 and –0.409


193

Compound 6 7⋅2MeOH



Crystal size [mm] 0.32 × 0.14 × 0.04 0.70 × 0.42 × 0.35

Crystal system Monoclinic Monoclinic

Space group P21 P21

a [Å] 8.5602(12) 12.0928(14)

b [Å] 12.704(2) 7.5347(8)

c [Å] 12.667(2) 12.1177(14)

α [°] 90 90

β [°] 92.05(2) 101.59(2)

γ [°] 90 90

V [Å3] 1376.6(4) 1081.6(2)

Z 4 2

ρ (calcd.) [g/cm3] 1.743 1.527


F(000) 728 512

T [K] 100(2) 100(2)








wR2 = 0.1295

R1 = 0.0269

wR2 = 0.0685


wR2 = 0.1349

R1 = 0.0281

wR2 = 0.0692

Flack parameter 0.01(5) -0.03(3)

Largest diff. peak and hole [e/Å3] 1.046 and – 0.926 1.066 and –0.721


194

Compound 8 9⋅H2O



Crystal size [mm] 0.52 × 0.35 × 0.28 0.40 × 0.28 × 0.12

Crystal system Orthorhombic Monoclinic

Space group P212121 P21

a [Å] 8.5384(4) 7.7734(6)

b [Å] 11.7713(8) 13.8012(10)

c [Å] 17.9943(12) 14.7311(11)

α [°] 90 90

β [°] 90 94.43(2)

γ [°] 90 90

V [Å3] 1808.6(2) 1575.7(2)

Z 4 2

ρ (calcd.) [g/cm3] 1.643 1.550


F(000) 912 744

T [K] 100(2) 100(2)








wR2 = 0.0511

R1 = 0.0346

wR2 = 0.0796


wR2 = 0.0519

R1 = 0.0441

wR2 = 0.0835

Flack parameter -0.01(2) 0.025(10)



195

Compound 10⋅MeOH 14



Crystal size [mm] 0.34 × 0.30 × 0.28 0.37 × 0.35 × 0.05

Crystal system Trigonal Orthorhombic


a [Å] 11.770(2) 9.7478(3)

b [Å] 11.770(2) 10.8387(3)

c [Å] 18.977(3) 23.1929(6)

α [°] 90 90

β [°] 90 90

γ [°] 120 90

V [Å3] 2276.7(7) 2450.41(12)

Z 6 4

ρ (calcd.) [g/cm3] 1.660 1.571


F(000) 1152 1188

T [K] 100(2) 100(2)








wR2 = 0.0893

R1 = 0.0261

wR2 = 0.0669


wR2 = 0.1088

R1 = 0.0274

wR2 = 0.0679




196

Compound 22 23



Crystal size [mm] 0.27 × 0.18 × 0.10 0.22 × 0.06 × 0.03

Crystal system Orthorhombic Monoclinic

Space group Pna21 Cc

a [Å] 15.347(3) 7.1786(3)

b [Å] 8.449(2) 32.62225)

c [Å] 22.454(4) 17.2691(8)

α [°] 90 90

β [°] 90 96.09(2)

γ [°] 90 90

V [Å3] 2911.5(10) 4021.3(4)

Z 8 8

ρ (calcd.) [g/cm3] 1.730 1.550


F(000) 1520 1904

T [K] 100(2) 100(2)








wR2 = 0.1269

R1 = 0.0389

wR2 = 0.0913


wR2 = 0.2085

R1 = 0.0522

wR2 = 0.1213



197

Compound 25 36

Empirical formula C24H26BrF6Mo2N6O4P C32H32FeN2O6


Crystal size [mm] 0.50 × 0.30 × 0.06 0.53 × 0.18 × 0.18

Crystal system Monoclinic Hexagonal

Space group P21/n P65

a [Å] 13.0262(9) 22.332(2)

b [Å] 9.4274(7) 22.332(2)

c [Å] 24.258(2) 10.3690(7)

α [°] 90 90

β [°] 99.15(2) 90

γ [°] 90 120

V [Å3] 2941.1(4) 4478.4(5)

Z 4 6

ρ (calcd.) [g/cm3] 1.986 1.327


F(000) 1728 1872

T [K] 100(2) 100(2)





Data / restraints / parameters 10168 / 0 / 397 4974 /1 / 372



wR2 = 0.0732

R1 = 0.0532

wR2 = 0.1337


wR2 = 0.0781

R1 = 0.0742

wR2 = 0.1567

Flack parameter -0.01(2)

Largest diff. peak and hole [e/Å3] 1.326 and –0.843 -1.189 and –0.309


198

Compound 39⋅0.25CH2Cl2 40⋅0.5CHCl3

Empirical formula C24H26FeN2O4⋅0.25CH2Cl2 C38H42FeN4O8⋅0.5CHCl3Formula weight 483.55 798.29

Crystal size [mm] 1.28 × 0.70 × 0.60 0.60 × 0.18 × 0.12

Crystal system Tetragonal Monoclinic


a [Å] 17.2236(14) 12.7711(8)

b [Å] 17.2236(14) 16.4209(10)

c [Å] 16.9834(12) 18.269(12)

α [°] 90 90

β [°] 90 93.44(1)

γ [°] 90 90

V [Å3] 5038.2(7) 3824(3)

Z 8 4

ρ (calcd.) [g/cm3] 1.275 1.386


F(000) 2020 1668

T [K] 100(2) 100(2)








wR2 = 0.1965

R1 = 0.0516

wR2 = 0.1266


wR2 = 0.2124

R1 = 0.0679

wR2 = 0.1551




199

Compound 41 45

Empirical formula C45H41BCoNO3 C43H39BCoNO


Crystal size [mm] 0.36 × 0.22 × 0.10 0.50 × 0.38 × 0.08

Crystal system Orthorhombic Triclinic


a [Å] 13.992(2) 9.493(1)

b [Å] 39.148(7) 9.925(1)

c [Å] 13.195(2) 9.947(1)

α [°] 90 97.90(3)

β [°] 90 92.80(3)

γ [°] 90 115.51(3)

V [Å3] 7228(2) 831.7(2)

Z 8 1

ρ (calcd.) [g/cm3] 1.311 1.309


F(000) 2992 344

T [K] 100(2) 100(2)








wR2 = 0.1206

R1 = 0.0434

wR2 = 0.0895


wR2 = 0.1480

R1 = 0.0560

wR2 = 0.0942

Flack parameter 0.02(2) 0.022(11)


200

References

[1] F. A. Carey, R. J. Sundberg, Advanced Organic Chemistry, part B: Reactions and

Synthesis, 3rd ed., Plenum Press, New York, 1990.

