Università degli Studi di Cagliari - people.unica.it · Università degli Studi di Cagliari...

Università degli Studi di Cagliari Dottorato di Ricerca in Chimica, XX ciclo

NUCLEAR MAGNETIC RESONANCE OF NUCLEAR MAGNETIC RESONANCE OF NUCLEAR MAGNETIC RESONANCE OF NUCLEAR MAGNETIC RESONANCE OF 129129129129XeXeXeXe

USED AS A PROBE USED AS A PROBE USED AS A PROBE USED AS A PROBE

FOR THE STRUCTURAL CHARACTERIZATION OFFOR THE STRUCTURAL CHARACTERIZATION OFFOR THE STRUCTURAL CHARACTERIZATION OFFOR THE STRUCTURAL CHARACTERIZATION OF

POROUS MATERIALS AND PROTEINSPOROUS MATERIALS AND PROTEINSPOROUS MATERIALS AND PROTEINSPOROUS MATERIALS AND PROTEINS

Supervisor Candidate

Prof. Mariano Casu Roberto Anedda

January 2008

i

Acknowledgments

I would like to spend few words to thank all the people who have helped and

guided me throughout the doctorate. Each of the people I have worked with has given an

important contribution to the work described here.

First of all I would like to thank my supervisor, Prof. Mariano Casu, for his

constant encouragement, guidance and freedom. His continuous support and the diligent

reviewing of the manuscript have been much appreciated.

All the people I have worked with in Cagliari have patiently discussed with me all

the details of the experimental work and scientific concepts related to the systems

studied. In this regard, I have to sincerely thank the groups of biologists from the

Department of Applied Sciences in Biosystems of the University of Cagliari. In

particular, Drs. Antonella Fais, Benedetta Era, Simona Porcu and Prof. Marcella Corda

for their constant and meticulous efforts in explaining a number of complex concepts of

biochemistry of myoglobins and hemoproteins in general. Thanks to Antonella, Simona

and Benedetta also for their truthful friendship.

Prof. Giovanni Floris, Drs. Rosaria Medda, Alessandra Padiglia, Anna Mura,

Silvia Longu and Francesca Pintus for their collaboration, guidance and for the helpuful

discussions on Amine Oxidases.

A significant work has been done thanks to the collaboration with Prof. Paolo

Ruggerone and Dr. Matteo Ceccarelli, two researchers of CNR-INFM SLACS,

Department of Physics, University of Cagliari and CNR-INFM CRS DEMOCRITOS,

SISSA, Trieste. Their extensive knowledge of biophysical processes together with their

experience in molecular dynamics simulations and the very helpful comments and

discussions regarding the whole work has been essential to the writing of this final

manuscript.

The Materials Structure and Function Group within the Steacie Institute for

Molecular Science of the National Research Council of Canada is greatly acknowledged.

First of all I want to thank John Ripmeester for the opportunity he has given me to

work at NRC Canada and to activate a fruitful scientific collaboration.

Among the people I have worked with in Canada, Dima Soldatov was one of the

most important teachers and mentors.

ii

I also show my appreciation to Igor Moudrakovski who has always patiently

answered my questions concerning NMR, for the careful review of the manuscript on

Xenon and dipeptides, for his kind friendship while I was in Canada. My

acknowledgments go also to Chris Ratcliffe for his suggestions on solid state NMR

measurements.

Steve Lang and Gennady Ananchenko were great office mates and friends, their

help and hospitality has been important and really appreciated.

My thanks go also to Long Li Lai from Taiwan, for his help in studying

dendrimers and viologen inclusion compounds, a work which is still in progress.

I show my gratitude to Kostia Udatchin for helping me with X-ray

crystallographic measurements.

Robin, Satoshi and his family, Rasnish and his wife, Phil Brown, Shane Pawsey,

were pleasant and amusing buddies.

I would like also to express my gratitude to Michaela Pojarova for her guidance

and for introducing me to Canadian life.

Finally, I have to thank the people closest to me, with whom I spend most of my

lifetime; my parents for their constant support and encouragement, this work would not

have been possible without their gratifying words and continuous support; my brother

and my sister since they have been very supportive and I learnt a lot from them; all my

friends and collegues in Cagliari; at last but not the least my sweet love, she is the most

beautiful source of motivation and perseverance.

iii

Preface (Italiano)

L’utilizzo dello Xenon ha conosciuto un enorme sviluppo negli ultimi anni.

E’ particolarmente affascinante a mio avviso notare oggi quali grandi progressi

abbia fatto la tecnica Xenon-NMR successivamente agli studi preliminari risalenti ai

primi anni ‘80.

L’applicazione di questa tecnica abbraccia oggi numerosi campi della scienza e

della tecnica: la ricerca fondamentale sui composti organici, inorganici e biologici sia allo

stato solido che in soluzione, la caratterizzazione dei materiali solidi porosi che trovano

impiego nell’industria e nell’alta tecnologia, le applicazioni in campo medico su sistemi

in vitro e in vivo.

La scienza dei materiali, prima fra tutte, ha tratto numerosi vantaggi da questa

tecnica, come testimoniano le numerose pubblicazioni scientifiche di grande rilievo

riguardanti la caratterizzazione strutturale di catalizzatori, setacci molecolari, dispositivi

per l’immagazzinamento dei gas, idrati, clatrati e composti di inclusione, materiali

nanostrutturati e nanocompositi, materiali stimuli-responsive, sistemi per drug delivery.

Sebbene le dimensioni, il volume dei pori e l’area superficiale di un materiale poroso

sono determinabili mediante TEM e principalmente attraverso l’adsorbimento di azoto

BET, queste tecniche non forniscono sufficienti informazioni sulla connettività e struttura

delle superfici interne dei pori.

In campo biologico, l’utilizzo dello Xenon come sonda ha permesso

l’individuazione e la caratterizzazione strutturale di cavita’ all’interno di proteine ed

enzimi fornendo importanti indicazioni sul processo di diffusione dei ligandi e substrati

in bio-macromolecole e sulla relazione struttura-funzione in sistemi biologici

relativamente complessi.

Il settore medico diagnostico ha giovato soprattutto degli sviluppi della tecnica

Xenon-MRI per l’acquisizione di immagini. In particolare, l’uso di tecniche che

permettono di incrementare il segnale NMR dello Xenon di diversi ordini di grandezza

(iperpolarizzazione) ha permesso di ottenere significativi e promettenti risultati nello

studio dei polmoni, nell’acquisizione di immagini angiografiche, nello studio del cervello

e nella diagnosi precoce dei tumori. Inoltre, come e’ noto, lo Xenon viene usato come

iv

anestetico generale, ma tuttoggi il meccanismo molecolare di azione di questi agenti e’

elusivo e ulteriori studi a riguardo sono necessari.

Il mio dottorato e’ stato svolto prevalentemente all’Universita’ di Cagliari con la

supervisione di Prof Mariano Casu. Una collaborazione con il National Research Council

of Canada mi ha permesso di lavorare per un anno nei laboratori dello Steacie Institute

for Molecular Science di Ottawa. Questa importante esperienza di collaborazione

scientifica, ancora attiva, ha avuto per supervisore John Ripmeester, uno dei pionieri di

questa tecnica, che guida un gruppo (Materials Structure and Function Group) di oltre 30

ricercatori di varia estrazione scientifica.

Questa tesi e’ organizzata in quattro capitoli che trattano i diversi sistemi studiati

durante il dottorato. In particolare, il primo capitolo introduce i concetti generali sui quali

la tecnica Xenon-NMR si basa e elenca alcuni obiettivi del progetto. La descrizione dei

sistemi studiati e’ riportata nel secondo capitolo. Nel terzo capitolo, Risultati e

Discussione, sono descritti e commentati i risultati sperimentali ottenuti. Alla fine di ogni

sottocapitolo della sezione Risultati e Discussione si traggono alcune conclusioni e

considerazioni generali, in particolare cercando di sottolineare le novita’ introdotte da

questo lavoro e discutere possibili sviluppi futuri della tecnica Xenon-NMR. Il quarto

capitolo e’ lasciato alla descrizione dei metodi sperimentali usati per preparare i campioni

analizzati, per acquisire i dati e della tecnica di iperpolarizzazione dello Xenon.

La tesi e’ scritta in lingua inglese, ormai diventata la lingua ufficiale della

comunicazione scientifica, con la speranza di permettere la lettura del lavoro ad un

gruppo piu’ numeroso ed eterogeneo di ricercatori.

v

Table of Contents

Chapter I

Introduction …………………………………………………………………………1

1.1 Generalities………………………………………………………....……2

1.2 Objectives…………………………………………………………..…….3

1.2.1 Proteins……………………………………………………….……..3

1.2.2 Microporous crystalline dipeptides……………….…………………5

1.3 NMR properties of Xenon…………………………….…………………6

1.4 Bibliography……………………………………………….……………11

Chapter II

The systems studied: void space in biomolecules……...……………14

2.1 Myoglobins: suitable model systems…………………………..………15

2.1.1 Function……………………………………………………………15

2.1.2 Structure……………………………………………………………16

2.1.3 Cavities in myoglobins…………………………………….………20

2.2 Copper-containing Amine Oxidases (AOs)……………..…………….24

2.2.1 Structure ………………………………………………..………….24

2.2.2 Hydrophobic cavities in AOs……………………………...……….26

2.2.3 Biological function: AOs’ catalytic process………………….……27

2.3 Biomaterials: microporous crystalline dipeptides……………...…….31

2.3.1 Developments of microporous materials……………………..……31

2.3.2 Characterization of bioorganic materials…………………………..32

2.3.3 Microporous dipeptides structure………………………….………36

2.3.4 129Xe NMR of dipeptides microporous crystals……………...…….39

2.4 Bibliography……………………………………………………….……41

vi

Chapter III

Results and discussion………………………………………………………….48

3.1 Myoglobins………………………………………………………...……49

3.1.1 129Xe NMR measurements in solutions of low-spin

(Fe3+ S=1/2) cyano-metmyoglobins…………………………….….52

3.1.2 129Xe NMR relaxation measurements

of CNMb solutions…………………………………………………62

3.1.3 1H NMR chemical shift in CNMbs from horse and pig……………70

3.1.4 NOE measurements used as a tool

to further assess His93 rotation relative to heme………………..…79

3.1.5 Thermodynamics of Xenon binding to

cyano-metmyoglobins from Xenon-induced 1H NMR

chemical shift variations……………………………………...……84

3.1.6 Myoglobins: CONCLUSIONS………………………………….…87

3.2 Copper containing Amine Oxidases enzymes:

Xenon-induced reactions………………………………………….……89

3.2.1 Lens Esculenta Amine oxidases (LSAO) in solution: 129Xe NMR chemical shifts …………………………………..……89

3.2.2 Spectral changes in the UV-vis region of LSAO solutions

induced by substrates and Xenon………………………………..…91

3.2.3 Involvement of a lysine residue

in the intra-molecular catalytic mechanism of LSAO………..……99

3.2.4 129Xe NMR of PKAO, ELAO and LSAO solutions…………...…105

3.2.5 Spectroscopic features induced by amine substrates

and Xenon in several AOs……………………………………..…107

3.2.6 Copper containing Amine Oxidases: CONCLUSION………...…113

3.3 Microporous Crystalline Dipeptides…………………………………115

3.3.1 Variable Temperature continuous flow HP 129Xe NMR:

General spectral features…………………………………….……115

3.3.2 Temperature dependence of the 129Xe CSA tensor……………..119

vii

3.3.2.1 Effect of channels loading on the CSA ………………...…124

3.3.2.2 Presence of specific sites (niches)……………………...…125

3.3.2.3 Effect of helicity and diameter of the channels on CSA…..125

3.3.2.4 Dynamics of Xe in the cross section of the pores

and CSA of 129Xe NMR signal…………………………….127

3.3.2 129Xe NMR isotropic chemical shifts

as a function of temperature………………………………………128

3.3.3 Thermodynamics of adsorption: the Langmuir model……………132

3.3.4 Signal intensities………………………………………………….138

3.3.5 Aging of dipeptide samples………………………………………143

3.3.6 Thermodynamics of adsorption in nanochannels.

General remarks (summary)…………………………………...…146

3.3.7 Dipeptides: CONCLUSIONS………………………………….…152

3.4 Bibliography……………………………………………………...……154

Chapter IV

Materials and methods……………………………………………….………164

4.1 Hyperpolarized Xenon: solving sensitivity problems………………...…165

4.1.1 Continuous-flow measurements………………………………..…167

4.1.2 Advantages (and drawbacks) implied in the use of

hyperpolarized and thermally polarized Xenon………………..…169

4.2 Myoglobins...................................................................................................171

4.3 Microporous Dipeptides..............................................................................174

4.4 Copper containing Amine Oxidases...........................................................175

4.5 Bibliography………………………………………………………….……178

Papers Published during the doctorate ............................................................180

CHAPTER I - INTRODUCTION

1

Chapter I

Introduction


2

1.1 Generalities

The scientific and technologic relevance of using Xenon atoms as probes for the

characterization of void spaces comprised within biological macromolecules and/or

porous materials appears clear when the extensive literature on this topic is considered.

It is even more evident from the most recent achievements the usefulness of combining

the sensitivity of Xenon and the versatility of a spectroscopic technique such as Nuclear

Magnetic Resonance in order to deeply characterize both structures and dynamics

involved in the host-guest systems.

Early studies proposed, discussed and demonstrated the usefulness of 129Xe NMR

in the characterization of void space in systems of different nature1,2. This technique is

useful in studying porous materials for gas sensing3, purification4, separation and

storage5,6, catalysis processes7,8. It has medical applications as well, as it allows for

acquisition of images of lungs, heart, kidneys and brain9-12 and helps in the challenging

studies that concern the understanding of the molecular mechanisms of the action of

general anesthetics13,14.

Recently, there has been renewed interest in Xenon NMR in view of using the

resonance of Xenon in the structural study of proteins15-24. In biochemistry, it is useful

for instance in characterizing ligand binding and diffusion within cavities of

biomolecules, just to mention one of its many applications in this field.

Moreover, among the most intriguing results, recent application of Xenon

biosensors made of conveniently functionalized Cryptophane-Xenon complexes has

allowed target-specific detections of specific proteins and oligonucleotide sequences by

means of Magnetic Resonance Spectroscopy and Imaging25-27.

Among the many advantages of combining Xenon as a biomolecular probe and

NMR as spectroscopic technique is that NMR parameters of nuclei belonging to both the

probe and the matrix can be studied. It has been demonstrated that, in fact, important

information on host-guest interactions in Xenon complexes can be derived both from

direct observation of 129Xe NMR signals and from Xenon-induced chemical shift changes

in 1H, 13C and 15N nuclei of the host compounds as well24,28-30. Moreover, polarization

transfer from hyperpolarized Xenon to protons by Spin Polarization-Induced Nuclear

Overhauser Effect (SPINOE) can be studied in suitable systems15,31-36. These latter


3

experiments, beyond confirming the results obtained by directly observing the guest

Xenon atoms, provide site-specific information, which is particularly valuable especially

when exchange processes average 129Xe NMR results and also when particularly complex

systems, such as large biomolecules, are under study.

1.2 Objectives.

The studies that will be described in the following have been devoted to test the ability of

Xenon as an efficient probe of void spaces. We have decided to approach the problem by

analyzing both porous crystals in the solid state and very flexible molecules such as

proteins in solution.

1.2.1. - Proteins

a) Myoglobin (Mb), a globular protein, has been selected to probe the ability of Xenon to

extract important information on their structure and function. Studying Xenon binding to

a model compound such as myoglobin represents a useful approach but at the same time

challenging due to the presence of different interaction sites within the protein, beside the

heme iron. The presence of four cavities37 has been directly evidenced by X-ray

diffraction on sperm whale Mb crystals pressurized by Xenon.

However, this early studies were carried out only on Mb crystals, which suffer of

the drawbacks related to the lower flexibility of the overall protein compared to the

protein in solution. Previous studies of Xe-Mb complexes in solution have been

performed on horse Myoglobins, confirming the XRD studies19,.

Clearly, hints on the structure of cavities in solution and on Xe-protein affinity

can be substantiated by comparatively studying myoglobins of different species as in a

recent NMR investigation on the pig and horse metmyoglobins38 (MMbs). There, the

combined use of the 129Xe chemical shift and the 129Xe spin lattice relaxation rate as a

function of Xenon and protein concentration has unraveled the influence of the structure

and/or hydrophobicity of a cavity on its Xenon occupancy.


4

To exploit more completely the peculiar properties of the NMR technique aiming

at a deeper description of the Mb cavities, we combine in the present study the analysis of 129Xe NMR chemical shifts and relaxation rates to an accurate and appropriate 1H NMR

characterization of the proteins. In order to substantiate and complete the conclusions

made by Corda et al.38, we extend the comparison between pig and horse Mbs to two

more Mbs, those of sheep and rabbit. In particular, we focus our attention on the protein

in the low-spin cyano form (CNMb), of which several 1H signals of the residues in the

active site have been already assigned.