[2] C. H. Wong, G. M. Whitesides; Enzymes in Synthetic Organic Chemistry, Pergamon

Press, 1994.

[3] R. B. Merrifield, J. Am. Chem. Soc. 1963, 85, 2149.

[4] M. Bodansky, A. Bodansky, The Practice of Peptide Synthesis, Springer-Verlag, Berlin,

1984.

[5] M. Bodansky, Principles of Peptide Synthesis, Springer-Verlag, Berlin, 1984.

[6] J. Jones, The Chemical Synthesis of Peptides, Clarendon Press, Oxford, 1991.

[7] D. Voet, J. G. Voet, Biochemistry, John Wiley & Sons, New York, 1990.

[8] L. Stryer; Biochemie, 4. Auflage, Spektrum Verlag, Heidelberg, 1990.

[9] R. W. Hay, Bio-Inorganic Chemistry, Ellis Horwood, Ltd., Chichester, 1984.

[10] I. Bertini, H. B. Gray, S. J. Lippard, J. S. Valentine, Bioinorganic Chemistry, University

Science Books, Mill Valley, CA, 1994.

[11] S. J. Lippard, J. M. Berg, Principles of Bioinorganic Chemistry, University Science

Books, Mill Valley, CA, 1994.

[12] A. Klug, D. Rhodes, Trends Biochem. Sci. 1988, 12, 465.

[13] J. Miller, A. D. McLachlan, A. Klug, EMBO J. 1985, 4, 1609.

[14] A. Liljas, K. K. Kannan, P.-C. Bergsten, I. Waara, K. Fridborg, B. Strandberg, U.

Carlbom, L. Järup, S. Lövgren, M. Petef, Nature 1972, 235, 131.

[15] K. K. Kannan, B. Notstrand, K. Fridborg, S. Lövgren, A. Ohlsson, M. Petef, Proc. Natl.

Acad. Sci. USA 1975, 72, 51.

[16] A. Volbeda, M.-H. Charon, C. Piras, E. C. Hatchikian, M. Frey, J. C. Fontecilla-Camps,

Nature 1995, 373, 580.

[17] A. Volbeda, E. Garcin, C. Piras, A. L. de Lacey, V. M. Fernandez, E. C. Hatchikian, M.

Frey, J. C. Fontecilla-Camps, J. Am. Chem. Soc. 1996, 118, 12989.

[18] R. P. Happe, W. Roseboom, A. J. Pierik, S. P. J. Albracht, K. A. Bagley, Nature 1997,

385, 126.

References

201

[19] K. A. Bagley, E. C. Duin, W. Roseboom, S. P. J. Albracht, W. H. Woodruff,

Biochemistry 1995, 373, 580.

[20] A. L. de Lacey, E. C. Hatchikan, A. Volbeda, M. Frey, J. C. Fontecilla-Camps, V. M.

Fernandez, J. Am. Chem. Soc. 1997, 119, 7181.

[21] L. M. Roberts, P. A. Lindahl, J. Am. Chem. Soc. 1995, 117, 2565.

[22] J. R. Lancaster, Science 1982, 216, 1324.

[23] J. J. G. Moura, I. Moura, B. H. Huyhn, H. J. Krüger, M. Teixeira, R. C. DuVarney, D. V.

DerVartanian, A. V. Xavier, H. D. Peck Jr., J. LeGall, Biochem. Biophys. Res. Commun.

1982, 108, 1388.

[24] S. P. J. Albracht, E.-G. Graf, R. K. Thauer, FEBS Lett. 1982, 140, 311.

[25] M. Y. Darensbourg, E. J. Lyon, J. J. Smee; Coord. Chem. Rev. 2000, 206-207, 533

[26] T. J. Kealy, P. L. Pauson, Nature 1951, 168, 1039.

[27] D. C. Hodgkin, J. Kamper, M. Mackay, J. Pickworth, K. N. Trueblood, J. G. White,

Nature 1956, 178, 64.

[28] K. Severin, R. Bergs, W. Beck; Angew. Chem. 1998, 110, 1722.

[29] W. S. Sheldrick, A. Gleichmann, J. Organomet. Chem. 1994, 470, 183.

[30] L. Y. Kuo, M. G. Kanatzidis, M. Sabat, A. L. Tipton, T. J. Marks, J. Am. Chem. Soc.

1991, 113, 9207.

[31] E. Cuignet, C. Sergheraert, A. Tartar, J. Organomet. Chem. 1980, 195, 325

[32] R. Epton, G. Marr, G. A. Willmore, D. Hudson, P. H. Snell, C. R. Snell, Int. J. Biol.

Macromol. 1981, 3, 395.

[33] E. Cuignet, M. Dautrevaux, C. Sergheraert, A. Tartar, B. Attali, J. Cros, Eur. J. Med.

Chim. Ther. 1982, 17, 203.

[34] J. Pospisek, S. Toma, I. Fric, K. Blaha, Coll. Czech. Chem. Commun. 1980, 45, 435.

[35] J. M. Osgerby, P. L. Pauson, J. Chem. Soc. 1958, 656.