A final and promising issue of the present work concerns the possible use of

specific Xenon-cavities interactions as probe to monitor the displacements of the

individual protons induced by the Xenon binding. The host-guest interaction is a subtle

and not negligible aspect of a spectroscopic technique, since it determines the accuracy

and validity of the data analysis. In particular, we examine the behaviour of the residues

in the proximal and distal cavities of pig and horse Mb and verify the potential

occupation of Xenon in the cavities close to the myoglobin active site.

b) It is generally believed that Xenon atoms can induce structural changes in some

of the cavities or channels that they are bound to, both in solution39 and in the solid

state40. Xenon has been used as a probe for dioxygen-binding cavities in copper AOs by

recording XRD data under pressure of Xenon gas41-43. Here is discussed our investigation

on the binding of Xenon to purified lentil (Lens esculenta) seedling copper/TPQ-amine

oxidase in solution. Upon pressurization with 10 atm of Xenon gas the enzyme can

generate the free radical intermediate in the absence of substrates outlining a process that

probably involves a lysine residue at the active site. The study has been extended to

highly purifed AOs from various sources and our results strongly support the hypothesis

that a lysine residue is implicated in the catalytic mechanism of plant enzymes.


5

1.2.2. - Microporous crystalline dipeptides

This study was aimed at the detailed characterization of sorption in the

microporous dipeptides. In particular, fundamental thermodynamic parameters and

molecular-scale peculiarities of the sorption process were in the focus and how these

characteristics relate to the structure of the micropores. Standard approach including the

determination of sorption isotherms appeared to be hardly suitable for the materials under

study. Thermodynamic parameters may be extracted from such data provided a series of

sorption isotherms for each material is available. At the same time, the materials of this

study appeared to change upon aging and the long experimental times required for the use

of standard procedures would have resulted in unreliable results. A fast method was thus

necessary but which would give, at the same time, reliable quantitative information.

Another requirement to the method was to be able to monitor changes on the molecular

level occurring in the pores. As it was demonstrated previously40, the pores are very

flexible and the pore structure revealed in the crystal structure examination of an empty

sorbent will not account for its sorption behavior as the pores become loaded with the

guest species.

In order to overcome the above problems, we have developed and demonstrated

here an entirely different, new approach based on the determination of sorption isobars

using continuous-flow hyperpolarized 129Xe NMR. Using this approach made possible

the first systematic study of gas sorption in microporous peptides with quantitative

thermodynamic description of the process and detailed analysis of specificities occurring

on molecular level between the flexible host matrix and the included guest species.

In this work, variable-temperature 129Xe NMR experiments using a continuous

flow of hyperpolarized Xenon were conducted for the eight microporous dipeptides. It is

demonstrated that quantitative information on the thermodynamics of the sorption

process can be extracted from these experiments as well as comprehensive knowledge on

the sorption events occurring on the molecular scale level. The present study reveals the

relation of the observed NMR parameters of absorbed Xenon with the thermodynamics of

sorption, geometry and dynamics of the micropores, and the structural features of the

cavity-Xenon intraporous association.


6

1.3 NMR properties of Xenon

Any distortion of the large electronic cloud of 129Xe is felt directly at the nucleus

and consequently affects the observed NMR parameters. The most sensitive parameter is

undoubtedly the chemical shift. The wide NMR spectral window typical of non-ligated 129Xe (∼350 ppm, which becomes ∼7500 ppm when also Xenon compounds are taken

into account) allows for detailed analysis of local environment around the observed

Xenon nuclei and facilitates simultaneous detection of 129Xe in different chemical

environments. Together with the ideal physico-chemical properties of Xenon, such as

inertness and large polarizability of the spherically symmetrical electronic cloud, it

should be pointed out that NMR sensitivity of naturally occurring 129Xe is quite good.

Due to the relatively high natural abundance of the isotope 129Xe, Xenon is approximately

32 times easier to observe than 13C, neglecting differences in relaxation times, and has

about 10-2 times the sensitivity of proton. Nevertheless, Xenon-NMR suffers of the

serious problem which is typical of all the nuclei that are traditionally studied by Nuclear

Magnetic Resonance: the low sensitivity that derives from low thermal polarization. In

order to face this problem, hyperpolarization techniques have been recently developed

which have allowed for obtainment of up to six orders of magnitude signal enhancement

by optical pumping [see Section 4.1].

The suitability of Xenon as a biomolecular probe is due to many of its physical

and chemical properties. Xenon is a monoatomic, non-toxic, chemically inert gas, small

enough (Van der Waals radius ∼ 2.2 Å) to be able to probe even very narrow pores and

cavities. Due to its hydrophobic properties, it is well suited to locate and explore

hydrophobic regions such as cavities and channels in biological systems; its large and

extremely polarizable electronic cloud makes it very sensitive to its local physical

environment: in particular, this sensitivity is readily detectable by analyzing NMR

spectroscopic parameters (chemical shift, relaxation times, line shapes and possibly

chemical shift anisotropy) which are mostly influenced by the atoms in the proximity of

observed nuclei.

Naturally occurring Xenon has nine stable isotopes, only two of which, 129Xe and 131Xe, have non-zero nuclear spin I, and are therefore detectable by means of Nuclear


7

Magnetic Resonance spectroscopy. 129Xe has I=1/2 and natural abundance of 26.4%,

while 131Xe has I=3/2 and natural abundance of 21.2%.

The observed NMR parameters are simultaneously influenced by several

concurrent factors. The chemical shielding is basically influenced by two contributions: a

diamagnetic contribution σd, merely determined by the fundamental electronic state of

the atom and a paramagnetic part, which depends on the excited electronic states, i.e. on

the symmetry of the valence electronic shell.

The paramagnetic contribution is zero when a spherically symmetrical

distribution characterizes the valence electronic shell of Xe atoms, which can be observed

only for the ideal situation of isolated Xe atoms (Xe in gaseous state at a pressure

extrapolated to zero). Due to the large and easily polarizable Xe electronic cloud, the

paramagnetic contribution to the chemical shielding is expected to play a significant role

in determining the observed chemical shift of Xenon when it is interacting physically

with its environment.

This high sensitivity makes Xenon a very useful probe for the characterization of

systems which it can interact with. However, while the central idea of early researchers

was to exploit Xenon’s sensitivity to get detailed structural and chemical information on

the systems studied, it soon appeared clear that the simultaneous presence of different

factors influencing in a variable manner the observed signal often leads to complex

outputs, which are sometimes difficult to be unambiguously interpreted.

Let us concentrate first on Xenon gas. Since the discovery of gaseous Xenon44,45

subsequent studies showed that over a wide range of densities the shift is expressed by

virial expansion of the Xe density46. The most precise values of virial coefficients were

obtained by Jameson et al.47.

Early studies concentrated on studying 129Xe NMR solutions in different solvents,

which clearly demonstrated the sensitivity of Xe chemical shift to physical environments.

For example, it was demonstrated by Jokisaari and coworkers48 although 129Xe gas to

solution shifts are linearly related to 13C gas to solution shifts of methane in the same

solvents, the entity of chemical shift variations proved that 129Xe is 27.1 times more

sensitive to physical interaction with solvent. As the shifts of dissolved gas merely

depend on Van der Waals interaction, they are reasonably expected to have higher values


8

for heavy atoms. 129Xe NMR of Xenon dissolved in a number of solvents and organic

and bioorganic ligands have been reviewed by Reisse49.

If we assume now that Xe interacts with more complex systems, where more than

one site (target) is available for Xenon atoms, (i.e. many chemically and or structurally

different surfaces Si are present), the observed 129Xe NMR chemical shift (δi) will be an

average between all the possible situations, weighted for the frequency of collisions.

δi = (Term for chemical nature of Si)·(Term for frequency of Xe-Si collisions)

The observed spectrum, therefore, will depend on the lifetime of Xe on each adsorption

site and two limiting conditions can be discussed:

If the lifetime of Xe on each Si is long in the NMR time scale, the spectrum will

be constituted by as many components (each with chemical shift δi) as there are target

types, the intensity being related to the number of targets of each type in the samples.

If instead the lifetime of Xe in each site is very short (fast exchange condition) all the

signals coalesce and the spectrum will therefore consist of only one component whose

chemical shift depends on the values of δi each weighted by the probability αi of the Xe-

Si collision:

δobs = Σ αiδi with Σ αi=1 [1.1]

In general, we can consider the observed 129Xe NMR chemical shift as influenced

simultaneously by the following contributions, as proposed by Fraissard50,51:

δobs = δref + δs + δXe-Xe + δSAS + δE + δM [1.2]

where δref is the chemical shift of Xe gas at zero pressure, which is considered the

reference value and fixed to zero; δs is due to interaction between Xe and the surface of

the sample, thus it provides structural and chemical information on each site where the

Xe atoms interact with (dimensions and shape of cage/channels, ease of Xe diffusion etc);

δXe-Xe arises from the interaction of two or more Xe atoms in cages, channels or pores


9

that can contain more than one Xe at the same time. This latter contribution evidently

depends on Xe density. Whenever strong adsorption sites (SAS) are present, Xe atoms

will spend a longer time in contact with them than the cage or channel walls, particularly

at low Xe concentrations.

δE and δM are related to the presence of electric and magnetic fields, that

sometimes arise from the presence of charged ions, paramagnetic metals and/or radical

species.

Longitudinal relaxation time of gaseous Xenon is in principle influenced, in a

homogeneous magnetic field, only by spin-rotation during collisions, according to the

relation

T1 ≈ 56/ρ [1.3]

Where ρ is Xe density in amagat and T1 is in hours. However, experimental results show

that in fact the measured T1 is generally less than the ideal value expressed by the

previous equation. This is basically due to collisions between Xe and the walls of the

sample. This problem was shown to be relevant when hyperpolarized Xenon is used, as

loss of polarization (i.e. longitudinal relaxation) causes considerable loss of the signal

previously enhanced by laser pumping. In this regard, it has been shown that pretreating

the pumping cell’s walls with polymeric coating (Surfasil-Pierce), T1 longer than 20

minutes can be obtained.

While the study of relaxation times has led to interesting achievements in the

characterization of different systems in solution, its employment for solid materials has

not provided the same results. It should be observed that although nuclear relaxation time

T1 of adsorbed Xenon should ideally provide interesting information about local Xenon

structure and dynamics, this study could be carried out reliably only on extremely pure

systems, as paramagnetic impurities are often present in real catalysts and other solid

porous systems. The effect of paramagnetic ions in enhancing nuclear relaxation is well

known, thus in presence of hyperfine coupling between unpaired electrons and nuclear

spins all the other possible sources of relaxation become negligible.


10

While in solid materials the presence of paramagnetic sites may generate

unwelcome problems, in solution of paramagnetic proteins it can give very useful

information. Hemoglobin and myoglobin were the first two proteins shown to bind

Xenon. Hemoglobin is currently more difficult to study via NMR because of its large

size, but exploitation of Xe NMR on myoglobins, which is by now relatively well

characterized, has given important insight on the structure of Xe-myoglobin complexes.

An example of 129Xe NMR relaxation studies in myoglobins containing the heme iron ion

in the high spin (S=5/2) form have been discussed by Locci et al. and Corda et al.23,38,

who discussed a method to obtain Xe-Fe distances from the analysis of T1.

We describe here Xenon-binding systems and we correlate results obtained by

NMR and other techniques to describe structural and dynamical characteristics of each

system and to discuss similarities and differences between them.

The systems which will be described throughout this thesis are myoglobins and

peroxidases in solution and solid microporous crystalline dipeptides.


11

1.4 Bibliography

1. Goodson BM, J. Magn. Reson. (2002), 155, 157–216

2. C. I. Ratcliffe, Ann. Rep. NMR Spectrosc., (1998), 36, 123-221

3. K Knagge, JR Smith, LJ Smith, JBuriak, D Raftery Solid State Nucl Magn Reson

(2006), 29(1-3), 85-89

4. Acosta RH, Agulles-Pedros L, Komin S, Sebastiani D, Spiess HW, Blumler P,

Phys Chem Chem Phys (2006), 8(36), 4182-4188

5. KJ Ooms, RE Wasylishen Micropor Mesopor Mater (2007), 103, 341–351

6. P. Sozzani, S. Bracco, A. Comotti, L. Ferretti, R. Simonutti Angew. Chem. Int.

Ed., (2005), 44, 1816-1820

7. Bonardet, J., Fraissard J., Gedeon A., Sprinuel-Huet M.. Catal. Rev. Sci. Eng.

(1999), 41, 115-225;

8. MJ. Annen, ME Davis BE Hanson Catal Lett (1990), 6, 331-339

9. SD Swanson, MS Rosen, BW Agranoff, KP Coulter, RC Welsh, TE Chupp Magn.

Reson. Med (2005), 38(5), 695 – 698

10. Driehuys B, Science (2006), 314, 432-433

11. Rosen MS, Chupp TE, Coulter KP, Welsh RC, Swanson SD Rev Sci Instr (1999),

70(2), 1546-1552

12. SD Swanson, MS Rosen, KP Coulter, RC Welsh, TE Chupp, Magn Reson Med

(1999), 42(6), 1137 – 1145

13. Eckenhoff RG, Johansson JS, Pharmacol. Rev (1997), 49(4), 343-367

14. JW Tanner, JS Johansson, PA Liebman, RG Eckenhoff Biochemistry (2001), 40,

5075-5080

15. G. Navon, Y.Q. Song, T. Room, S. Appelt, R.E. Taylor, A. Pines, Science (1996),

271, 1848– 1851

16. J. Wolber, A. Cherubini, A.S. Dzik-Jurasz, M.O. Leach, A. Bifone, Proc. Natl.

Acad. Sci. U. S. A. (1999), 96, 3664–3669

17. C.R. Bowers, V. Storhaug, C.E. Webster, J. Bharatam , A. Cottone III, R. Gianna,

K. Betsey, B.J. Gaffney, J. Am. Chem. Soc. (1999), 121, 9370–9377

18. A. Stith, T.K. Hitchens, D.P. Hinton, S.S. Berr, B. Driehuys, J.R. Brokeman, R.G.

Bryant, J. Magn. Reson. (1999), 139, 225– 231;


12

19. S.M. Rubin, M.M. Spence, B.M. Goodson, D.E. Wemmer, A. Pines Proc. Natl.

Acad. Sci. (2000), 97, 3472–9475 ;

20. E. Locci, Y. Dehouck, M. Casu, G. Saba, A. Lai, M. Luhmer, J. Reisse, K. Bartik,

J. Magn. Reson. (2001), 150, 167–174;

21. S.M. Rubin, M.M. Spence, A. Pines, D.E. Wemmer, J. Magn. Reson. (2001),

152, 79– 86;

22. S.M. Rubin, M.M. Spence, I.E. Dimitrov, E.J. Ruiz, A. Pines, D.E. Wemmer, J.

Am. Chem. Soc. (2001), 123, 8616–8617;

23. E. Locci, M. Casu, G. Saba, A. Lai, J. Reisse, K. Batik, ChemPhysChem (2002),

3, 812– 814;

24. S.M. Rubin, S.Y. Lee, E.J. Ruiz, A. Pines, D.E. Wemmer, J. Mol. Biol. (2002),

322, 425– 440

25. Schröder L, Lowery TJ, Hilty C, Wemmer DE, Pines A, Science (2006),

314(5798), 466-449;

26. Q. Wei, G.K. Seward, P.A. Hill, B. Patton, I. Dimitrov, N.N. Kuzma, I.J.

Dmochowski, J. Am. Chem. Soc., (2006), 128, 13274-13283;

27. V. Roy, T. Brotin, J.P. Dutasta, M.H. Charles, T. Delair, F. Mallet, G. Huber, H.

Desvaux, Y. Boulard, P. Berthault. ChemPhysChem (2007), 8(14), 2082-2085

28. Gröger C, Möglich A, Pons M, Koch B, Hengstenberg W, Kalbitzer HR, Brunner

E, J. Am. Chem. Soc. (2003), 125, 8726-8727;

29. TJ Lowery, SM Rubin, EJ Ruiz, A Pines DE Wemmer, Angew Chem Int Ed

(2004), 116, 2-4;

30. L Dubois , P Da Silva , C Landon , JG Huber , M Ponchet , F Vovelle , P

Berthault , H Desvaux J Am Chem Soc. (2004), 126 (48), 15738-15746

31. Room, T.; Appelt, S.; Seydoux, R.; Hahn, E. L.; Pines, A. Phys. Rev. B (1997),

55, 11604-11610.

32. Raftery, D.; MacNamara, E.; Fisher, G.; Rice, C. V.; Smith, J. J. Am. Chem Soc.

(1997), 119, 8746-8747

33. Pietrass, T.; Seydoux, R.; Pines, A. J. Magn. Reson. (1998), 133, 299-303.

34. MacNamara, E.; Rice, C. V.; Smith, J.; Smith, L. J.; Raftery, D. Chem. Phys. Lett.

(2000), 317, 165-173.


13

35. Song, Y.-Q. Concepts Magn. Reson. (2000), 12, 6-20.

36. C Landon, P Berthault, F Vovelle, H Desvaux, Protein Science (2001), 10, 762-

770

37. RF Tilton, ID Kuntz, GA Petsko, Biochemistry (1984), 23, 2849-2857

38. Corda M, Era B, Fais A, Casu M, Biochim. Biophys. Acta (2004), 1674,182-192

39. Moglich A, Koch B, Gronwald W, Hengstenberg W, Brunner E & Kalbitzer HR,

Eur J Biochem, (2004), 271, 4815–4824

40. Soldatov DV, Moudrakovsky IL, Grachev EV & Ripmeester JA J Am Chem Soc

(2006), 128, 6737–6744

41. Duff AP, Cohen AE, Ellis PJ, Kuchar JA, Langley DB, Shepard EM, Dooley DM,

Freeman HC & Mitchell Guss J Biochemistry (2003), 42, 15148–15157;

42. Duff AP, Trambaiolo DM, Cohen AE, Ellis PJ, Juda GA, Shepard EM, Langley

DB, Dooley DM, Freeman HC, Mitchell Guss J J Mol Biol (2004), 344, 599–

607;

43. BJ Johnson, J Cohen, RW Welford, AR Pearson, K Schulten, JP Klinman, CM.

Wilmot J Biol Chem (2007), 282(24), 17767–17776

44. Proctor WG, Yu FE, Phys Rev (1950), 78, 471;

45. Proctor WG, Yu FE, Phys Rev (1951), 81, 20

46. Brinkmann D, Brun E, Staub HH, Helv Phys Acta (1962), 35, 431

47. Jameson AK, Jameson CJ, Gutowski HS, J Chem Phys, (1970), 53, 2310

48. Diehl P, Ugolini R, Suryaprakash N, Jokisaari J, Magn Reson Chem (1991), 29,

1163

49. Reisse New J Chem (1986), 10, 665

50. Fraissard J, Ito T, Zeolites (1988), 8, 350 and references therein;

51. Springuel-Huet MA, Bonardet JL, Fraissard J, Appl Magn Reson (1995), 8, 427

CHAPTER II – DESCRIPTION OF THE SYSTEMS STUDIED

14

Chapter II

The systems studied: void space in biomolecules


15

2.1 Myoglobins: suitable model systems

Myoglobins (Mb) are intracellular hemoproteins that reversibly bind molecular oxygen

(O2). They are expressed in the myocytes of cardiac tissues and in striated muscular

fibers of type I and II vertebrates1.

Earliest studies on myoglobin were carried out by Millikan in late ‘30s and

resulted in a comprehensive review2 in which he assembled a significant body of

knowledge to establish that myoglobin is formed adaptively in tissues in response to the

demand for oxygen. Subsequent important results achieved by many authors further

assessed myoglobin structure and function and this class of proteins is nowadays one of

the most studied systems and commonly believed to be a very useful model compound to

investigate the important issue of structure-function relationship. Additionally, the

extensive experimental results on myoglobin make it a prime example for testing the

applicability of various theoretical techniques to proteins3,4.

2.1.1. - Myoglobin function.

Functionally, myoglobin is well accepted as an O2-storage protein in muscle,

capable of releasing O2 during periods of hypoxia or anoxia. Myoglobin is also thought to

buffer intracellular O2 concentration when muscle activity increases and to facilitate

intracellular O2 diffusion by providing a parallel path that augments simple diffusion of

dissolved O2. It has been extensively demonstrated that the function of myoglobin is

carried out with remarkable variability with genetic origin of the polypeptide chain; the

role of many aminoacids in modulating the process of ligand binding and diffusion has

been deeply investigated by means of computational methods and experimental analysis

on point mutants of myoglobins5,6. Several studies have demonstrated that myoglobin

carries out many other important functions beyond serving as an O2 reservoir and

transporter. In this regard, it was recently proposed7 that oxymyoglobin (MbO2) can also

play the role of intracellular scavenger of nitric oxide (NO), thus protecting respiration in

skeletal muscles and heart. NO, in fact, is known to reversibly inhibit Cytocrome-c

oxidases, the terminal enzyme of the mitochondrial respiratory chain8-12. It was also

pointed out that myoglobin supports oxidative phosphorilation13.


16

Another important ligand with high affinity for myoglobins is carbon monoxide

(CO). Carbon monoxide binds coordinately to heme iron atoms in a manner similar to

that of oxygen, but the binding of carbon monoxide to heme is much stronger than that of

oxygen. The preferential binding of carbon monoxide to heme iron is largely responsible

for the asphyxiation that results from carbon monoxide poisoning. Several amino acids

play important role in regulating the binding of different ligands to myoglobins and in

determining the selectivity of this protein. Among all, distal histidine E7 (His64) in

vertebrate myoglobins has been strongly conserved during evolution and is thought to be

important in fine-tuning the ligand affinities of these proteins14,15.

2.1.2. - Myoglobin structure.

Myoglobin has relatively small size (Mr ∼17 600) and it is formed by a single

polypeptidic chain of 153 aminoacids and a protoporphyrin IX heme prosthetic group, a

tetrapirrole to which is bound an iron atom, identical to that of hemoglobins. The iron

atom (green ball in Figure 2.1) forms five coordination bonds in the deoxy form of

myoglobin, four of which with the nitrogen atoms (in blue) belonging to the tetrapyrrole

and one with the Nε of the imidazolic ring of proximal His93(F8) (in yellow), which has a

particular relevance in stabilizing the heme group.

Figure 2.1: Heme group (Fe-protoporphyrin IX) in deoxy-myoglobin.


17

The sixth coordination position, at the opposite side of the heme plane with respect to

proximal histidine, can be occupied either by oxygen in oxymyoglobin or by other

potential ligands such as CO (Carboxy-Mb), H2O (Meta-myoglobin), N3 (Azide-Mb), CN

(Cyano-Meta-Mb) and NO (Nitroso-Mb).

The polypeptidic chain is arranged in eight separate right handed α-helices,

designated A through H, that are connected by short non helical regions in a highly

conserved globular fold. Amino acid R-groups packed into the interior of the molecule

are predominantly hydrophobic in character while those exposed to the solvent on the

surface of the molecule are generally hydrophilic, thus making the molecule relatively

water soluble. The heme prosthetic group is buried within a hydrophobic cleft of the

globin, sandwiched between the E and F helices. In particular, the heme group is placed

between two Histidine residues that significantly influence the overall function of

myoglobin: the proximal His93(F8) and the distal His64(E7). The helices B, C and E

(purple region in Figure 2.2) form the so called distal side of the active site, while the

proximal side is lined by residues belonging to the helix F (yellow in Figure 2.2).

Figure 2.2: Helices B, C and E (in purple) form the distal region of myoglobins, while helix F (in yellow)

lines the proximal side. In the picture are evidenced proximal histidine (His93) in orange, heme group in

red and iron atom in green.


18

The protoporphirin ring, embedded within the folded globin, is stabilized by Van der

Waals forces or through hydrophobic interactions with the non-polar side chains of the

residues lining the active site16. In particular some residues, Leu89 (F4), His97 (FG3),

Ile99 (FG5) and Leu104 (G5), have been demonstrated to be crucial in stabilizing the

heme group firmly linking it to the globin: Ile99 and Leu104 residues are localized in the

internal region of the heme pocket, close to the prosthetic group and are believed to act as

barriers to the water molecules of the solvent; the residue Leu89 has the same function,

being at the entrance of the hydrophobic active site. Substitutions in these positions have

significant effect on the dissociation of heme17. On the other side of the prosthetic group,

in the distal region, His97 forms a hydrogen bond with 7-propionate, thus dividing the

interior cavity from the solvent. Propionic groups in positions 6 and 7 of the heme ring

are exposed to the solvent and interact with polar residues on the surface of the protein.

Mb can exist in both the reduced iron(II) state as well as oxidized iron(III) state,

and both the diamagnetic and paramagnetic derivatives have been the subject of intense

physicochemical studies to elucidate the mechanism of control of ligand binding18. The

electronic configuration the valence shell of Fe is 3d64s2 while that of Fe2+ ion is 3d6 and

that of Fe3+ ion is d5. When Fe is coordinated with heme within the protein, a splitting of

the energy level characterizes each coordination environment according to the ligand

field theory [Scheme 2.1].

Scheme 2.1: Ligand field theory predicts the splittings of the d orbitals of heme iron.


19

High-spin iron(III) represents the resting state form of many heme proteins. In

this form the heme iron is pentacoordinated with proximal His ligand, or hexacoordinated

with a water molecule in the sixth coordination site. High spin ferric heme proteins have 6A ground state. The electronic configuration of a high-spin d5 system is shown in

Scheme 2.2 reported below:

Scheme 2.2: Electronic configuration of a high-spin d5 system

The low-spin ferric heme cyano-metmyoglobins (CNMMbs) represent an

important subclass of paramagnetic metalloproteins whose 1H NMR spectral parameters

contain a wealth of structural information. Due to the low paramagnetism of these

systems characterized by S = 1/2, they represented the material for the first applications

of new NMR techniques to paramagnetic proteins.

Scheme 2.3: Electronic configuration of a low-spin d5 system


20

NMR assignments for the hyperfine-shifted resonances in this system have relied

primarily on comparisons with model compounds19-21 analysis of paramagnetic

relaxations22,23 and by isotope labeling of heme protons24.

In 1983-1985 some papers appeared which represented the first application of 1D

NOE to a paramagnetic metalloprotein25-28. Moreover, it has been demonstrated that

approximately 95% of the protons within 7.5 Å of the ferric iron of CNMMb can be

assigned on the basis of 2D NOESY29 by exploiting the X-ray crystal coordinates30 to

interpret the cross peaks. It was also demonstrated that these assignments can be used to

determine the orientation of the magnetic axes in solution29.

2.1.3. - Cavities in myoglobin.

Myoglobin has been the first protein to be crystallized and resolved at atomic

resolution31.

The pioneering crystallographic studies of Shoenborn and coworkers evidenced

the presence of cavities within Sperm Whale myoglobin (SW Mb) crystals able to bind

Xenon under moderate pressure32,33. These Xenon complexes were in fact obtained by

subjecting native protein crystals to relatively low gas pressure (2-2.5 bars) and it was

suggested that Xenon atoms are bound to pre-existing atomic-sized cavities in the interior

of these globular proteins through weak Van der Waals forces. Subsequent work carried

out by Tilton and co-workers34 clearly showed that the number and the occupancies of

Xenon binding sites vary with the applied pressure. Thus, at a pressure of 7 bars, only

one fully occupied principal binding site, the proximal cavity (Xe1), was found in Sperm

Whale met-myoglobin crystals, and three additional secondary cavities were

characterized having lower Xenon occupancy. These four sites, referred to as Xe1, Xe2,

Xe3 and Xe4, are shown in Figure 2.3 below.


21

Figure 2.3: Crystal structure of Sperm Whale Myoglobin complexed with Xenon. In red is represented the

prosthetic heme group, in orange, licorice style, is shown proximal His93(F8) and yellow balls are Xenon

atoms bound to the four principal hydrophobic cavities referred to as Xe1, Xe2, Xe3 and Xe4.

Xenon 1 is bound in the proximal pocket and is essentially fully occupied. Xenon

2 binds directly below the proximal cavity in a relatively small cleft near the bottom of

the heme group and it was shown to have very close contacts with protein atoms, in

particular Cε1 and Cδ1 of Phe138(H14). Xenon 3 is located in a cavity lined by residues

belonging to the E-F corner and to the H helix near the surface of the protein. Finally,

Xenon 4 is on the distal side directly below the O2 binding site. It was shown by Tilton

and coworkers that, according to X-ray crystal structures, Xenon binds to myoglobin with

only very little perturbation of the local environment. It was observed that in fact non-

neglibible positional changes correspond to aminoacids such as Leu89 and Ile142, which

have close contacts with Xenon into the main binding site Xe1. One more feature that

characterizes Xenon binding to myoglobin according to Tilton et al. is the observation of

an overall decrease (about 13%) of the temperature factors of both backbone and side

chains atoms, some particular regions, such as helices close to Xe1, being more affected

than others. It should be pointed out in this regard that thermal factors are of interest as


22

they measure dynamic and conformational disorder and this led the authors to the

conclusion that atomic motions in myoglobin are decreased upon Xenon inclusion.

However, although proximal region was shown to be much more influenced by

this restricted motion, the overall flexibility of protein atoms was affected, that can be

explained only by considering a likely ligand-induced restriction of the number of

conformational substates, i.e. a “freezing” of ligand-stabilized substates. The presence of

different conformational states of myoglobin, each having particular activation energy,

was postulated by Austin and coworkers35 and further confirmed by means of FT-IR on

CO-Mb36 and two dimensional infra-red based vibrational echo experiments37.

The role of hydrophobic voids in biomolecules is the source of a very intense

scientific debate. Early researchers referred to those internal cavities using the term

“packing defects” and the uncertainty on their actual relevance in favoring the dynamics

of protein molecules lasted until recently. However, these cavities exist at the expense of

considerable cost in free energy, so that it is unlikely that they are mere packing defects.

Moreover, binding of Xenon to myoglobin has been shown to significantly affect the

functionality of the protein38,39

The hypothesis that cavities are important for the conformational flexibility of

protein molecules is further supported by the observation that conformational states of

myoglobin are restricted by Xenon binding to protein cavities40. At the same time, the

observation of binding of small ligands to internal voids buried within proteins would

contrast with the static representation of cavities as voids closed off by the protein atoms

if concerted movements of the protein backbone and side chains were not taken into

account. It is in fact commonly believed that diffusion of ligands into proteins is made

possible by transient formation of pores, channels and pathways which are not observed

in the average picture usually obtained by crystallographic structures.

Cavities, therefore, are generally hydrophobic, are able to bind exogenous ligands

such as Xenon, which are stabilized by non covalent specific interactions32. Cavities,

moreover, permit proteins to have a stable structure and to perform their function, as they

represent the best compromise between thermodynamic stability and flexibility: it has

been suggested in fact that they are important in catalyzing reactions41,42 and in tracing

pathways for the diffusion of ligands to and from the active site.


23

Both molecular dynamics simulations34,43,44 and laser photolysis studies as well as

time-resolved crystallography45-47 have enlightened the key role of Mb cavities in ligand

dynamics. Within this framework, the competition with Xenon in occupying these

strategic sites has also been exploited for testing possible routes followed by the ligands

inside the protein45,48.

One among the most intriguing features to be understood about Xenon binding to

protein cavities is the diversity of possible binding sites. They can be channel-pores and

pockets transiently exposed to the solvent and they can be buried inaccessible cavities as

well. Moreover Xenon is found to bind to inter- as well as to intra-molecular sites49.

Many studies have exploited Xenon-induced variation of NMR chemical shift of

nuclei which the Xenon comes in contact with when it is included in cavities or

channels50-54. Commonly, 15N, 13C, and 1H NMR chemical shift variations are considered

in order to extract thermodynamical parameters of binding and to detect and characterize

the sites where the complexed Xenon resides51,53-55. Such studies, however, are usually

applied to diamagnetic systems, and, to our knowledge, no attempt has been presented in

literature so far to explain similar results in paramagnetic biomolecules. Deepening the

knowledge of paramagnetic interactions in model proteins as myoglobin is, however,

relevant, as many heme-proteins, along with many other metallo-proteins, exist in their

paramagnetic states.

Analyzing the observed chemical shift and its variations in paramagnetic

biological compound is more challenging than in diamagnetic molecules because many

different contributions, essentially related to the presence of the unpaired electron in the

atomic orbitals of the metal ion, finally influence the experimental result.

The influence of guest molecules on proton NMR spectra of myoglobins was

observed in early studies56,57; however, despite it was clearly demonstrated that the

presence of cyclopropane and Xenon within the internal cavities of myoglobins caused

modifications in the experimental proton NMR spectra no detailed explanation of that

result was attempted. This is probably due to the poorly resolved and not yet assigned

spectra obtained and to the limited knowledge of these systems at that time. Further

studies, mainly NMR and EPR measurements58,59 together with model calculations60,


24

successively expanded the investigation and showed the relevance of parameters such as

Fe-1H dipolar through-space interactions, orientation of unpaired spin density and

electron and spin delocalization from iron to porphyrin orbitals in establishing the

structure of the heme cavity and consequently influencing the observed proton spectra61-

63.

2.2 Copper-containing Amine Oxidases

Copper/quinone–containing amine oxidases [amine:oxygen oxidoreductase

(deaminating)(copper containing); EC 1.4.3.6] (Cu/TPQ AOs) are found in bacteria,

yeasts, fungi, plants and mammals.

2.2.1. - Structure.