[36] H. Brunner, W. König, B. Nuber, Tetrahedron Asymm. 1993, 4, 699.

[37] J. C. Brunet, E. Cuignet, H. Gras, P. Marcinal, A. Mocz, C. Sergheraert, A. Tartar, J.

Organomet. Chem. 1981, 216, 73.

[38] A. Ricouart, P. Maes, T. Battmann, B. Kerdelhue, A. Tartat, C. Segheraert, Int. J. Peptide

Protein Res. 1988, 32, 56.

[39] P. Maes, A. Ricouart, E. Escher, A. Tartar, C. Sergheraert, Coll. Czech. Chem. Commun.

1988, 53, 2914.

[40] H. Köpf and P. Köpf-Maier, Angew. Chem. Int. Ed. Engl. 1979, 18, 477.

References

202

[41] H. Köpf and P. Köpf-Maier, Z. Naturforsch. 1979, 34b, 805.

[42] P. Köpf-Maier, T. Klapötke, J. Cancer Res. Clin. Oncol. 1992, 118a, 216.

[43] B. K. Keppler, Metal Complexes in Cancer Chemotherapy, VCH Weinheim 1993.

[44] S. P. Fricker (Ed.), Metal Compounds in Cancer Therapy, Chapman & Hall, New York,

1994.

[45] H. El Amouri, Y. Besace, J. Vaissermann, G. Jaouen, J. Organomet. Chem. 1996, 515,

103.

[46] O. Brosch, Ph. D. Thesis, Ruhr-Universität Bochum, 1999.

[47] A. Voller, A. Barlett, D. Bidwell (Ed.), Immunoassays for the 80s, MTP Press, Ltd.,

Lancaster, 1981.

[48] W. M. Hunter, J. E. T. Corrie (Ed.), Immonuassays for Clinical Chemistry, Churchill

Livingstone, Edinburgh, 1983.

[49] W. R. Butt (Ed.), Practical Immunoassays. The State of the Art, Marcel Dekker, Inc.,

New York, 1984.

[50] S. B. Pal (Ed.), Immunoassay Technology, vol. 2, Walter de Gruyter, Berlin, 1986.

[51] D.W. Chan, M. T. Perlstein (Ed.), Immunoassay. A Practical Guide, Academic Press,

New York, 1987.

[52] J. P. Gosling, Clin. Chem. 1990, 36, 1422.

[53] A. E. Bolton, M. Hunter, Biochem. J. 1973, 133, 529.

[54] B. M. Jaffe, H. R. Behrmann (Ed.), Methods of Hormone Radioimmunoassay, 2nd ed.,

Academic Press, New York, 1979.

[55] J. L. Vaitukaitis, Clin. Chem. 1985, 31, 1749.

[56] S. Schwarz, P. Berger, G. Wick, Clin. Chem. 1985, 31, 1322.

[57] M. Cais, S. Dani, Y. Eden, O. Gandolfi, M. Horn, E. E. Isaacs, Y. Josephy, Y. Saar, E.

Slovin, L. Snarsky, Nature 1977, 270, 535.

[58] M. Cais, E. Slovin, L. snarsky, J. Organomet. Chem. 1978, 160, 223.

[59] P. Cheret, P. Brosnier, Res. Comm. Chem. Pathol. Pharmacol. 1986, 54, 237.

[60] I. Hammilä, Clin. Chem. 1985, 31, 359.

[61] E. P. Diamandis, Clin. Biochem. 1988, 21, 139.

[62] E. P. Diamandis, V. Bhayana, K. Conway, Clin. Biochem. 1988, 21, 291.

[63] R. Bador, H. Dechaud, F. Claustrat, C. Desuzinges, Clin. Chem. 1987, 33, 48.

[64] G. Barnard, F. Kohen, H. Mikola, T. Lövgren, Clin. Chem. 1989, 35, 555.

[65] I. Hammilä, S. Holttinen, K. Petterson, T. Lövgren, Clin. Chem. 1987, 33, 2281.

References

203

[66] A. D. Ryabov, Angew. Chem. 1991, 103, 945.

[67] S. Weber, W. Purdy, Anal. Lett. 1979, 12, 1.

[68] P. Tijssen, Practice and Theory of Enzyme Immunoassays, Elsevier, Amsterdam, 1985.

[69] G. A. Robinson, G. Martinazzo, G. C. Forrest, J. Immunoassay 1986, 7, 1.

[70] T. Yao, G. A. Rechnitz, Biosensors 1987, 3, 307.

[71] K. D. Gleria, H. A. O. Hill, C. J. McNeil, M. J. Green, Anal. Chem. 1986, 58, 1203.

[72] G. A. Robinson, G. Martinazzo, G. C. Forrest, Biosensors 1986, 2, 45.

[73] G. A. Robinson, V. M. Cole, G. C. Forrest, Biosensors 1987, 3, 147.

[74] H. Eckert, M. Koller, J. Liq. Chromatogr. 1990, 17, 3399.

[75] H. Eckert, M. Koller, Z. Naturforsch. 1990, 45b, 1709.

[76] M. Koller, German appl. Patent Nr. 39 19 317, 28 April 1989.

[77] C. J. Neill, I. J. Higgins, J. V. Bannister, Biosensors 1987, 3, 199.

[78] G. Jaouen, A. Vessières, S. Top, M. Savignac, A. A. Ismail, I. S. Butler, Organometallics

1987, 6, 1985.

[79] A. Vessières, S. Top, A. A. Ismail, I. S. Butler, M. Louer, G. Jaouen, Biochemistry 1988,

27, 6659.

[80] M. Gruselle, P. Deprez, A. Vessières, S. Greenfield, G. Jaouen, J. P. Larue, D.

Thouvenot, J. Organomet. Chem. 1989, 359, C53.

[81] I. S. Butler, A. Vessières, M. Salmain, P. Brossier, G. Jaouen, Appl. Spectrosc. 1989,

1497.