Amine oxidases are homodimers: each subunit (molecular mass ∼ 70-90 kDa)

contains an active site composed of a tightly bound Cu(II) and a quinone of 2,4,5-

trihydroxyphenilalanine (TPQ or TOPA). The protein-derived cofactor TPQ is generated

by an endogenous tyrosine residue through a self-catalytic reaction with copper divalent

ions and molecular oxygen64 and has a crucial role in the catalytic process of Copper

amine oxidases, defined for plant amine oxidases.

Figure 2.4: Biogenesis of TPQ

P r o t e i n

O H

O H 2 O

O P r o t e i n

O

+ + 2O 2 + Cu(II)

TPQ

+ H 2 O 2 + Cu(II) + OH -


25

So far, six AOs have been successfully crystallized and their structure has been

resolved by XRD65-70.

Figure 2.5: Crystal structure of a eukaryotic (pea seedling) Copper containing amine oxidase. The picture

shows the structure of dimers where TPQ (in red) cofactors and Copper ions (in orange) are evidenced

The Copper ion is coordinated with the imidazol groups of three conserved

histidine residues and with two water molecules, arranged in a distorted square base

pyramidal geometry (Fig. 2.6).

Figure 2.6: Structure of Copper sites: three histidines (residues His603, His442 and His444) and a water

molecule (the oxygen is a red sphere) are shown. The fifth position is expected to be occupied by another

water molecule to form a distorted square base pyramidal conformation.


26

2.2.2. - Hydrophobic cavities in AOs.

XRD structures of AOs show, similarly to what has been observed in many other

proteins and enzymes, that the active site is buried in a cavity not directly accessible from

the solvent. X-ray crystal structures of Copper-AOs bound to Xenon are available from

bacteria (Arthrobacter globiformis), yeast (Pichia Pastoris), plant (Pisum Sativum), and

mammalian sources (bovine serum albumine oxidases)70,71.

A recent investigation of a Copper-containing Amine Oxidases from Hansenula

polymorpha by means of a combination of XRD analysis on HPAO single-crystals in

presence of Xenon gas, kinetics and computational approaches have given evidences for

the existence of at least four binding sites for Xe inside these AOs .

Figure 2.7: In figure are shown the four Xe sites found in HPAO by means of X-ray Crystallography. In

the picture is shown the structure of a dimer where one of the monomers is whitened. Xenon atoms are

represented as yellow spheres, Copper ions are depicted in orange and TPQ cofactor is evidenced in

licorice style.

Xe1

Xe2

Xe3

Xe4


27

2.2.3. - Biological function: AOs’ catalytic process.

Copper AOs catalyze the conversion of two substrates, primary amines and

molecular oxygen, to aldheydes and hydrogen peroxide, respectively.

The oxidative deamination occurs by transfer of two electrons from amines to molecular

oxygen72.

The ping–pong catalytic mechanism of Cu/TPQ AOs can be basically divided into two

half–reactions: the first, referred to as reductive half–reaction, involves the oxidation of

amine to aldehyde and the formation of a reduced form of the TPQ cofactor:

Eox + R–CH2–NH3+ → Ered + R–CHO [2.1]

The second half–reaction, known as oxidative half–reaction, involves the reoxidation of

the enzyme with the simultaneous release of ammonia and hydrogen peroxide:

Ered + O2 + H2O → Eox + NH4+ + H2O2 [2.2]

A number of biochemical investigations have been carried out in order to shine a

light on the molecular mechanisms implied in both biogenesis of TPQ and catalytic cycle

of Copper AOs, but the debate is still ongoing. In particular, while quite definite results

tend to confirm the reductive half reaction, somewhat unclear appears the mechanism of

activation of the molecular oxygen in the oxidative step of the cycle, which remains

subject of intense study73-76.

Somewhat contentious has appeared the role of Copper ion in the catalytic process

of these amine oxidases and this issue has been the focus of recent controversy. As

Copper AOs contain Cu(II) ion in the active site, it was suggested that Cu(I) ion is likely

responsible of reacting with O2 to give Cu(II)-superoxide77. In anaerobic conditions, in

fact, amine induced reduction of Copper AOs shows an equilibrium between Cu(II)-

aminoquinol and Cu(I)-semiquinone with yelds of Cu(I)-semiquinone depending on the

particular enzyme source78. A catalytic mechanism proposed for plant AOs is reported in

the scheme below79,80.


28

+

CuII

(I)

O

OO

NH3CH2 R

CuI

CuII

+ H2O

-

-

H++

H+

O

NH2

OH

R CHO

OH

HO

NH2

(II)

O2H2O2-

NH4-

+(III)

Figure 2.8: Left: scheme of the proposed catalytic process of Copper amine oxydases defined for plant

AOs. On the right, the active sites where Copper ions (orange) and TPQ cofactors (cyan=carbons;

red=oxygens; blue=nitrogens) are evidenced.

Following the reaction scheme reported in the left side of Figure 2.8, three

principal steps can be described. First (I), the amine substrate reacts with the TPQ

cofactor of the oxidized enzyme to give the Shiff base Cu(II)-quinone ketimine, a short

lived species which is rapidly converted, through the formation of an unstable Cu(II)-

carbanion, another Cu(II)-quinolaldimine Shiff base and release of the aldehyde, into the

reduced Cu(II)-aminoquinol derivative (II), which binds an ammonia molecule ( a more

detailed reaction scheme will be discussed in the Results and discussion, see sections

3.2.2 and 3.2.3) .

It was suggested and demonstrated by EPR measurements that Cu(II)-

aminoquinol (II) forms the yellow-colored intermediate Cu(I)-semiquinolamine (III)

radical in anaerobic conditions78. This latter species, observable only in absence of

oxygen, contains the substrate-derived nitrogen which is covalently bound to the aromatic

ring system and is characterized by a typical UV-vis spectrum having characteristic

absorption bands at 464, 434 and 360 nm,81,82 similar to that shown below:


29

Figure 2.9: UV-vis absorption spectrum of the yellow-colored intermediate Cu(I)-semiquinolamine radical

observed in anaerobic conditions.

Spectroscopic features have been explained by considering both electronic

transitions associated with the quinone and also the possible influence of Copper-cofactor

charge transfer (LMCT transitions) was hypothesized. Ligand field transitions of the

Cu(II) ion in amine oxidases in presence of exogenous ligands that stabilize Cu(I), such

as CN-, have been characterized by means of circular dichroism83.

Both forms of the reduced enzyme (II and III) can further react with molecular

oxygen (if present) to release hydrogen peroxide and ammonia, thereby regenerating the

Cu(II)-quinone species84,85.

Understanding the molecular mechanisms implied in enzymatic activities is of

fundamental relevance. In particular, the role of many aminoacids in the overall catalytic

process of Amine oxidases has been suggested especially on the base of crystal

structures. In this case, a critical issue would be the study and substantiation of possible

pathways involved in the processes of migration and binding of molecular oxygen, as this

species is certainly involved in the enzymatic redox reactions.

Among all the methods that are usually adopted in order to study O2 migration pathways

inside the cavities Xenon atoms represent ideal probes which can be used to investigate

the interior of proteins and enzymes. Because of their analogous properties in size and


30

hydrophobicity, any region that binds Xenon is usually assumed to be favourable for

O2.87

Although protein crystallographers have used Xenon derivatives in order to get

isomorphous form of protein crystals and thus acquire diffraction phase parameters

aiming to elucidate biological activity of proteins and enzyme, it should not be forgotten

that Xenon can sometimes participate in biological reactions.

Although only very few examples in literature give evidence of the relevant role

of Xenon in triggering and/or catalyzing biological reactions, neglecting this possibility a

priori may lead to erroneous conclusion. In this regard, recently the role of Xenon in

increasing electron spin intersystem crossing rates in chemical and enzymatic reactions

with radical pair intermediates has been discussed88-94


31

2.3 Biomaterials: microporous crystalline dipeptides

2.3.1. - Developments of microporous materials.

Microporous materials are currently the subjects of widespread studies because of

the availability of well defined and ordered void space of different sizes and shapes and

due to the variety of possible applications. Current and future possible applications of

these materials include industrial catalysis, gas sensing and storage, isolation and

purification technologies, stabilization of pharmaceuticals, biological molecules and

reactive species, inertization of hazardous waste materials.

Until mid 1990s there were basically two types of microporous materials, namely

inorganic and carbon-based materials.

In the case of microporous inorganic solids the largest two subclasses are the

aluminosilicates and aluminophosphates. Several related crystalline oxides such as

silicoaluminophosphates, metallosilicates, metalloaluminophosphates, but also porous

chalcogenides, halides and nitrides have been discovered.

Carbon-based materials represent another important example of widespread used

microporous materials. A principal negative aspect is however that in this type of

materials microporosity is usually very disordered and a very detailed systematic study of

sorption process is often disadvantaged.

Development of organically-based microporous materials assembled from

building blocks represents a very actual issue in material science, supramolecular

chemistry and crystal engineering, mainly due to the remarkable diversity of possible

modes of assembly and to the multitude of final structural motifs attainable. These

materials offer many important advantages with respect to the more commonly used

inorganic counterparts such as zeolites, clays and various metal oxides. In organic-related

systems, chemical and structural modulations can be introduced in small increments over

a wide range to create a desirable property or function. At the same time, the size and

geometry of the free void space are crucial factors in determining the properties of

microporous materials. In many cases, the species included within the pores are weakly

linked by non-covalent bonds to the host matrix, frequently by only Van der Waals

interactions. Therefore selectivity, total capacity, thermodynamics and kinetics of the


32

inclusion process are often merely determined, and may be therefore predicted and/or

regulated, by knowing the pore geometry.

In the literature, a significant number of examples can be found of the synthesis

and characterization of molecular details of the void space in organic95-97, hybrid metal-

organic97-100 and protein/peptide97,101-104 solid microporous frameworks.

Particularly interesting results have been recently obtained on the characterization of void

space in flexible pore systems such as biozeolites, a new group of microporous materials

based on peptides.

The first peptide-based system whose building blocks form nanotubes in the solid

phase were described in 1975 for cyclic α−β−α−β peptides105. Ghadiri and coworkers

have continued in 1990s to use cyclic peptides with eight to twelve residues106. β-sheet-

like intermolecular hydrogen bonds between the peptidic units formed in both cases

tubular structures.

Other research groups since then have kept working at the synthesis and

characterization of peptide-based nanotubes in crystalline compounds and the variety of

structure modulations is evident107,108.

The first example of much smaller peptide-chains forming nanotubes in the solid

state was reported by Görbitz and Gundersen in 1996 and was represented by dipeptides

L-Val-L-Ala as building units109 which form helical nanochannels resulting from head-to-

tail hydrogen bonds between functional groups of dipeptide molecules and with

hydrophobic inner walls. Other dipeptides were then studied: it seems to be a common

pattern among oligopeptides that, taking into account their combinatorial diversity might

entirely revolutionize the domain of engineering microporous solids.

2.3.2. - Characterization of bioorganic materials

Studies on oligo-peptides formed by long aminoacid chains have shown that these

materials usually show complex behavior and are often difficult to characterize in detail:

consequently, building-blocks formed by smaller peptides can be considered as to

represent simpler model systems to test the fundamental properties of these systems.

Moreover, short-chain peptides are cheaper and it is easier to deal with them for

crystallizations and structural refinement. Sorption studies have already revealed size-


33

matching molecular recognition by the peptidic channels102 and the design of peptide

sorbents that are highly selective to a particular guest will likely be possible taking into

account their diversity.

Gas sorption was tested for some dipeptides nanochannels and it was found that

their frameworks show a high sorption capacity and high selectivity for inert species like

Xenon. This permitted pioneer researchers to refer to these biomaterials with the term

biozeolites. Further advantages of biozeolites are their biocompatibility and

environmental friendliness. Moreover, microporous dipeptides display not only

similarities to inorganic sorbents (uniform pore geometry, ordered porosity,

thermodynamic stable phases), but also some characteristics which are typical of

proteins, such as flexibility of the pores and structural softness. Microporous

oligopeptide-based chains are in fact considered very suitable models of biological ion

channels and may be used for further understanding the mechanisms involved in the

inclusion of ligands in proteins.

Dipeptides represent in fact a rare opportunity to study ordered microporous

biologically-related solid materials that maintain, perhaps to a less extent, the

characteristic flexibility of protein in solution: this could result in a system which

possesses the characteristic stability of crystals and the specificity (selectivity) peculiar of

proteins and enzymes. In fact, the flexibility of the dipeptides channels is well known 97,102,110-112 that “dipeptides structures can be considered as belonging to a class of “soft”

sorbents which tend to adapt their structure relative to the presence, concentration and

chemical nature of the guest species”.

While the development of new synthetic processes enables the design of tailored

materials with pores of known volume and geometry, the continuous improvements of

the characterization techniques make easier the understanding of the correlation between

their structural and functional properties. A successful characterization of nanoporous

materials relies upon the establishment and optimization of suitable techniques that are

able to highlight even subtle details of the structure under investigation. It is reasonable

to think that the selectivity toward different guests that often characterizes nanoporous

channels can be significantly influenced by small differences in the geometry, chemical

composition and flexibility of void space. This is particularly true if a biological


34

functionality is involved: the understanding of the molecular mechanisms underlying the

action of biological ion-channels and receptors is still the subject of vivid interest and

their study gathers the efforts of researchers belonging to different scientific branches. At

the same time, the synthesis and characterization of organically-based soft nanoporous

materials has recently paved the way for a new material science, due to a number of

distinctions that these sorbents show compared to the more diffused inorganic porous

materials such as zeolites, aluminophosphates, activated carbons, silicas, clays, etc113.

Previuos studies on nanochannels of crystalline dipeptides97,112 have shown that

the information elicited about the porosity of such materials appears to be strictly

dependent on the technique employed for their characterization. Although the results

obtained by means of different methods have found to be roughly coincident, it appeared

clear that the dynamics of the host matrix might play a crucial role in determining the

sorption of these soft materials and the reliability of the characterization techniques as

well. In particular, the consistency of single crystal XRD structures with respect to other

characterization techniques has been objected due to the average and static nature of the

information provided97. In this respect, a complete and comprehensive description of soft

materials would require the combination of different and complementary techniques and

the comparison between the results obtained. In order to gain an exhaustive

understanding of the structure and properties of flexible nanoporous materials, therefore,

the improvement and testing of methods that are able to give insights on the dynamic

nature of the sorption mechanism appear to be essential.

The specificity of the dipeptides nanochannels respect to the complexation of

simple chemical species was already observed by Gorbitz et al111. In particular, a fully

retention of I2 into the nanochannels space of LS dipeptides was observed and local

structural adaptations were suggested in order to explain the low degree of channels

filling: however, the efforts, by means of TGA/DSC technique, addressed to the

characterization of the release process of I2 were unsuccessful.

In this regard, is known that most of biological-related nanochannels can be switched

between an open and a closed state and that this transition can be triggered by the binding

of a ligand. In the case of microporous dipeptides complexes the characterization of


35

possible structural rearrangements and the study of their reversibility could be an

interesting challenge.

Other experimental results and calculations based on simple dipeptides systems

suggest the influence of guest Xenon atoms on modification of the internal pore

structure112, however a compelling explanation of these phenomena has never been

attempted.

Moreover, the diffusion through such ultramicropores and the possible trapping of

small organic and inorganic moieties into the nanopore space, along with the capability of

dipeptides to sustain guest solvent exchange and full removal, make the deep

characterization of these phenomena a very interesting challenge.

Despite the fundamental interest on this materials and their numerous possible

applications108, a full characterization of the structure-function correlation and dynamics

of such systems is still not available.


36

2.3.3. - Microporous dipeptides structure.

In this thesis eight different dipeptide structures (all LL-isomers) will be discussed

in particular: Ala-Val (AV), Val-Ala (VA), Leu-Ser (LS), Ala-Ile (AI), Val-Val (VV),

Ile-Ala (IA), Ile-Val (IV), Val-Ile (VI). Formulas of all dipeptides are represented in

Figure 2.10.

Figure 2.10: Molecular structure of the dipeptides units studied

The crystal structures of the eight dipeptides can be divided into two isostructural

groups: seven of them, formed by the hydrophobic residues Alanine (Ala), Valine (Val)

and Isoleucine (Ile), represent in fact an isostructural series, having hexagonal unit cell

with P61 space group. In this structure, the channels are formed by helical H-bonded

assembly of the single dipeptide molecules. As a typical example, H-bonding scheme is

shown for VV in Figure 2.11.

H 3 N +O

N H C O O -

H 3 N+ O

N H C O O -

H 3 N +O

N H C O O

-

H 3 N +O

N H C O O

-

H 3 N +O

N H C O O -

H 3 N +O

N H C O O -

H 3 N+

O NH C O

-

H3 N +O

NH C O O

-

O H

AV

VA

AI

VV

IA

IV

VI

LS

O


37

Figure 2.11: Fragment of the crystal structure of VV (hexagonal, space group P61) showing H-bonding

helical assembly of the dipeptide molecules surrounding the channel (view along the channel). Two

translational periods are shown. N and O atoms are drawn as small black and white balls, respectively. H-

bonds are designated with grey lines.

The channels are one-dimensional and isolated from each other. Single file

diffusion of Xenon atoms inside these nanochannels appears very reasonable from the

analysis of channel diameters by means of XRD crystal structures and other independent

techniques97,114 and it was recently demonstrated by Bowers and coworkers114 by means

of NMR measurements of a continuous flow of hyperpolarized 129Xe. The interior of

channels is lined by hydrophobic parts of the aminoacids and the void space inside the

nanochannels is essentially determined by the size of hydrocarbon fragments of the

dipeptide molecules. The diameter of the nanotubes (which does not equal the inner

diameter of the channels) is identified by the a parameter of the unit cell and varies

from14.2 Å to 10.4 Å and the translation period varies from 10 to 10.4 Å and includes a

screw rotation.

Leu-Ser dipeptides (LS) form a different structure. Crystals have hexagonal unit

cells and P65 space group. Similarly to the other dipeptides, LS units form tubular

channels with approximately 15 Å diameters by means of intermolecular H-bonds.

VV


38

Translation periods are of about 6.5 Å. However, in this case, the inner walls are lined by

hydrophobic fragments of only leucyl residues, as it is shown in Figure 2.12.

Figure 2.12: Fragment of the crystal structure of LS (hexagonal, space group P65). Atoms are represented

as previously described (see Fig 2.11).

Also for LS crystals, channels are isolated from each other. A significant

difference with respect to the other series of dipeptides is that the helicity of LS channels

is left-handed, while in AV, VA, AI, IA, VV, IV and VI the helicity is right-handed. This

characteristic is very important as it has been suggested that crystalline dipeptides can be

used in the separation of organic compounds by chiral recognition. Moreover, while in

LS each isolated nanotube is formed by a single chain of dipeptide molecules, in the other

dipeptides a channel wall is formed of 50% by one nanotube and for the other 50% by six

adiacent nanotubes.

LS


39

2.3.4. - 129Xe NMR of dipeptides microporous crystals. 129Xe NMR has proven to be an extremely sensitive technique for characterization

of void spaces in porous materials and proteins and the suitability of this technique to

provide dynamically averaged information is well known.

Since the earliest 129Xe NMR experiments on characterization of porous

materials, a number of different theories have been proposed, aimed at finding a good

correlation between the NMR parameters of the guest Xenon atom probes (chemical

shifts, line shapes, relaxation times) and the actual structure of the hosts. However,

although high sensitivity of the easily polarizable Xenon electronic cloud to its physical

environment potentially provides plenty of information, the experimental NMR

parameters are often influenced by a number of different concurrent contributions.

Calculations based on Lennard-Jones potentials115,116 demonstrated that, in fact, a single,

simple correlation between 129Xe NMR chemical shift and void space is not expected and

the chemical shift of Xenon adsorbed in microporous materials is a complicated function

of void space geometry, sorption energy and temperature. Considerable progress has been

made in the derivation of useful parameters for sorption from 129Xe NMR spectroscopy

of large pore sytems (d > ~ 1nm). In those cases, the line shape is dominated by

exchange and dynamics, and parameters pertaining to pore size, adsorption energy, etc,

can be derived using a simple model that uses the temperature dependence of the

chemical shift.

Rather consistent 129Xe NMR results concerning the δ-D correlation were in fact

obtained by adopting a semi-empirical approach117 based on the adsorption

thermodynamics to study a number of mesoporous amorphous silica gels with a range of

mean pore diameters from 2 to 40 nm and this approach has been successively extended

to micropores over the 0.5-2 nm range118.

Moreover, a fair number of empirically-derived equations119,120 were proposed in

the early works with the purpose of tentatively explain the correlation between isotropic

Xe chemical shifts (δ) and the pore diameter (D) mainly in zeolites and clathrates. Such

an empirical approach is still largely in use 119,121-125. Although the application of

empirical equations to a number of systems seems to support the proposed models126, this


40

approach is of only little help in understanding the physical nature of the correlation

between the 129Xe chemical shift and the pore size. Besides, their general validity has

been several times objected127-129. The analysis by 129Xe NMR of compounds having pore

sizes comparable or smaller than the Xenon Van der Waals diameter would be very

useful to complement the picture drawn so far and would eventually point out further

possible advantages or drawbacks of the technique

It appears therefore clear from what has been said above, that a more detailed

description of the correlation between the physical nature of the sorption process in

nanochannels and the 129Xe NMR outputs is highly desirable.

The study of model systems with well known chemical composition and ordered

porosity can help to deeper analyze how different effects contribute to the experimental

results, allows extracting general rules on the sorption mechanism and can be useful in

providing hints for the interpretation of more complex systems. Dipeptides nanotubes

represent a very suitable class of compounds for this purpose: as previous studies

confirm97, in fact, they form one-dimensional nanochannels with cylindrical or nearly

cylindrical cross-section and diameter of the same order of magnitude of that of the

Xenon atom (ranging from about 3 Å for VI to about 5.4 Å for AV). The inner walls of

the channels are formed by hydrocarbon fragments of both residues and are essentially

hydrophobic. Moreover, the self-assembling process of different dipeptides allows a fine

modulation of the diameter and the helicity of channels, giving the unique opportunity to

compare 129Xe NMR experimental results obtained from samples having the same

chemical composition but different pore geometry and vice versa.


41

2.4 Bibliography

1. Wittemberg JB, Wittemberg BA, Ann. Rev. Physiol., (1989), 51, 857-878

2. Millikan GA, Physiol Rev (1939), 19, 503-523

3. Frauenfelder H, Wolynes PG, Science (1985), 229, 337

4. Ceccarelli M, Anedda R, Casu M, Ruggerone P, Proteins, structure function and

genetics (2007), in press

5. R Varadarajan, A Szabo, SG Boxer Proc Natl Acad Sci USA (1985), 82, 5681-

5684

6. BA Springer, SG Sligar Proc Natl Acad Sci USA (1987), 84, 8961-8965

7. Brunori M., Trends in Biochem. Sci (2001), 26(1), 21-23

8. Cleeter MWJ, Cooper JM, Darley-Usmar VM, Moncada S, Shapira AHV, FEBS

Lett, (1994), 345, 50-54

9. Merx MW, A. Godecke, U. Flogel, J Schrader, FASEB J., (2005), 19, 1015-1017

10. Giuffre’ A, E Forte, M Brunori, and P Sarti, FEBS Lett, (2005), 579, 2528-2532

11. Flogel U, MW Merx, A Godecke, UKM Decking, J Schrader, Proc. Natl. Acad.

Sci. U.S.A. (2001), 98, 735-740

12. Garry DJ, A Meeson, Z Yan, RS Williams Cell Mol Life Sci (2000), 57, 896-898

13. Wittemberg JB, Wittemberg BA, J Exp Biol (2003), 206, 2011-2020

14. Collman, J. P. Accts Chem. Res. (1977), 10, 265−273

15. Perutz, M. F. in Molecular Basis of Blood Diseases (eds Stammatoyanopoulos,

G., Nienhaus, A. W., Leder, P. & Majerus, P. W.) 127−178 (Saunders,

Philadelphia, 1987)

16. Hargrove MS, Olson JS, Biochemistry (1996), 35, 11310-11318

17. Liong CE, “Structural and functional analysis of proximal pocket mutants of

sperm whale myoglobins” PhD Thesis, Rice University, Houston, TX. 1999

18. Antonini & Brunori, Hemoglobin and myoglobin in their reactions with ligands,

North Holland, Amsterdam, 1971

19. La Mar, G. N.; Budd, D. L.; Goff, H. M. Biochem. Biophys. Res. Commun.

(1977), 77, 104

20. Chacko, V. P.; La Mar, G. N. J. Am. Chem. Soc. (1982), 104, 7002


42

21. LaMar,G.N.; deRopp, J. S.;Chacko,V.P.;Satterlee, J.D. Biochim. Biophys. Acta

(1982), 708, 317

22. Mayer, A.; Ogawa, S.; Shulman, R. G.; Yamane, T.; Cavaleiro, J. A. S.; Rocha

Gonsalves, A. M. d’A.; Kenner, G. W.; Smith, K. M. J. Mol. Biol. (1974), 86, 749

23. Cutnell, J. D.; La Mar, G. N.; Kong, S. B. J. Am. Chem. Soc. (1981), 103, 3567

24. La Mar, G. N.; Davis, N. L.; Parish, D. W.; Smith, K. M. J. Mol. Biol. (1983),

168, 887

25. Unger, S. W.; Lecomte, J. T. J.; La Mar, G. N. J. Magn. Reson.(1986), 64, 521

26. Johnson, R. D.; Ramaprasad, S.; La Mar, G. N. J. Am. Chem. Soc. (1983), 105,

7205

27. Ramaprasad, S.; Johnson, R. D.; La Mar, G. N. J. Am. Chem. Soc. (1984), 106,

3632

28. Ramaprasad, S.; Johnson, R. D.; La Mar, G. N. J. Am. Chem. Soc. (1984), 106,

5330

29. Emerson, S. D.; La Mar, G. N. Biochemistry (1990), 29, 1545

30. Kuriyan, J.; Wilz, S.; Karplus, M.; Petsko, G. A. J. Mol. Biol. (1986), 192, 133

31. Kendrew JC, RE Dickerson, BE Strandberg, RG Hart, DR Davies, DC Phillips

and VC Shore, Nature, (1960), 185, 422-427

32. Shoenborn BP, Watson HC, Kendrew JC, Nature (London) (1965), 208, 760-762

33. Shoenborn BP, J Mol Biol (1969), 45, 297-303

34. RF Tilton, ID Kuntz, GA Petsko Biochemistry (1984), 23, 2849-2857

35. Austin RH, Beeson KW, Eisenstein L, Frauenfelder H, Gunsalus IC, Biochemistry

(1975), 14, 5355-5373

36. Hong MK, Braunstein D, Cowen BR, Frauenfelder H, Iben IE, Mourant JR,

Ormos P, Sholl R, Shulte A, Steinback PJ, Biophys J. (1990), 58, 429-436

37. K. A. Merchant, D. E. Thompson, Q. Xu, R. B. Williams, R. F. Loring, M. D.

Fayer Biophys J. (2002), 82, 3277-3288

38. Settle W, “Function of the myoglobin is influenced by anesthetic molecules” in

Guide to molecular Pharmacology-Toxicology Featherstone, RM (ed.) New York

Marcel Dekker, 1973


43

39. Nishihara Y, Sakakura M, Kimura Y, Terazima M, J. Am. Chem. Soc. (2004),

126, 11877-11888

40. T Prangé, M Schiltz, L Pernot, N Colloc'h, S Longhi, W Bourguet, R Fourme

Proteins, structure function and genetics, (1998), 30, 61-73

41. Brunori, M. and Gibson Q.H.. EMBO Reports (2001), 21, 674-679

42. Brunori, M., Bourgeois D., Vallone B. J. Struct. Biol. (2004), 147, 223-234

43. Tilton, R.F., Jr., Singh U.C., Weiner S.J., Connolly M.L., Kuntz I.D., Jr., Kollman

P.A., Max N. Case D.A. J. Mol. Biol. (1986), 192, 443-456

44. Elber R, Karplus M., J. Am. Chem. Soc. (1990), 112, 9161–9175

45. Scott EE, Gibson QH, Biochemistry (1997), 36, 11909-11917

46. Scott EE, Gibson QH, Olson JS, J. Biol. Chem. (2001), 276, 5177-5188

47. Tetreau C, Di Primo C, Lange R, Tourbez H, Lavalette D, Biochemistry (1997),

36, 10262-10275

48. Tetreau, C., Blouquit Y., Novikov E., Quiniou E., and Lavalette D. Biophys J.

(2004), 86, 435-447

49. Blobel, J., Schmidl, S., Vidal, D., Nisius, L., Bernadó, P., Millet, O., Brunner, E.,

Pons, M., J. Am. Chem. Soc. (2007), 129, 5946-5953

50. Lowery TJ, Doucleff M, Ruiz EJ, Rubin SM, Pines A, Wemmer DE, Protein

Science (2005), 14, 848-855



52. Landon C, Berthault P, Vovelle F, Desvaux H, Protein Science, (2001), 10, 762-

770

53. SM Rubin, SY Lee, EJ Ruiz, A Pines, DE Wemmer, J. Mol. Biol. (2002), 322,

425-440

54. Gröger C, Möglich A, Pons M, Koch B, Hengstenberg W, Kalbitzer HR, Brunner

E, J. Am. Chem. Soc. (2003), 125, 8726-8727

55. TJ Lowery, SM Rubin, EJ Ruiz, A Pines, DE Wemmer, Angew Chem Int Ed

(2004), 116, 2-4

56. A Mayer, S Ogawa, RG Shulman, T Yamane, JAS Cavaleiro, AM dA Rocha

Gonsalves, GW Kenner, KM Smith, J. Mol Biol (1974), 86, 749-756


44

57. RG Shulman, J Peisach, BJ Wyluda, J Mol Biol (1970), 48, 517-523

58. GN La Mar, FA Walker, J. Am. Chem. Soc. (1973), 95(6), 1782-1790

59. MP Byrn, BA Katz, NL Keder, KR Levan, CJ Magurany, KM Miller, JW Pritt,

CE Strouse, J. Am. Chem. Soc., (1983), 105(15), 4916-4922

60. RG Shulman, SH Glarum, M Karplus, J. Mol. Biol. (1971), 57, 93-115

61. Wu, Y.; Chien, E. Y. T.; Sligar, S. G.; La Mar, G. N. Biochemistry, (1998),

37(19), 6979-6990

62. K Rajarathnam, GN La Mar, ML Chiu, SG Sligar, J. Am. Chem. Soc. (1992),

114(23), 9048-9058

63. Nguyen, B. D.; Xia, Z.; Yeh, D. C.; Vyas, K.; Deaguero, H.; La Mar, G. N., J.

Am. Chem. Soc, (1999), 121(1), 208-217

64. SM Janes, D Mu, D Wemmer, AJ Smith, S Kaur, D Maltby, AL Burlingame, and

JP Klinman, Science (1990), 248, 981-987

65. Parson MR, Convery MA, Wilmot CM, Yadav KDS, Blakeley V, Corner AS,

Phillips SEV, McPherson MJ & Knowles PF, Structure (1995), 3, 1171–1184

66. Kumar V, Dooley DM, Freeman HC, Mithchell Guss J, Harvey I, McGuirl MA,

Wilce MCJ & Zubak VM, Structure (1996), 4, 943–955

67. Wilce MCJ, Dooley DM, Freeman HC, Mitchell Guss J, Matsunami H, McIntire

WS, Ruggiero HC, Tanizawa K & Yamaguchi H, Biochemistry (1997), 36,

16116–16133

68. Li R, Klinman JP, Scott Mathews F, Structure (1998), 6, 293–307

69. Duff AP, Cohen AE, Ellis PJ, Kuchar JA, Langley DB, Shepard EM, Dooley DM,

Freeman HC & Mitchell Guss J, Biochemistry (2003), 42, 15148–15157

70. Lunelli M, Di Paolo ML, Biadene M, Calderone V, Battistutta R, Scarpa M, Rigo

A & Zanotti G, J Mol Biol (2005), 346, 991–1004

71. Duff, A. P., Trambaiolo, D. M., Cohen, A. E., Ellis, P. J., Juda, G. A., Shepard, E.

M., Langley, D. B., Dooley, D. M., Freeman, H. C., and Guss, J. M., J. Mol. Biol.