[82] G. Jaouen, A. Vessières, Pure Appl. Chem. 1989, 61, 565.

[83] A. Vessières, C. Vaillant, M. Salmain, G. Jaouen, J. Steroid Biochem. 1989, 34, 301.

[84] A. Vessières, C. Vaillant, M. Gruselle, D. Vichard, G. Jaouen, J. Chem. Soc., Chem.

Commun. 1990, 837.

[85] A. M. Bond, E. Mocellin, C. B. Pascual, P. Wedkanja, G. Jaouen, Appl. Organomet.

Chem. 1990, 4, 557.

[86] M. Salmain, A. Vessières, G. Jaouen, I. S. Butler, Anal. Chem. 1991, 63, 2323.

[87] A. Vessières, M. Salmain, V. Philomin, G. Jaouen, Immunonanal. Biol. Spec. 1992, 31,

9.

[88] M. Salmain, A. Vessières, P. Brossier, I. S. Butler, G. Jaouen, J. Immunol. Methods

1992, 148, 65.

[89] H. El Amouri, A. Vessières, D. Vichard, S. Top, M. Gruselle, G. Jaouen, J. Med. Chem.

1992, 35, 3130.

References

204

[90] A. Varenne, M. Salmain, C. Brisson, G. Jaouen, Bioconjugate Chem. 1992, 3, 471.

[91] A. Vessières, S. Top, C. Vaillant, D. Osella, J. P. Mornon, G. Jaouen, Angew. Chem., Int.

Ed. Engl. 1992, 31, 753.

[92] G. Jaouen, A. Vessières, Acc. Chem. Res. 1993, 26, 361.

[93] V. Philomin, A. Vessières, M. Gruselle, G. Jaouen, Bioconjugate Chem. 1993, 4, 419.

[94] M. Salmain, M. Gunn, A. Gorfti, S. Top, G. Jaouen, Bioconjugate Chem. 1993, 4, 425.

[95] D. Osella, G. Dutto, G. Jaouen, A. Vessières, P. R. Raithby, L. De Benedetto, M. J.

McGlinchey, Organometallics 1993, 12, 4545.

[96] M. Salmain, K. L. Malisza, S. Top, G. Jaouen, M.-C. Sénéchal-Tocquer, D. Sénéchal, B.

Caro, Bioconjugate Chem. 1994, 4, 655.

[97] A. Varenne, A. Vessières, P. Brossier, G. Jaouen, Res. Commun. Chem. Pathol.

Pharmacol. 1994, 84, 81.

[98] A. Varenne, A. Vessières, M. Salmain, P. Brossier, G. Jaouen, J. Immunol. Methods

1995, 186, 195.

[99] S. Top, H. El. Hafa, A. Vessières, J. Quivy, J. Vaissermann, D. W. Hughes, M. J.

McGlinchey, J.-P.Mornon, E. Thoreau, G. Jaouen, J. Am. Chem. Soc. 1995, 117, 8372.

[100] A. Piórko, R. G. Sutherland, A. Vessières-Jaouen, G. Jaouen, J. Organomet. Chem.

1996, 512, 79.

[101] A. Kazimierczak, J. Zakrzewski, M. Salmain, G. Jaouen, Bioconjugate Chem. 1997, 8,

489.

[102] A. Varenne, A. Vessières, M. Salmain, G. Jaouen, Appl. Spectrosc. 1998, 52, 1383.

[103] A. Vessières, M. Salmain, P. Brossier, G. Jaouen, J. Pharm. Biomed. Anal. 1999, 21,

625.

[104] F. A. Cotton, J. Kraihanzel, J. Am. Chem. Soc. 1962, 84, 4432.

[105] A. Varenne, A. Vessières, M. Salmain, S. Durand, P. Brossier, G. Jaouen, Anal.

Biochem. 1996, 242, 172.

[106] M. Salmain, A. Vessières, A. Varenne, P. Brossier, G. Jaouen, J. Organomet. Chem.

1999, 589, 92.

[107] M. Wenzel, C. Klinge, J. Labelled Compd. Radiopharm. 1994, 34, 981.

[108] F. Lipmann; Acc. Chem. Res. 1973, 6, 366.

[109] H.-D. Jakubke; Peptide-Chemie und Biologie, 1st ed., Spektrum Verlag, Heidelberg,

1996.

References

205

[110] J. C. Sheehan, G. P. Hess, J. Am. Chem. Soc. 1955, 77, 1067.

[111] Y. S. Klausner, M. Bodansky, Synthesis 1972, 453.

[112] J. Meienhofer; The mixed carbonic anhydride method of peptide synthesis, Academic

Press, London, 1979.

[113] R. B. Merrifield, J. Am. Chem. Soc. 1964, 86, 304.

[114] W. C. Chan, P. D. White (Eds.), Fmoc Solid Phase Peptide Synthesis, Oxford

University Press, New York, 2000.

[115] Novabiochem Catalog and Peptide Synthesis Handbook, Callbiochem-Novabiochem,

1999

[116] J. Hughes, T. W. Smith, H. W. Kosterlitz, L. A. Fothergill, B. A. Morgan, H. R. Morris,

Nature 1975, 258, 577.

[117] R. Simantov, S. H. Snyder, Life Sci. 1976, 18, 781.

[118] S. Udenfried, J. Meienhofer (Eds.), The Peptides: Analysis, Synthesis, Biology,

Academic Press, New York, Vol. 6, pp. 324-389, 1984.

[119] G. D. Smith, J. F. Griffin; Science 1978, 199, 1214.

[120] A. Aubry, N. Birlirakis, M. Sakarellos-Daitsiotis, C. Sakarellos, M. Marraud, J. Chem.

Soc., Chem. Commun. 1998, 963.

[121] A. Aubry, N. Birlirakis, M. Sakarellos-Daitsiotis, C. Sakarellos, M. Marraud,

Biopolymers 1989, 28, 27.

[122] L. Gylbert, Acta Cryst., Sect. B 1973, 29, 1630.

[123] A. Camerman, D. Mastrapaolo, I. L. Karle, J. Karle, N. Camerman; Nature 1983, 306,

447.

[124] I. L. Karle, D. Mastrapaolo, A. Camerman, N. Camerman, Acta Cryst., Sect. B 1983,

39, 625.

[125] J. F. Griffin, D. A. Langs, G. D. Smith, T. L. Blundell, I. J. Tickle, S. Bedarkar, Proc.

Natl. Acad. Sci. USA 1986, 83, 3272.

[126] B. P. Roques, C. Garbay-Jaureguiberry, R. Oberlin, R. Anteunis, A. K. Lala; Nature

1976, 262, 778.

[127] C. R. Jones, W. A. Gibbons, V. Garsky; Nature 1976, 262, 779.

[128] H. E. Bleich, A. R. Day, R. J. Freer, J. A. Glasel, Biochem. Biophys. Res.

Commun.1977, 74, 592.

[129] J. L. De Coen, C. Humblett, M. H. J. Koch; FEBS Lett. 1977, 73, 38.

References

206

[130] M. A. Khaled, M. M. Long, W. D. Thompson, R. J. Bradley, G. B. Brown, D. W. Urry;

Biochem. Biophys. Res. Commun. 1977, 76, 224.

[131] T. Ishida, M. Kenmotsu, Y. Mino, M. Inoue, T. Fujiwara, K.-I. Tomita, T. Kimura, S.

Sakakibara; Biochem. J. 1984, 218, 677.

[132] J. Soos, I. Berzetei, S. Bajusz, A. Z. Ronai, Life Sci. 1980, 27, 129.

[133] M. Cousins, M. L. H. Green; J. Chem. Soc. 1963, 889.

[134] R. B. King; Inorg. Chem. 1966, 5, 2242.

[135] A. Davison, W. C. Rode; Inorg. Chem. 1967, 6, 2124.

[136] J. W. Faller, M. Incorvia; Inorg. Chem. 1968, 7, 840.

[137] J. W. Faller, C.-C. Chen, M. J. Mattina, A. Jakubowski; J. Organomet. Chem. 1973, 52,

361.

[138] R. G. Hayter, J. Organomet. Chem. 1968, 13, P1.

[139] H. T. Dieck, H. Friedel; J. Organomet. Chem. 1968, 14, 375.

[140] W. P. Hart, D. W. Macomber, M. D. Rausch; J. Am. Chem. Soc. 1980, 102, 1196.

[141] D. W. Macomber, W. P. Hart, M. D. Rausch; Adv. Organomet. Chem. 1982, 21, 1.

[142] N. J. Coville, K. E. du Plooy, W. Pickl; Coord. Chem. Rev. 1992, 116, 1.

[143] B. El Mouatassim, H. El Amouri, M. Salmain, G. Jaouen, J. Organomet. Chem. 1994,

479, C18.

[144] B. El Mouatassim, H. El Amouri, J. Vaissermann, G. Jaouen; Organometallics 1995,

14, 3296.

[145] N. J. Long, Metallocenes, 1st ed.; Blackwell Science, Oxford, 1998.

[146] A. Gorfti, M. Salmain, G. Jaouen, M. J. McGlinchey, A. Benouna, A. Mousser;

Organometallics 1996, 15, 142.

[147] J. W. Faller, D. F. Chodosh, D. Katahira; J. Organomet. Chem. 1980, 187, 227.

[148] W. E. Vanarsdale, J. K. Kochi; J. Organomet. Chem. 1986, 317, 215.

[149] S. Wan, M. J. Begley, P. Mountford; J. Organomet. Chem. 1995, 489, C28.

[150] M. J. Begley, P. Mountford, P. J. Stewart, D. Swallow, S. Wan; J. Chem. Soc., Dalton

Trans. 1996, 1323.

[151] M. Green, T. D. McGrath, R. L. Thomas, A. P. Walker; J. Organomet. Chem. 1997,

532, 61.

[152] N. W. Murrall, A. J. Welch; Acta Cryst. 1984, C40, 401.

[153] C. L. Perrin, T. J. Dwyer; Chem. Rev. 1990, 90, 935.

[154] J. W. Faller, B. C. Whitmore; Organometallics 1986, 5, 752.

References

207

[155] P. S. Braterman; Metal Carbonyl Spectra, Academic Press, London, 1975.

[156] W. Beck, W. Petri, J. Meder; J. Organomet. Chem. 1980, 191, 73.

[157] H.-J. Meder, W. Beck; Z. Naturforsch. 1986, 41b, 1247.

[158] M. D. Curtis, O. Eisenstein; Organometallics 1984, 3, 887.

[159] P. Espinet, R. Hernando, G. Iturbe, F. Villafane, A. G. Orpen, I. Pascual; Eur. J. Inorg.

Chem. 2000, 1031.

[160] B. J. Brisdon, A. A. Woolf; J. Chem. Soc., Dalton Trans. 1978, 291.

[161] M. D. Curtis, N. A. Fotinos; J. Organomet. Chem. 1984, 272, 43.

[162] B. J. Brisdon, M. Cartwright, A. G. W. Hodson, M. F. Mahon, K. C. Molloy; J.


[163] K.-B. Shiu, C.-J. Chang, S.-L. Wang, F.-L. Liao; J. Organomet. Chem. 1991, 407, 225.

[164] J. W. Faller, D. A. Haitko, R. D. Adams, D. F. Chodosh, J. Am. Chem. Soc. 1979, 101,

865.

[165] E. M. Holt, S. L. Holt, K. J. Watson; J. Chem. Soc., Dalton Trans. 1973, 2444.

[166] D. A. Cooper, S. J. Rettig, A. Storr, J. Trotter; Can. J. Chem. 1986, 64, 1643.