(2004), 344, 599–607

72. Floris G and Finazzi Agro’ A, Amine Oxidases. In Encyclopedia Biological

Chemistry (Lennarz and Lane eds) Vol 1, 85-89, Academic Press Inc. New York,

NY, 2004


45

73. Dove, J. E., and Klinman, J. P. Adv. Protein Chem. (2001), 58, 141–174

74. Dooley, D. M., Scott, R. A., Knowles, P. F., Colangelo, C. M., McGuirl, M. A.,

and Brown, D. E. J. Am. Chem. Soc. (1998), 120, 2599–2605

75. Hirota, S., Iwamoto, T., Kishishita, S., Okajima, T., Yamauchi, O., and Tanizawa,

K. Biochemistry (2001), 40, 15789–15796

76. Padiglia, A., Medda, R., Lorrai, A., Paci, M., Pedersen, J. Z., Boffi, A., Bellelli,

A., Agro, A. F., and Floris, G. Eur. J. Biochem. (2001), 268, 4686–4697

77. Lewis, Tolman Chem Rev (2004), 104, 1047-1076

78. Medda R, Padiglia A, Bellelli A, Pedersen JR, Finazzi Agro’ A, Floris FEBS Lett

(1999), 453, 1-5

79. M. Mure, Acc. Chem. Res. (2004), 37, 131–139

80. E. Agostinelli, F. Belli, L. Dalla Vedova, S. Longu, A. Mura, G. Floris, Eur. J.

Inor. Chem. (2005), 9, 1635–1641 and references therein

81. Dooley DM, McGuirl MA, Brown DE, Turowski PN, McIntire WS, Knowles PF,

Nature (1991), 349, 262-264

82. Medda R, Padiglia A, Pedersen JZ, Rotilio G, Finazzi Agro’ A, Floris G,

Biochemistry (1995), 34, 16375-16381

83. DM. Dooley, WS. McIntire, J M A. McCuirl, CE. Cote, and JL. Bates, J. Am.

Chem. Soc. (1990), 112(7), 2782-2789

84. Medda, R., Padiglia, A., Bellelli, A., Sarti, P., Santanche’, S., Finazzi Agro’ , A.

and Floris, G. Biochem. J. (1998), 332, 431-437

85. Klinman, J.P. and Mu, D. Annu. Rev. Biochem. (1994), 63, 299-344

86. Su, Q. and Klinman, J.P. Biochemistry (1998), 37, 12513-12525

87. Scott, E. E., Gibson, Q. H., and Olson, J. S. J. Biol. Chem. (2001), 276, 5177–

5188

88. Carroll, F. A.; Quina, F. H. J. Am. Chem. Soc. (1976), 98, 1

89. Morgan, M. A.; Pimentel, G. C. J. Phys. Chem. (1989), 93, 3056

90. Maeda, M.; Graf, U.; Niikura, H.; Okamoto, M.; Hirayama, S. Chem. Phys. Lett.

(1996), 257, 175

91. Tanaka, F.; Hirayama, S.; Shobatake, K. Chem. Phys. Lett. (1992), 195, 243

92. Kim, T.-S.; Choi, Y. S.; Yoshihara, K. Chem. Phys. Lett. (1995), 247, 541


46

93. Anderson, M. A.; Grissom, C. B. J. Am. Chem. Soc. (1996), 118, 9552

94. MA Anderson; Y Xu, CB Grissom J. Am. Chem. Soc. (2001), 123, 6720-6721

95. Tedesco, C.; Immediata, I.; Gregoli, L.; Vitagliano, L.; Immirzi, A.; Neri, P.

CrystEngComm (2005), 7, 449-453

96. Brunet, P.; Demers, E.; Maris, T.; Enright, G. D.; Wuest, J. D. Angew. Chem.,

Int. Ed. (2003), 42, 5303-5306

97. Soldatov, D. V.; Moudrakovski, I. L.; Grachev, E. V., Ripmeester J. A. J. Am.

Chem. Soc. (2006), 128, 6737-6744

98. Lipkowski, J.; Soldatov, D. V. J. Inclusion Phenom. (1994), 18, 317-329

99. Soldatov, D. V.; Ripmeester, J. A.; Shergina, S. I.; Sokolov, I. E.; Zanina, A. S.;

Gromilov, S. A.; Dyadin, Yu. A. J. Am. Chem. Soc. (1999), 121, 4179-4188

100. Soldatov, D. V.; Grachev, E. V.; Ripmeester, J. A. Cryst. Growth Des. (2002),

2, 401-408

101. Margolin, A. L.; Navia, M. A. Angew. Chem., Int. Ed. (2001), 40, 2204- 2222;

102. Soldatov, D.V.; Moudrakovski, I. L.; Ripmeester, J. A. Angew. Chem., Int. Ed.

(2004), 43, 6308-6311

103. Malek, K.; Odijk, T.; Coppens, M. O. Nanotechnology (2005), 16, S522- S530

104. Smith, R. D.; Hu, L.; Falkner, J. A.; Benson, M. L.; Nerothin, J. P.; Carlson, H.

A. J. Mol. Graphics Modell. (2006), 24, 414-425, and refs therein

105. Karle IL, Handa BK, Hassall CH, Acta Cryst (1975), B31, 555-560

106. MR Ghadiri, JR. Granja, RA. Milligan, DE. McRee, N Khazanovich Nature

(1993), 366, 324-327

107. Soldatov, D.V.; Ripmeester, J.A. Stud. Surf. Sci. Catal. (2005), 156, 37-54

108. Bong D.T.; Clark T.D.; Granja J.R. and Ghadiri M.R. Angew. Chem. Int. Ed.

(2001), 40, 988-1011

109. Görbitz CH and Gundersen E, Acta Cryst (1996), C52, 1764-1767

110. Gorbitz CH, Acta Cryst. (2002), B58, 849-854

111. Gorbitz CH, Nilsen, M.; Szeto, K.; Tangen, L.W. Chem. Commun. (2005),

4288-4290

112. Moudrakovski, I.; Soldatov, D.V.; Ripmeester, J.A.; Sears, D.N.; Jameson, C.J.

Proc Natl Acad Sci USA (2004), 101, 17924-17929


47

113. Soldatov DV, Ripmeester JA, “Organic zeolites” in Nanoporous materials IV,

Stud in Surf sci Catal (2005), 156, 37-54

114. C.-Y. Cheng and C.R. Bowers, (2007), 8(14), 2077-2081

115. Ripmeester, J.A.; Ratcliffe, C.I, J. Phys. Chem. (1990), 94, 7652-7656

116. Cheung, T.T.P., J. Phys. Chem. (1995), 99, 7089-7095

117. Terskikh VV, Moudrakovski IL, Mastikhin VM, J. Chem. Soc. Faraday Trans.

(1993), 89, 4239-4243

118. Terskikh VV, Moudrakovski IL, Breeze SR, Lang S, Ratcliffe CI, Ripmeester

JA, Sayari A, Langmuir, (2002), 18, 5653-5656

119. Bonardet, J., Fraissard J., Gedeon A., Sprinuel-Huet M, Catal Rev-Sci. Eng.

(1999), 41(2), 115-225

120. Ripmeester JA, Ratcliffe CI, Tse, J. Chem. Soc. Faraday Trans. (1988), 84(11),

3731-3745

121. Ueda T, Eguchi T, Nakamura, Wasylishen R, J. Phys. Chem. B, (2003), 107,

180-185

122. Kato N., Ueda T., Omi H., Miyakubo K., Eguchi T., Phys Chem Chem Phys

(2004), 6, 5427-5434

123. Hiejima Y., Kanakubo M., Minami K., Aizawa T., Nanjo H., Ikushima Y., e-J.

Surf. Sci. Nanotech. (2005), 3, 338-340

124. Sozzani P., Bracco S., Comotti A., Mauri M., Simonutti R., Valsesia P., Chem

Commun. (2006), 1921-1923

125. Romanenko, Py, d’Espinose de Lacaillerie, Lapina, Fraissard J. Phys. Chem. B,

(2006), 110, 3055-3060

126. Demarquay J, Fraissard J, Chem. Phys. Lett. (1987), 136(3,4), 314-318

127. Ripmeester JA, Ratcliffe CI, Tse, J. Chem. Soc. Faraday Trans. (1988), 84(11),

3731-3745

128. Jameson CJ, De Dios A, J Chem Phys (2002), 116(9), 3805

129. Barrie P.J., Klinowski J., Prog. NMR Spectrosc. (1992), 24, 91

CHAPTER III – RESULTS AND DISCUSSION

48

Chapter III

Results and discussion


49

3.1 Myoglobins

Clearly, hints on the characterization of cavities and Xe-protein affinity can be

achieved by comparatively studying myoglobins extracted from different species or

mutated myoglobins. For instance, internal structural differences between myoglobins

extracted and purified from different animal species can be highlighted by exploiting the

sensitivity of Xe to its local environment.

Four myoglobins, from horse, rabbit, pig and sheep hearts are described here.

They have been chosen because they differ for several residues, mostly located at the

protein surface and only one residue differentiates their hydrophobic interior. Namely,

the residue in position 142, which lines the proximal cavity (Xe1), is an Isoleucine in

horse and rabbit Mbs, whilst pig and sheep Mbs have a Methionine at the same position.

This substitution introduces changes in both the size and shape of proximal cavity and in

its hydrophobic character: Methionine, indeed, is not as highly hydrophobic as Isoleucine

(hydrophaty indexes: Ile = 4.5, Met = 1.9)1 and additionally it is larger in size.

Figure 3.1: Molecular structure of amino acids methionine and isoleucine

It was shown by means of crystal structures of myoglobins from sperm whale in

presence of 7 atm of Xenon that the proximal cavity is the main binding site for Xenon

inside the protein and that only another site is populated at any time other than Xe1 at 7


50

atm of Xe.2 Moreover, the presence of both specific (Xe-cavities) and non-specific

(between Xe and protein surface and/or other environments that are sampled by Xenon

with weak diffusion-mediated interactions) interactions between Xenon and myoglobins

has been previously demonstrated and the effect of non-specific binding with the protein

surface on 129Xe NMR parameters was quantitatively assessed3-6. Equilibrium constants

for Xenon horse myoglobin, metmyoglobin and cyanomyoglobin complexes were also

derived from Xenon adsorption measurements and it was suggested that in metmyoglobin

two suitable binding sites exist for Xenon and that equilibrium constants vary, depending

on temperature, between 85 and 200 M-1 for one cavity and between 1 and 10 M-1 for the

other7. It was suggested afterwards by Locci and coworkers4, by analyzing 129Xe NMR

linewidths in myoglobins of high spin state (Fe3+ S=5/2), that Xe1 is, at first

approximation, the main binding site in horse Mb as it was previously observed in crystal

structures of sperm whale Mb-Xe complexes2. This is very reasonable if one considers

that the amino-acid compositions of the internal cavities in the two proteins are identical.

All these results suggested that proximal cavity in sperm whale and horse

myoglobin combines proper size, shape and chemical composition that make of Xe1 the

cavity for which Xenon atoms have the highest affinity compared to the other (at least

three) internal cavities, and that also surface residues can influence observed 129Xe NMR

parameters.

Additionally, the previous achievements just summarized led to speculate that

structure and amino acid composition of the proximal cavity within myoglobins could

represent crucial factors in determining Xe binding to myoglobin and that they could

have particular relevance in regulating ligand diffusion within the protein, i.e. protein

functionality.

It was successively shown by Corda et al8 that careful analysis of 129Xe NMR

chemical shifts and relaxation rates in aqueous solutions of high-spin Met-myoglobins

from pig can provide useful insight on the distribution of Xe within the internal cavities

of Mbs. In particular it was found that while Xe1 is the main binding site in horse Mb, at

least another cavity was suggested to significantly influence Xe binding to pig Mb8. The

same study also evidenced that the 13 residues that differentiate the protein surfaces of

pig and horse Mbs do not specifically influence observed 129Xe NMR parameters,


51

suggesting that the interaction of Xenon with the surfaces of the two proteins is almost

identical.

In order to more clearly highlight structural differences between proximal cavities

of myoglobins from horse and pig, some residues lining the proximal cavity in horse and

rabbit Mbs (a) and in pig and sheep Mbs (b) are shown in Figure 3.2.

(a) (b)

Figure 3.2: Residues lining the Xe1 cavity in (a) horse (Leu89, Ala90, His93, Leu104, Phe138, Ile142,

Heme) and (b) pig (Leu89, Ala90, His93, Leu104, Phe138, Met142, Heme) Mbs. The figures are obtained

with the visualization program VMD9.


52

3.1.1 - 129Xe NMR measurements in solutions of low-spin (Fe3+ S=1/2) cyano-

metmyoglobins.

Figure 3.3 collects the 129Xe NMR spectra of solutions of pig, sheep, horse and

rabbit CNMbs. Spectra of ∼1mM solutions of the four myoglobins, pressurized with 3

atm of Xenon overpressure in high-pressure tubes, were acquired at 25°C. Chemical

shifts are referenced to the chemical shift of 129Xe dissolved in buffer solution, in the

same experimental conditions.

-50510129129129129Xe (ppm )Xe (ppm )Xe (ppm )Xe (ppm )

sheepsheepsheepsheep

pigpigpigpig

horsehorsehorsehorse

rabbitrabbitrabbitrabbit

Figure 3.3: 129Xe NMR spectra of Xenon (3 atm of overpressure, 25°C) in ~1 mM solution (phosphate

buffer 0.1 M, 20% D2O) of cyano-metmyoglobins from horse, rabbit, sheep and pig.

Very similar NMR parameters, i.e. isotropic chemical shifts (δiso) and linewidths

(fwhm), characterize 129Xe signals of Xenon in solutions of CNMbs from pig and sheep

(pig : CNMb δiso=3.7 ppm, fwhm=46 Hz; sheep : δiso=3.3 ppm, fwhm=36 Hz).

Analogously, 129Xe spectra of horse and rabbit CNMbs exhibit essentially

identical chemical shifts and line widths: δiso=1.4 ppm and fwhm=55 Hz in horse CNMb

solution and δiso=1.2 ppm and fwhm=50 Hz for rabbit CNMb solution.


53

It may be useful at this point to recall again that these four myoglobins differ for

several residues mostly located at the surface of the proteins and that they can be grouped

into two groups according to the only difference concerning the amino acid composition

of internal cavities: in particular, horse and rabbit Mbs have in position 142 an Isoleucine

whereas pig and sheep Mbs have a methionine at the same position. Clearly, 129Xe NMR

is able to evidence this difference, also indicating that 129Xe NMR parameters are not as

much influenced by non-specific interactions with the protein surface as they are affected

by specific interactions with internal cavities.

These preliminary results were in fact not unexpected. It was previously shown by

Locci et al4 that 129Xe NMR chemical shift in horse high-spin metmyoglobins is affected

by two concurrent opposite contributions with respect to 129Xe resonance in water: a

downfield contribution, attributed to non-specific interactions between Xenon and the

protein surface, and an upfield contribution coming from the hyperfine dipolar interaction

between the Xenon bound within the proximal cavity and the high-spin paramagnetic

Fe3+.

Moreover, it was pointed out by Corda et al8 that a large diamagnetic component

characterizes the observed 129Xe signal in pig metmyoglobins (MMbs) compared to what

is observed in horse MMbs; however, it is difficult to separate the single contributions

and assign them to particular local sites, due to the dynamically averaged nature of 129Xe

NMR measurements. This latter work additionally demonstrated that the contributions to

the observed 129Xe chemical shift from non-specific interactions between Xenon and

metmyoglobins from horse and pig are very similar while the contribution due to specific

interactions Xe-cavity are quite different, showing a higher affinity of Xenon towards

horse MMbs.

In order to more deeply investigate the influence of proximal cavity structure and

composition on the affinity of Xenon to myoglobins, we present here the changes in 129Xe NMR chemical shift observed in solutions of myoglobins from horse, rabbit, pig

and sheep as a function of increasing Xe concentration (overpressure).

In Fig. 3.4 the variation of the observed 129Xe chemical shift is reported as a function of

the ratio nXe/Vl between the total number of moles of Xenon in the NMR tube, nXe, and

the volume of the CNMb solution, Vl.


54

Figure 3.4 clearly shows the sensitivity of 129Xe chemical shift to the nature of

proximal sites in the different myoglobins analyzed. In particular, ∆δ 129Xe, i.e. chemical

shift changes resulting from increasing Xenon concentration in the four CNMb solutions,

group the four myoglobins into two well distinct classes: noticeably, horse and rabbit

CNMb show different behavior than pig and sheep CNMbs, certainly reflecting different

interactions of these two groups of CNMbs with Xenon. In horse and rabbit CNMb

solutions, the 129Xe chemical shift moves downfield with increasing Xenon

concentration, while a reverse trend is observed in pig and sheep CNMbs. For the highest

concentrations of Xenon investigated the 129Xe ∆δ seem to asymptotically converge to

the same value for all the four CNMb solutions.

1

1.5

2

2.5

3

3.5

4

0 0.5 1 1.5 2

nXe/Vl (M)

∆δ∆δ ∆δ∆δ 12

9 Xe

(ppm

)

Pig

Sheep

Rabbit

Horse

Figure 3.4: Variation (in ppm) of the observed 129Xe chemical shift (∆δ 129Xe) as a function of the total

number of moles of Xenon in the NMR tube divided by the volume of solution (nXe/Vl), in ~1 mM

solutions of cyano-metmyoglobins from pig, sheep, rabbit and horse. Shifts are referenced to the chemical

shift of 129Xe dissolved in buffer.


55

In order to characterize the binding of Xenon to the CNMbs in aqueous solution

the observed 129Xe chemical shift variation with Xenon concentration was analysed

adopting a two-site model4,8.

inout XeXeCNMb ↔+ [3.1]

According to this model Xenon exchanges rapidly between specific sites within

the interior of the protein (Xein) and all the remaining possible environments (Xeout),

mainly represented by the solvent and the protein surface. The process is characterized by

an equilibrium binding constant K

][][

][

CNMbXe

XeK

out

in= [3.2]

Under these conditions, the observed chemical shift δobs is given by the following

equation:

l

inoutinoutobs Xe

Xe

][

][)( ⋅−+= δδδδ [3.3]

Here δin is the chemical shift of 129Xe trapped inside the protein, δout the chemical

shift of 129Xe in the other possible environments, [Xe]in the concentration of the

complexed Xenon (Xe bound to Mb), and [Xe]l the total concentration of Xenon

dissolved in solution.