[167] S. J. Rettig, A. Storr, J. Trotter; Can. J. Chem. 1988, 66, 97.

[168] C. Borgmann, C. Limberg, L. Zsolnai; J. Chem. Soc., Chem. Comm. 1998, 2729.

[169] A. J. Graham, R. H. Fenn; J. Organomet. Chem. 1970, 25, 173.

[170] J. Sandström, Dynamic NMR Spectroscopy, Academic Press, London, 1982.

[171] K.-B. Shiu, K.-S. Liou, C. P. Cheng, B.-R. Fang, Y. Wang, G.-H. Lee, W.-J. Vong;


[172] M. Bühl, Chem. Eur. J. 1999, 5, 3514.

[173] M. Brookhart, M. L. H. Green; J. Organomet. Chem. 1983, 250, 395.

[174] B. J. Brisdon, K. A. Conner, R. A. Walton, Organometallics 1983, 2, 1159.

[175] B. J. Brisdon, S. K. Enger, M. J. Weaver, R. A. Walton, Inorg. Chem. 1987, 26, 3340.

[176] B. J. Ransil; Rev. Mod. Phys. 1960, 32, 245.

[177] P.S. Braterman; Struct. Bonding 1972, 10, 57.

[178] P. S. Braterman; Struct. Bonding 1976, 26, 1.

[179] H. Günther, NMR Spectroscopy, John Wiley & Sons, New York, 1980.

[180] C. W. Haigh, R. B. Mallion, Org. Magn. Reson. 1972, 4, 203.

[181] G. C. K. Roberts (Ed.), NMR of Macromolecules, Oxford University Press, New York,

1993

References

208

[182] G. Snatzke (Ed.); Optical and Rotary Dispersion and Circular Dichroism in Organic

Chemistry, Heyden and Son, Ltd., London, 1967.

[183] P. Crabbé; Optical and Rotary Dispersion and Circular Dichroism in Organic

Chemistry, Holden Day, San Francisco, 1965.

[184] H. Kelm, H.-J. Krüger; Eur. J. Inorg. Chem. 1998, 1381.

[185] M. Rossignoli, P. V. Bernhardt, G. A. Lawrance, J. Chem. Soc., Dalton Trans. 1997,

4247.

[186] S. P. van Kouwenberg, E. H. Wong, G. R. Weisman, E. J. Gabe, F. L. Lee, P. Jackson,

Polyhedron 1989, 8, 2333.

[187] F. A. Cotton, G. A. Rusholme, A. Shaver, J. Coord. Chem. 1973, 3, 99.

[188] K.-B. Shiu, J. Y. Lee, Y. Wang, M.-C. Cheng, S.-L. Wang, F.-L. Liao, J. Organomet.

Chem. 1993, 453, 211.

[189] C. P. Marbella, J. H. Enemark, J. Organomet. Chem. 1982, 226, 57.

[190] T. J. Desmond, F. J. Laylor, G. Ferguson, M. Parvez, J. Organomet. Chem. 1984, 277,

91.

[191] I. T. MacLeod, E. R. T. Tiekink, C. G. Young, J. Organomet. Chem. 1996, 506, 301.

[192] G. Haselhorst, S. Stoetzel, A. Strassburger, W. Walz, K. Wieghardt, B. Nuber, J. Chem.

Soc., Dalton Trans 1993, 83.

[193] N. L. Armanasco, M. V. Baker, M. R. North, B. W. Skelton, A. H. White, J. Chem.

Soc., Dalton Trans 1998, 1145.

[194] A. Chandrasekeran, S. S. Krishnamurthy, M. Nethaji, Curr. Sci. 1991, 60, 700.

[195] A. Chandrasekeran, S. S. Krishnamurthy, M. Nethaji, J. Chem. Soc., Dalton Trans

1994, 63.

[196] G. G. Aleksandrov, V. N. Babin, A. P. Kharchenikov, Y. T. Struchkov, N. S.

Kochetkova, J. Organomet. Chem. 1984, 266, 106.

[197] C.-Y. Kuo, Y.-S. Fuh, J.-Y. Shiue, S. J. Yu, G.-H. Lee, S.-M. Peng; J. Organomet.

Chem. 1999, 588, 260.

[198] R. W. Hay, I. Fraser, G. Ferguson; J. Chem. Soc., Chem. Commun. 1987, 1715.

[199] R. W. Hay, I. Fraser, G. Ferguson; J. Chem. Soc., Dalton Trans. 1989, 2183

[200] T. M. Bockman, J. K. Kochi; New J. Chem. 1992, 16, 39.

[201] F. A. Cotton, R. M. Wing; Inorg. Chem. 1965, 4, 314.

[202] P. Chaudhuri, K. Wieghardt, Y.-H. Tsai, C. Krüger, Inorg. Chem. 1984, 23, 427.

References

209

[203] G. Backes-Dahmann, W. Herrmann, K. Wieghardt, J. Weiss, Inorg. Chem. 1985, 24,

485.

[204] S. Trofimenko, J. Am. Chem. Soc. 1966, 88, 1842.

[205] D. Burdinski, Ph. D. Thesis, Ruhr-Universität-Bochum, Germany, 1998.

[206] K. Wieghardt, W. Schmidt, B. Nuber, J. Weiss; Chem. Ber. 1979, 112, 2220.

[207] T. J. Atkins, J. E. Richman, W. F. Oettle, Org. Synth. 1978, 58, 86.

[208] K. Wieghardt, M. Hahn, W. Swiridoff, J. Weiss; Angew. Chem. 1983, 95, 499.

[209] R. Hazama, K. Umakoshi, A. Ichimura, S. Ikari, Y. Sasaki, T. Ito; Bull. Chem. Soc. Jpn.

1995, 68, 456.

[210] A. Kojima, S. Ooi, Y. Sasaki, K. Z. Suzuki, K. Saito, H. Kuroya, Bull. Chem. Soc. Jpn.

1981, 54, 2457.

[211] K. Z. Suzuki, Y. Sasaki, S. Ooi, K. Saito; Bull. Chem. Soc. Jpn. 1980, 53, 1288.

[212] M. Minelli, J. H. Enemark, R. T. C. Brownlee, M. J. O’Connor, A. G. Wedd, Coord.

Chem. Rev. 1985, 68, 169.

[213] D. Rehder, Coord. Chem. Rev. 1991, 110, 161.

[214] R. T. C. Brownlee, M. J. O’Connor, B. P. Shehan, A. G. Wedd, J. Magn. Res. 1985, 64,

142.

[215] K. Wieghardt, M. Gutmann, P. Chaudhuri, W. Gebert, M. Minelli, C. G. Young, J. H.

Enemark, Inorg. Chem. 1985, 24, 3151.

[216] A. F. Masters, R. T. C. Brownlee, M. J. O’Connor, A. G. Wedd, J. D. Cotton, J.

Organomet. Chem. 1980, 195, C17.

[217] E.C. Alyea, A. Somogyvari, in H. F. Barry and P. C. H. Mitchell (Eds.), Proceedings of

the Climax Fourth International Conference on the Chemistry and Uses of Molybdenum,

Climax Molybdenum Company, Ann Arbor, MI, 1982, p. 46.

[218] T. M. Bockman, J. K. Kochi, New. J. Chem. 1992, 16, 39.

[219] J. H. MacNeil, A. W. Roszak, M. C. Baird, K. F. Preston, A. L. Rheingold,


[220] T. Beissel, B. S. P. C. Della Vedova, K. Wieghardt, R. Boese; Inorg. Chem. 1990, 29,

1736.

[221] M. D. Curtis, K.-B. Shiu, W. M. Butler, J. C. Huffman, J. Am. Chem. Soc. 1986, 108,

3335.

[222] K.-B. Shiu, L.-Y. Lee, J. Organomet. Chem. 1988, 348, 357.

[223] M. D. Curtis, K.-B. Shiu, Inorg. Chem. 1985, 24, 1213.

References

210

[224] S. Dick, A. Weiss; Z. Naturforsch. 1997, 52b, 188.

[225] D. D. Cox, S. J. Benkovic, L. M. Bloom, F. C. Bradley, M. J. Nelson, L. Que, Jr., D. E.

Wallick, J. Am. Chem. Soc. 1988, 110, 2026.

[226] N. Niklas, O. Walter, R. Alsfasser, Eur. J. Inorg. Chem. 2000, 1723.

[227] R. Alsfasser, Personal communication.

[228] G. J. Kubas, L. S. van der Sluys, R. J. Angelici (Ed.); Inorganic Synthesis, John Wiley

& Sons: New York, 1990, Vol. 28, 29.

[229] K. Schögl, Mh. Chem. 1957, 88, 601.

[230] R. S. Herrick, R. M. Jarrett, T. P. Curran, D. R. Dragoli, M. B. Flaherty, S. E.

Lindyberg, R. S. Slate, L. C. Thornton, Tetrahedron Lett. 1996, 30, 5289.

[231] M. Oberhoff, L. Duda, J. Karl, R. Mohr, G. Erker, R. Fröhlich, M. Grehl;


[232] H.-B. Kraatz, J. Lusztyk, G. D. Enright; Inorg. Chem. 1997, 36, 2400.

[233] J. F. Gallagher, P. T. M. Kenny, M. J. Sheeny; Acta Cryst. 1999, C55, 1257.

[234] J. F. Gallagher, P. T. M. Kenny, M. J. Sheeny; Inorg. Chem. Comm. 1999, 2, 200.

[235] H.-B. Kraatz, D. M. Leek, A. Houmam, G. D. Enright, J. Lusztyk, D. D. M. Wayner; J.


[236] A. Nomoto, T. Moriuchi, S. Yamazaki, A. Ogawa, T. Hirao; Chem. Commun. 1998,

1963.

[237] T. Moriuchi, A. Nomoto, K. Yoshida, T. Hirao; Organometallics 2001, 20, 1008.

[238] A. Togni, T. Hayashi (Eds.); Ferrocenes, Weinheim, New York, 1995.

[239] T. Moriuchi, A. Nomoto, K. Yoshida, T. Hirao, J. Organomet. Chem. 1999, 589, 50.

[240] T. Moriuchi, A. Nomoto, K. Yoshida, A. Ogawa, T. Hirao, J. Am. Chem. Soc. 2001,

123, 68.

[241] J. E. Sheats, M. D. Rausch; J. Org. Chem. 1970, 35, 3245.

[242] M. C. Grossel, M. R. Goldspink, J. A. Hriljac, S. C. Weston, Organometallics 1991, 10,

851.

[243] L. Lin, A. Berces, H.-B. Kraatz; J. Organomet. Chem. 1998, 556, 11.

[244] T. Moriuchi, I. Ikeda, T. Hirao, Organometallics 1995, 14, 3578.

[245] T. A. Albright, J.K. Burdett, M.-H. Whangbo; Orbital Interactions in Chemistry, John

Wiley and Sons, New York, 1985.

[246] C. D. Hall, I. P. Danks, S. C. Nyburg, A. W. Parkins, N. W. Sharpe, Organometallics

1990, 9, 1602.

References

211

[247] C. N. Kirsten, T. H. Schrader, J. Am . Chem. Soc. 1997, 119, 12061.

[248] J. C. Green, Struct. Bond. 1981, 43, 37.

[249] B. Corain, B. Longato, G. Favero, D. Ajò, G. Pilloni, U. Russo, F. R. Kreissel, Inorg.

Chim. Acta 1989, 157, 259.

[250] R. M. G. Roberts, J. Silver, J. Organomet. Chem. 1984, 263, 235.