Within the framework of this thermodynamic model [Xe]in can be expressed as a

function of K, and thus the variation of δobs with (nXe/Vl) is a function of δin, δout and K.

For numerical application few reasonable approximations can be introduced without

significantly alter the results: it is in fact assumed that Xenon is an ideal gas, that Henry’s

law holds, and that [Xe]out corresponds to the total concentration of Xenon dissolved in

the buffer solution, i.e. [Xe]out = [Xe]l. Due to the low protein concentration used in the


56

experiment and the expected weak binding constant these approximations seem

reasonable.

The just described model can be exploited to fit the experimental variations of 129Xe NMR chemical shifts as a function of nXe/Vl in Figure 3.4 and obtain the values of

δin, δout and K. Table 3.1 collects the results of the fitting procedure.

Table 3.1: Pig, sheep, horse and rabbit CNMb equilibrium constants (K) and 129Xe chemical shifts (δin, δout)

obtained from fitting the thermodynamic model to the experimental data. The errors are fitting errors.

[CNMb] (mM) K (M -1) δδδδin (ppm) δδδδout (ppm)

Horse (1.01)

Rabbit (1.03)

Pig (0.94)

Sheep (0.90)

158±40

131±22

48±8

36±6

-18±4

-22±3

55±10

51±9

2.7±0.2

2.7±0.2

2.1±0.1

2.2±0.1

Within the experimental errors, the binding constants K relative to the complexes

of Xenon with the horse (158 M-1) and rabbit (131 M-1) CNMbs are in good agreement

with that reported in the literature7 for the sperm whale CNMb (145 M-1). The

comparison between binding constants and chemical shifts that characterize bound and

unbound Xenon species in the four myoglobins analyzed seem to group them into two

well defined classes: pig CNMb- sheep CNMb on one side and horse CNMb – rabbit

CNMb on the other, suggesting that differences at the level of the interior structure of

myoglobins are significant in determining 129Xe NMR results. This classification of the

four myoglobins considered into two well distinct groups is further confirmed by the

value of the chemical shift (δin) associated to the 129Xe bound within the cavities of the

proteins, obtained by means of the thermodynamic model applied to fit the experimental

results. δin has in fact very similar negative values for 129Xe bound to horse (δin =-18

ppm) and rabbit (δin = -22 ppm) CNMbs, while positive values characterize Xe-sheep

CNMb (δin = 51 ppm) and Xe-pig CNMb (δin = 55 ppm) complexes. Thus, the two

groups of myoglobins differ in δin of about 70 ppm. Considering, as previously stated,

that the distance-dependent hyperfine dipolar interaction between Xenon and


57

paramagnetic heme Iron causes an upfield shift of the observed 129Xe signal we could

suggest that Xenon is much more influenced by the paramagnetic ion in myoglobins from

horse and rabbit than it is in myoglobins from pig and sheep. Clearly, 129Xe NMR

resonances are mirroring differences in the Xenon binding inside the proteins. In

particular, these observations likely originate from a higher affinity of Xenon toward the

proximal site of myoglobins from horse and rabbit. This hypothesis is further confirmed

by the binding constants K extracted for the four myoglobins: the equilibrium constant K

measured in the horse and the rabbit are about three times larger than in the pig and the

sheep Table 3.1. These differences can be ascribed either to changes in the size, shape

and residue composition of the cavities affecting the chemical shift of the bounded 129Xe

and the binding constant or to residues at the surface of the proteins, which differentiate

the four myoglobins analyzed.

The values of δout of Table 3.1 can be essentially attributed to non-specific

interactions of Xenon with the protein surface and Xenon in the bulk4,6. Remarkably, the

∆δout, normalized to 1mM protein, in the CNMbs from pig (2.2 ppm), sheep (2.4 ppm),

horse (2.6 ppm) and rabbit (2.6 ppm) are very alike, suggesting similar interaction

schemes of Xenon with the external surface of the proteins. All these data are in good

agreement with those found for pig (2.41 ppm) and horse (2.35 ppm) MMb4,8, reinforcing

the idea that an interaction of Xenon with the external surface of the protein can be

considered to a good approximation independent from the particular surface composition

of the studied species.

In order to more clearly define the critical factors that characterize the chemical shifts

of 129Xe observed in solutions of the different Mbs considered in this work, we need to

describe in more detail the nature of specific Xenon-protein interactions in paramagnetic

proteins.

The NMR chemical shift of 129Xe bound to the protein, δin, is the sum of three

different contributions: diamagnetic, pseudocontact (or dipolar) and Fermi contact

contributions.

δin = δdia + δdip + δcon [3.4]


58

Here, δdia is due to the interactions of Xenon with the residues lining the cavity where

it is trapped in, and depends on cavity structure and shape10-13. δdip is related to the

hyperfine dipolar interactions between the Xenon and the low spin paramagnetic Fe3+ of

the prosthetic heme group and δcon is the contribution of the unpaired electron spin

density of the paramagnetic ion delocalized at the observed nucleus. When non-

covalently bound Xenon is considered, in fast exchange with the bulk solution, δcon is

zero.

Analogously, δdip is zero when the heme iron is in the oxidation state 2+ and has

spin state S=0 (no unpaired electrons). This latter situation arises when diamagnetic

carboxy myoglobin (CO-Mb) species are considered and experiments on this species

allow obtaining estimates of the diamagnetic contribution to the 129Xe NMR chemical

shift.

The high information content of low-spin cyano-myoglobins is due, among other

reasons, to the presence of substantial magnetic anisotropy that imposes significant

dipolar shift to nuclei near the active site. This property is very useful especially because

it allows the exploitation of the iron paramagnetism to probe the geometry of nonbonded

nuclei (residues and/or ligands).

In details, δdip in CNMbs is determined by the anisotropic magnetic moment of

the iron atom and can be written as a function of the polar coordinates (R, ϑ, ϕ) of the

targeted Xenon with respect to a coordinate system in which the magnetic susceptibility χ

is diagonal. The resulting equation is14:

∆+∆−= ),,(2

3),,(

3

1RFRF

N rhrhaxaxdip ϕϑχϕϑχδ [3.5]

with axial and rhombic magnetic anisotropies of the magnetic susceptibility tensor χ,

respectively

∆χax = χzz – ½( χxx+ χyy)

∆χrh = χxx – χyy


59

geometric factors

3

2 )cos3(

RFax

ϑ= and 3

2 2cossin

RFrh

ϕϑ=

N is Avogadro’s number and (R, ϑ, ϕ) are the polar coordinates of the targeted

nuclei with respect to a Fe-centred coordinate system.

Equation [3.5] therefore becomes:

3

22rh

3

2ax 2cossin

2

1cos3

3 RNRNdip

ϕϑχϑχδ ∆−−∆−= [3.6]

The magnetic coordinate system adopted following Nguyen et al.14 has the

magnetic z-axis passing through the iron atom and the His93 Cα atom, tilted ~15° from

the heme normal toward the heme δ-meso position. Also the values of ∆χax and ∆χrh, 2.54

x 10-8 m3/mole and -0.62 x 10-8 m3/mole, respectively, are taken from the work on sperm

whale CNMb by Nguyen et al.14. This latter assumption is justified by the similar heme

geometry and stereochemistry of all these CNMbs15.

It is usually adopted, in order to localize nuclei within myoglobins, a coordinate

system in which the tensor χ is diagonal. Thus, the atomic coordinates (x,y,z) extracted

by X-ray crystallography are replaced by new coordinates (x’,y’,z’) of the new coordinate

system. If Γ(α,β,γ) represents the Euler rotation matrix transforming the X-ray derived

coordinates system defined above to one in which the magnetic susceptibility tensor χ is

diagonal (with components χxx, χyy, χzz), equation [3.6] becomes

),,(),,(2

3),,(

3

1 γβαϕϑχϕϑχδ Γ⋅

∆+∆−= RFRFN rhrhaxaxdip [3.7]

The Euler rotation matrix is defined by the angles α,β, and γ, according to the scheme

below.


60

Figure 3.5: Schematic representation of the Euler rotation angles α, β and γ which transform the coordinate

system x,y,z in a new system referred to as ‘new x’, ‘new y’, ‘new z’.

Equation [3.7] shows that the chemical shift experienced by each nucleus depends

on its position with respect to the iron, on the magnitude of the anisotropies ∆χax and

∆χrh, which in turn relate to the electronic structure of the heme iron, and on the

orientation of the magnetic susceptibility tensor.

Ideally, if the position of Xenon atoms with respect to the paramagnetic center

Fe3+ were known, we could obtain the dipolar contribution to the observed 129Xe

chemical shift by just applying equation [3.7]. However, due to the fast exchange of

Xenon in the NMR times scales, this calculation is in fact unfeasible. Nevertheless,

knowledge of the geometric factors which characterize the four internal cavities of

myoglobins furnishes useful indication of what cavities could reasonably participate to

the binding process in Xe-myoglobin complexes.

The polar coordinates of the Xenon bound to the cavities Xe1 (R = 5.30 Å; θ =

35.75; φ = 72.20), Xe2 (R = 9.50 Å; θ = 77.14; φ = 70.85), Xe3 (R = 15.09 Å; θ =

62.69; φ = 83.58) and Xe4 (R = 8.46 Å; θ = 44.98; φ = 76.60) can be extracted by

analysing of the crystallographic structure of sperm whale myoglobin (pdb 1J52). In

particular, as it has been already stated8, according to equation [3.7] the 129Xe shift would

shift upfield for Xe1 and Xe4, and downfield for Xe2 and Xe3 relative to the chemical

shift of 129Xe shift in buffer.


61

It is worth noting that by using different crystallographic structures (pdb 1MBC,

1YMB and 1MYG) of myoglobins from horse and sperm whale found in literature,

computing the location of the centre of gravity (CG) of the cavity by exploiting the

program VOIDOO (see Materials and Methods’ section) and following the procedure

outlined above we obtained similar results. Moreover, the calculation of paramagnetic

contribution to the observed 129Xe shift from geometric constraints by means of equation

[3.7] suggest that, as it was in fact expected, δdip has decreasing value in the series Xe1

>> Xe2 ≥ Xe4 >> Xe3.

As it is shown in Figure 3.3, in horse CNMb and rabbit CNMb 129Xe signal is

shifted to higher fields than in solutions containing CNMbs from sheep and pig.

However, all myoglobins induce an upfield shift to the 129Xe resonances even if to

slightly different extents.

Moreover, the values estimated for δin in horse and rabbit CNMbs via the two-site

thermodynamic model are upfield shifted with respect to Xenon in the buffer. These

observations, together with the calculations just mentioned would suggest that the

chemical shift of 129Xe is dominated by the dipolar interaction between Xenon and the

unpaired electron of low spin paramagnetic Fe3+ and typify the Xenon binding in the

proximal cavity, in agreement with crystallographic studies on structurally similar sperm

whale Mb2.

In pig and sheep CNMbs δin is downfield shifted, although it results upfield

shifted compared to Xenon binding in the diamagnetic COMbs of the same species. This

latter paramagnetic shift has been explained by the presence of Xenon in the Xe1 cavity

and potentially in the Xe4 cavity of the pig CNMb8. The fast exchange regime of Xenon

in the NMR chemical shift time scale does not allow discriminating certainly between the

paramagnetic contributions from these two sites.


62

Table 3.2:* Inaccessible Cavities computed using the program VOIDOO for horse MMb (PDB code =

1YMB) and pig MMb (PDB code = 1MYG)

Cavities V a

Ra θθθθa

GFb Lining

Horse Xe1 43.7 5.3 42.9 40.9 Leu89, Ala90, His93, Leu104, Phe138, Ile142,

Heme Xe2 48.9 9.4 84.5 -11.7 Leu72, Leu104, Ile107, Ser108, Ile111, Leu135,

Phe138, Arg 139, Heme Xe3 59.9 15.6 74.1 -2.6 Trp7, Ile75, Leu76, Lys78, Lys79, Gly80,

His82, Ala134, Leu137, Phe138 Xe4 19.7 8.6 147.4 17.7 Gly25, Val28, Leu29, Gly65, Val68, Leu69

Pig Xe1 30.3 4.9 49.7 21.7 Leu89, His93, Leu104, Phe138, Met142, Heme

Xe2 23.4 9.6 92.8 -11.2 Leu72, Leu104, Ile107, Ser108, Ile111, Leu135, Phe138, Heme

Xe3 46.6 15.7 73.9 -2.0 Trp7, Ile75, Leu76, Lys78, Lys79, Gly80, His82, Ala134, Leu137, Phe138

Xe4 25.8 8.4 147.3 18.9 Gly25, Val28, Leu29, Gly65, Val68, Leu69, Ile107

a) Volume (Å3), distance from centre of gravity and iron (Å) and angles (deg) obtained from VOIDOO. b)

Geometric Factor, (3cos2θ-1)R-3 · 1026 m3. *From ref 8

3.1.2. - 129Xe NMR relaxation measurements of CNMb solutions.

Longitudinal relaxation times T1 have been measured for ~ 1 mM solutions of

CNMbs extracted from pig, sheep, rabbit and horse. For each solution, a single T1 value

fitted the experimental data, indicating that Xenon is in fast exchange, in the 129Xe

relaxation time scale, between all available environments. The 129Xe relaxation rate R1 =

1/T1 measured for buffer solution pressurized with 2 atm of Xenon is R1(buff) ~ 2x10-3 s-

1. This value, as expected, is significantly affected by the presence of low-spin

paramagnetic cyano-metmyoglobins: in particular we measured, in the same experimental

conditions, values of R1(h)=0.95±0.07 s-1 for horse CNMb, R1(r)=0.96±0.10 s-1 for rabbit

CNMb, R1(p)=0.31±0.03 s-1 for pig CNMb and R1(s)=0.36±0.06 s-1 for sheep CNMb.

The increase of 129Xe longitudinal relaxation rates of Xenon in solutions

containing low-spin (Fe3+, S=1/2) paramagnetic myoglobins in the cyano form, however,

is less pronounced than that previously measured for 129Xe in solutions of high-spin


63

(Fe3+, S=5/2) Met-myoglobins (horse R1=11.4±0.1 s-1, pig R1=1.4±0.1 s-1),8 thus

substantiating the sensitivity of 129Xe NMR not only to the presence of the paramagnetic

metal ion but even to its spin state.

In order to more clearly understand and explain these experimental data we

should identify and quantitatively assess all the contributions that determine the observed 129Xe relaxation rates and in addition formulate appropriate considerations on the

experimental conditions and on the features that characterize Xe-myoglobin complexes.

It is evident that the presence of the paramagnetic ion (Fe3+, S=1/2 in CNMbs)

plays a fundamental role in determining the value of the observed spin-lattice relaxation

rate R1,obs. The contribute R1,p, merely due to hyperfine interactions between Xenon and

the unpaired electron of the paramagnetic heme iron, is superimposed to that (R1,0) given

by the diamagnetic interactions between Xenon and globins and between Xenon and the

buffer solution in which Xe and myoglobins are dissolved.

In presence of a large excess of free Xenon compared to Xenon bound to protein

cavities, as it is in our case, longitudinal relaxation rate can be expressed by the equation

[3.8]:

0,1,1,1 RRR pobs += [3.8]

Here, the observed spin-lattice 129Xe NMR relaxation rate R1,obs is given by the

sum of paramagnetic spin-lattice relaxation rate R1,p and relaxation rate R1,0, respectively

corresponding to the relaxation rate measured in presence and the absence of Fe3+

species.

Indicating τM the residence time of Xenon bound to Mb cavities, T1M the spin-

lattice relaxation time of the Xenon in the bound state, p the ratio between the total

concentration of metal ion and that of dissolved Xenon, and n the number of moles of

Xenon interacting with each paramagnetic site (i.e. with each myoglobin), equation [3.8]

becomes:

)( 10,1,1,1

MMobsp T

npRRR

τ+=−= [3.9]


64

Clearly, np represents the molar fraction of Xenon complexed inside the protein

cavities

l

in

l

in

Xe

Xe

Xe

Fe

Fe

Xenp

][

][

][

][

][

][=⋅= [3.10]

In the absence of contact interactions, as is the case of Xe-myoglobin complex,

the diffusion averaged nuclear relaxation of the weakly bound 129Xe is enhanced by the

motions of both the electron spin and of the entire complex relative to the direction of the

external magnetic field. Therefore, both the longitudinal (T1M-1) and the transverse (T2M

-1)

relaxation rates of the nuclei bound in the proximity of paramagnetic sites are affected by

these motions. When such complexes are considered, where non-bound Xe freely

diffuses through protein cavities, paramagnetic enhancement of nuclear relaxation of 129Xe is merely dominated by distance-dependent dipole-dipole interactions.