[251] M.-H. Koepp, H. Wendt, H. Strehlow; Z. Elektrochem., Ber. Bunsenges. Phys. Chem.

1960, 64, 483.

[252] D. R. van Staveren, T. Weyhermüller, N. Metzler-Nolte, Organometallics 2000, 19,

3730.

[253] D.R. van Staveren, E. Bothe, T. Weyhermüller, N. Metzler-Nolte, J. Chem. Soc., Chem.

Comm. 2001, 131.

[254] A. D. Becke; Phys. Rev. A 1988, 38, 3098.

[255] J. P. Perdew, Phys. Rev. B 1986, 33, 8822.

[256] J. P. Perdew, Phys. Rev. B 1986, 34, 7406.

[257] D. Andrae, U. Häußermann, M. Dolg, H. Stoll, H. Preuß; Theor. Chim. Acta 1990, 77,

123.

[258] A. Bergner, M. Dolg, W. Küchle, H. Stoll, H. Preuß; Mol. Phys. 1993, 80, 1431.

[259] T. H. Dunning, J. Chem. Phys. 1970, 53, 2823.

[260] J. P. Perdew, in Electronic Structure of Solids (Eds.: P. Ziesche and H. Eischrig),

Akademie Verlag, Berlin, 1991.

[261] J. P. Perdew, Y. Wang, Phys. Rev. B 1992, 45, 13244.

[262] S. Huzinaga, M. Klobukowski, J. Mol. Struct. 1988, 167, 1.

[263] W. J. Hehre, R. Ditchfield, J. A. Pople, J. Chem. Phys. 1972, 56, 2257.

[264] P. C. Hariharan, J. A. Pople, Theor. Chim. Acta. 1973, 28, 213.

[265] W. Koch, M. C. Holthausen, A Chemist‘s Guide to Density Functional Theory, Wiley-

VCH, Weinheim, 2000.

[266] M. J. Frisch, J. A. Pople et al., Gaussian98, Gaussian, Inc., Pittsburgh PA, 1998.

212

Curriculum Vitae

Name: Dave Richard van Staveren

Birth date: 10-07-1976

Place of birth: Rotterdam, The Netherlands

Nationality: Dutch

Education

1988-1994: Voorbereidend Wetenschappelijk Onderwijs at Christelijke

scholengemeenschap “Melanchton” in Rotterdam, The Netherlands

1994-1999: Chemistry studies at the Rijksuniversiteit Leiden, Leiden, The

Netherlands

26-08-1999: M. Sc. degree obtained from the Rijksuniversiteit Leiden, Leiden, The

Netherlands

From july 1999: Ph. D. student at the Max-Planck-Institut für Strahlenchemie, Mülheim

/ Ruhr and the Ruhr-Universität Bochum

Research experience

09-1996 – 05-1998 M. Sc. research project at the Rijksuniversiteit Leiden, Leiden, The

Netherlands, in the group of Prof. Dr. Jan Reedijk

05-1997 – 08-1997 Summer research at Grinnell College, Grinnell, Iowa, U.S.A., in the

group of Dr. Martin Minelli

01-1999 – 07-1999 Research at the Max-Planck-Institut für Strahlenchemie, Mülheim /

Ruhr, Germany, as an exchange student in the group of Dr. Nils

Metzler-Nolte

07-1999 – Ph. D. research at Max-Planck-Institut für Strahlenchemie, Mülheim /

Ruhr, Germany, in the group of Prof. Dr. Nils Metzler-Nolte

213

List of publications

1) D. R. van Staveren, G. A. van Albada, J. G. Haasnoot, S. Gorter, J. Reedijk

Synthesis, characterization and luminescence properties of lanthanide(III) compounds

with 2,6-di(hydroxymethyl)pyridine and the related 2,6-di(hydroxymethyl-d2)pyridine

ligand. X-ray structure of [Sm(2,6-di(hydroxymethyl)pyridine)3](NO3)3

Inorg. Chim. Acta 2000, 302, 1104-1108.

2) D. R. van Staveren, J. G. Haasnoot, A.-M. Manotti-Lanfredi, S. Menzer, P. J.

Nieuwenhuizen, A. L. Spek, F. Ugozzoli, J. Reedijk

Synthesis and X-ray crystal structures of two novel dinuclear methoxo-bridged Er-

complexes with 2,2'-bipyridine and 1,10-phenanthroline

Inorg. Chim. Acta 2000, 307, 81-87.

3) D. R. van Staveren, T. Weyhermüller, N. Metzler-Nolte

The Mo(ηηηη-allyl)(CO)2 Moiety as a Robust Marker Group in Bio-Organometallic

Chemistry. Unusual Crystal Structure of the Phenylalanine Derivative Mo(C5H4-CO-

Phe-OMe)(ηηηη-allyl)(CO)2

Organometallics 2000, 19, 3730-3735

3) D. R. van Staveren, G. A. van Albada, J. G. Haasnoot, H. Kooijman, A.-M. Manotti-

Lanfredi, P. J. Nieuwenhuizen, A. L. Spek, F. Ugozzoli, T. Weyhermüller, J. Reedijk,

Increase in coordination number of lanthanide complexes with 2,2´-bipyridine and 1,10-

phenanthroline by using ββββ-diketonates with electron-withdrawing groups

Inorg. Chim. Acta 2001, 315, 163-171.

5) D. R. van Staveren, E. Bothe, T. Weyhermüller, N. Metzler-Nolte

Variable temperature electrochemistry as a powerful method for conformational

investigations on the fluxional organometallic complex Mo(His-Nεεεε-C2H4CO2Me)(ηηηη-

allyl)(CO)2 (His = Nδδδδ, N, O-L-Histidinate)

J. Chem. Soc., Chem. Comm. 2001, 131-132.

Documents

Robust transition metal markers for labelling of … transition metal markers for labelling of peptides via solid phase synthesis methods Dissertation for the degree of Doktor der