The unpaired spin density of Fe3+ electrons is not usually confined to the metal

ion, but it is delocalized over the heme group. It has been stated14,16, however, that the

just mentioned π delocalization is in fact negligible in low-spin Fe(III) complexes, so that

the effects of the paramagnetic ion can be reliably approximated by a point-dipole model

even for nuclei that are very close to the iron.

Equation [3.11] represents the Solomon equation that describes dipolar

longitudinal relaxation of 129Xe bound to myoglobin, in a fast exchange condition

between all the cavities17,18 and is given by:

++

++

=2c

2S

c2c

2Xe

c6

2B

2e

2Xe

2

0M1 τω1

τ7τω1

τ3r

µg)γ1S(Sπ4µ

152

/T1 [3.11]

where r is the distance Fe-Xe, S the electron spin quantum number (S=1/2 for the low

spin Fe3+ in CNMb), γXe the gyromagnetic ratio of the observed 129Xe nuclei, µB the Bohr

magneton, ge the electron g value, µ0 the vacuum permeability, ωS and ωXe the electronic


65

and nuclear (129Xe) angular Larmor frequencies in radiants per second at the operating

magnetic field, respectively19.

The correlation time (τc) is the reciprocal of a rate constant which value is

affected by different contributions, according to the following equation [3.12]:

τc-1 = τM

-1 + τr-1 + τs

-1 [3.12]

where τs-1is the electronic relaxation rate , τr

-1 is the contribution from molecular rotation,

and τM-1 takes into account the chemical exchange of the probe nucleus (Xe).

Equation [3.11] has been extensively used for determining the distance between a

paramagnetic metal ion and a probe nucleus in absence of significant trough-bond

interactions 5,20-22.

The rate constants that contribute to define the correlation time have been

previously estimated for several species and their values can be found in literature. The

electronic term τs ranges between 10-11 and 10-12 s for low spin Fe3+,23,24 while the

isotropic rotational correlation time for a molecule of the size of myoglobin ranges

between 10-8 and 10-9 s; τM should not be larger than 10-5 s.25 For low spin Fe3+ CNMb

the correlation time τs (~10-11-10-12 s) is expected to dominate τc. The value τc of 6 x 10-9

s was derived from the 15N T1 and T2 relaxation times measured for 25 α-helical residues

of a diamagnetic COMb sample16.

When 129Xe spin-lattice relaxation times T1 are measured in aqueous solutions

containing the four low-spin CNMbs from horse, rabbit, pig and sheep different

behaviors are observed and again horse CNMb and rabbit CNMb show similar behavior,

but different from that of pig CNMb and sheep CNMb.

In order to clearly define a picture of what is observed let us consider just CNMb

from horse and pig as representative species, and discuss T1 values obtained at variable

pressures of Xenon into tubes containing solutions of these two myoglobins.

Figure 3.6 shows T1 values measured at increasing Xenon pressure in the tubes

containing ~1 mM solution of CNMbs from pig and horse.


66

0

1

2

3

4

0 2 4 6 8 10 12

T1 (

sec)

atm Xenon

pig CNMb

horse CNMb

Figure 3.6: T1 values measured at increasing Xenon pressures in tubes containing solutions of CNMbs

from pig and horse

Clearly, an increase in the T1 of 129Xe corresponds to an increase of Xenon

pressure in the tube containing a solution of horse CNMbs, while a reverse trend is

observed for the solution of pig CNMb. It is interesting to point out again that CNMbs

from pig and sheep show similar behavior and the same can be said for the couple rabbit

CNMb and horse CNMb (data not shown). A detailed examination of the data obtained

can be carried out by exploiting the thermodynamic model adopted for the interpretation

of 129Xe chemical shifts. This model, in fact, allows discerning between interactions of

Xenon with the internal region of the proteins and all the other available environments.


67

0

0.2

0.4

0.6

0.8

1

1.2

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07

R1o

bs(s

-1)

[Xe]in

/[Xe]tot

pig

horse

rabbit

sheep

Figure 3.7: Variation of the observed 129Xe spin lattice relaxation rate as a function of the ratio of

concentration of Xenon complexed within internal cavities and total Xenon concentration in ~1 mM

solution ([Xe]in/[Xe] tot) of CNMbs from pig, sheep, horse and rabbit.

In Figure 3.7 are plotted the 129Xe R1obs corresponding to Xenon in horse, rabbit,

sheep and pig CNMbs as a function of the ratio [Xe]in/[Xe]tot.

This ratio can be estimated by using7 K = 145 M-1 for horse and rabbit, K = 48 M-1 for

pig and K = 36 M-1 for sheep, and the two-site thermodynamic model previously

described [see page 55].

While in horse and rabbit CNMb the 129Xe relaxation rate increases with

increasing concentration of Xenon bound to the internal cavities, a slight decrease is

observed for the 129Xe relaxation rate as a function of Xenon bound ([Xe]in/[Xe]tot) in pig

CNMb and sheep CNMbs.

An additional very important consideration must be done at this point as adopting

the variable [Xe]in/[Xe]tot may introduce misunderstandings in regard to its correlation

with Xenon pressure in the tube: it is important to note that due to the relatively low

concentration of myoglobins in solution, increasing Xenon pressure results in a decrease


68

of [Xe]in/[Xe]tot, as bound Xenon species ([Xe]in) have a very low concentration with

respect to non-bound Xe, while Xenon dissolves in solution in the amount of ∼4 mM/atm.

We can now hypothesize that the residence time of Xenon within the proximal cavity of

the four myoglobins analyzed is the same and, precisely, identical to that previously

measured25 (~ 10-5 s) for sperm whale met-myoglobin. Thus, 1/(T1M+τM) in equation [3.9]

reduces to 1/T1M and this value, which is the longitudinal relaxation rate of 129Xe bound

to the Mb, can be determined for the 129Xe in the four Mbs analyzed.

As expected, the extracted values are grouped in two blocks: horse-rabbit couple

together as well as pig-sheep do. More precisely, 1/(T1M) = 9.18 ± 0.28 s-1 and R1,0 = 0.45

± 0.01 s-1 for the horse CNMb, 1/(T1M) = 9.41 ± 0.71 s-1 and R1,0 = 0.39 ± 0.03 s-1 for the

rabbit CNMb, 1/(T1M) = -0.64 ± 0.24 s-1, R1,0 = 0.35 ± 0.01 s-1 for the pig and 1/(T1M) = -

0.29 ± 0.13 s-1, R1,0 = 0.36 ± 0.01 s-1 for the sheep.

The negative values of 1/(T1M) elicited for the latter group are, of course, non

physical, thus indicating that the starting hypothesis on the residence time of Xenon

within proximal cavity is not reliable.

The increasing 129Xe relaxation rate actually observed as a function of increasing

Xenon pressure in pig and sheep CNMbs (see Figure 3.7) suggests that the Xenon

population in the proximal cavity increases significantly as a function of Xenon pressure.

Because of the 1/r6 dependence in the relaxation rate (equation [3.11]), Xenon in the

proximal cavity, with a CG-Fe distance of 4.9 Å (as determined by the program

VOIDOO) in the pig MMb8, mainly contributes to the average relaxation rate, being the

other cavities 8.4, 9.7 and 15 Å far from the iron centre.

The unphysical values extracted from fitting the relaxation rates measured on pig

and sheep CNMbs might therefore arise from a different distribution of Xenon among

their internal cavities than that in horse and rabbit CNMbs. In the former pair of Mbs this

ratio does not favour the Xe1 as much as in the latter.

The results previously obtained by observing 129Xe NMR parameters4,5,8 together

with those discussed here clearly indicate important differences in the affinity of Xenon

towards internal cavities of myoglobins, they are infected by some problems, mostly

derived from the fast exchange condition of Xenon between all available environments.


69

Testing the ability of Xenon as an efficient biomolecular probe of cavities in

globular proteins and, particularly, in myoglobins (Mbs) is crucial to extract important

information on their structure and function, but at the same time challenging due to the

presence of different interaction sites besides the heme iron. 129Xe NMR measurements

so far performed have given therefore exchange-averaged results and lacked of the site-

specific information needed in order to more clearly understand the role played by

structure, shape and composition of proximal cavity in regulating Xenon affinity.

Among the techniques adopted in the characterization of myoglobins, NMR has

proven its suitability in investigating these model systems. The possibility to easily tune

the oxidation state of the metal ion (Fe) in myoglobins gives the unique opportunity to

study interactions of different physical nature in the same protein, which are sensed by

both the guest and the host itself.

In particular, for low-spin cyano myoglobins (CNMbs) most of 1H signals of the

residues in the active site have been unambiguously assigned26. These low-spin systems

have been the subject of extensive study because of the excellent resolution and narrow 1H NMR lines even for the protons close to the iron27-30. It is well known that both

contact and pseudocontact shifts contribute to the observed proton NMR chemical shifts

in paramagnetic species: a number of studies have been carried out so far aiming to

understand and quantitatively separate these two contributions in myoglobins and model

heme compounds22,31-33. Moreover, many efforts have been spent to determine the effect

of the orientation of the magnetic axes of the susceptibility tensor χ and of their

modifications in point mutants of CNMbs on the dipolar contribution to the observed

NMR chemical shifts14,34-39.

The extremely high information content of low-spin (S=1/2) CNMbs can be

exploited with the aim of deepening the role played by internal cavities in these proteins.

The presence of significant magnetic anisotropy imposes large dipolar shift to nonbonded

residues in the active site and allows exploiting iron paramagnetism to probe the

geometry of distal and proximal regions of the heme cavity, thus providing valuable

information on possible ligand-induced modifications of the host matrix. Most

importantly, 1H NMR measurements on myoglobin solutions pressurized with Xenon gas


70

allow obtaining site-specific information on the residues close to the iron, i.e. lining the

active site.

3.1.3. - 1H NMR chemical shift in CNMbs from horse and pig.

Figure 3.8 collects 1H NMR spectra of CNMbs from horse and pig and the effect

of 10 atm of Xenon on the chemical shifts of hyperfine-shifted protons. It is evident, even

at a first sight, that the presence of 10 atm Xenon overpressure causes marked changes in

the proton NMR spectra of CNMbs.

Further discussion is needed to point out important distinctions that must be made

to explain the results concerning different protons and myoglobins from different sources.

Tables 3.3 and 3.4 collect the 1H chemical shift of CNMb from horse and pig in

the absence and presence of 10 atm of Xenon, compared to the chemical shift of the

corresponding protons in sperm whale CNMb26.

1216202428 -10-8-6-4-20

Pig

Pig+Xe

Horse

Horse+Xe

5-C

H3

8-C

H3

1-C

H3

His

93 N

δH

His

93 C

εH2-

Hα

Phe

43 C

ζH

His

93 N

pH

Ile99

Cγ1

H

Ile99

CδH

3Ile99

Cγ2

H3

Val

68 C

αHThr

67 C

γH3

Leu1

04 C

δH3

1H (ppm)1H (ppm)

Figure 3.8: 1H NMR spectra in the low-field (from 11 ppm to 29 ppm) and high-field (from -10 ppm to 0

ppm) regions for ~1 mM solutions of CNMbs from horse and pig in the absence and presence of 10 atm of

Xenon overpressure. Labelling of some signals is made on NOESY spectra and based on the assignments

of Emerson and La Mar26.


71

Signal assignment is based on the published data on the sperm whale CNMb and

on two-dimensional spectra COSY, TOCSY and NOESY14,26. Some of the assigned

peaks of horse CNMb and pig CNMb are indicated in the mono-dimensional 1H NMR

spectra.

As observed for Sperm Whale CNMb26, 1H NMR spectra of Cyano-

MetMyoglobins from pig and horse show three methyl signals in the region 18-13 ppm,

assigned to 5-CH3, 1-CH3, and 8-CH3 heme groups. In the spectral region from 23 ppm to

9 ppm many one-proton signals are observed and attributed to α-type protons of the heme

substituents and to protons pertaining to the proximal histidine. Some resonances of other

residues present in the active site are also detected in this spectral region, due to the

magnetic anisotropy of the low-spin iron system. In the low-frequency region signals are

resolved in the -1 to -10 ppm range.


72

Table 3.3: 1H NMR chemical shifts relative to Heme and His93 protons observed in pig CNMb and horse

CNMb solutions in the absence and the presence of 10 atm of Xenon. The 1H chemical shifts of sperm

whale CNMb are reported for comparison26.

Chemical shift (ppm) Resonances Proton Sperm Whale Horse Horse/Xe Pig Pig/Xe

Heme 1-CH3 18.62 18.37 17.87 18.49 18.41 3-CH3 4.76 4.39 3.82 4.04 3.88 5-CH3 27.03 27.16 26.54 27.52 27.24 8-CH3 12.88 13.49 12.17 13.20 12.75 2-Hα 17.75 17.88 18.43 17.75 17.95 2-Hβc -1.73 -1.49 -1.76 -1.36 -1.52 2-Hβt -2.55 -2.42 -2.64 -2.37 -2.50 4-Hα 5.50 5.53 5.75 5.72 5.62 4-Hβc -1.95 -1.77 -1.24 -1.97 -1.73 4-Hβt -0.77 -0.59 -0.10 -0.81 -0.61 6-Hα 9.18 9.27 9.84 9.23 9.47 6-Hα’ 7.35 7.52 8.10 7.86 8.00 6-Hβ 1.67 1.58 1.78 1.26 1.39 6-Hβ’ -0.48 -0.41 -0.76 7-Hα 1.13 1.43 1.64 1.14 1.20 7-Hα’ -0.45 -0.33 -0.01 -0.35 -0.27 7-Hβ 1.55 1.49 1.32 1.33 1.41 7-Hβ’ 0.78 0.50 0.62 0.71 0.65 Hα 4.40 4.41 2.60 4.32 4.21 Hβ 2.09 2.38 2.94 2.55 2.69 Hγ 5.98 6.08 6.73 6.32 6.05 Hδ 4.40 4.13 3.79 3.59 3.79

His93 CαH 7.51 7.41 7.49 7.49 7.48 CβH 11.68 11.54 11.13 11.61 11.59 CβH’ 6.34 6.46 6.33 6.45 6.45 NpH 13.20 13.78 13.57 13.92 13.88 CδH -4.70 -4.80 -4.58 -5.05 -4.98 NδH 20.11 21.20 21.37 21.15 21.13 CεH 19.20 18.90 19.82 18.65 18.66


73

Table 3.4: 1H NMR chemical shifts of residues lining the active site of pig and horse CNMbs in the

absence and presence of 10 atm of Xenon in the NMR tube. The 1H chemical shifts of sperm whale CNMb

are also reported26.

Chemical shift (ppm) Residues Proton

Sperm Whale Horse Horse/Xe Pig Pig/Xe Leu29 Cδ2H 5.53 5.63 5.56

CγH 3.90 3.76 3.90 4.07 4.03 Phe33 CεH 8.32 8.36 8.33 8.31 8.27 Phe43 CεH 12.58 12.42 12.38 12.41 12.40

CζH 17.27 17.16 17.21 16.96 16.91 Phe46 CδH 7.69 8.23 8.19 His64 NpH 8.45 8.68 8.67

CδH 11.61 12.40 12.37 12.43 12.34 NεH 23.70 23.46 23.31 22.57 22.56

Val67 CαH 2.37 2.42 CβH 0.87 0.89 CγH3’ 0.09 0.10 CγH3 -1.93 -1.86

Thr67 CαH 2.47 2.52 2.51 CβH 2.66 2.79 2.78 CγH3 -1.59 -1.63 -1.64

Val68 CαH -2.55 -2.21 -1.78 -2.32 -2.19 CβH 1.42 2.03 2.08 2.00 Cγ2H3 -0.97 -0.61 -0.54 -0.53 -0.60 Cγ1H3 -0.81 -0.61 -0.54 -0.53 -0.60

Ala71 CαH 3.48 3.49 3.55 2.85 2.90 CβH3 -0.12 -0.18 -0.08 -0.20 -0.20

Leu89 Cδ2H3 3.25 3.18 2.87 CαH 8.53 8.64 8.64

Ala90 CαH 6.50 6.46 6.52 6.55 6.54 CβH3 2.63 2.66 2.66 2.79 2.78

Ser92 NH 11.04 11.04 11.03 11.01 11.00 His97 CδH 11.07 11.03 11.01 11.02 11.00

CγH 6.83 6.82 6.83 Ile99 CδH3 -3.83 -3.63 -3.34 -3.74 -3.63

CγH3 -3.46 -3.19 -3.18 -3.30 -3.28 CγH’ -1.91 -1.77 -1.22 CγH -9.60 -9.18 -8.71 -9.44 -9.18

Leu104 Cδ2H3 -1.49 -1.31 -1.45 -1.36 -1.37 CδH3 0.07 0.12 0.04 CγH 0.07 0.06 0.15 0.35 0.34

Ile107 CγH3 -0.25 -0.36 -0.25 -0.26 -0.27 CδH3 0.37 0.34 0.25 0.38 0.37


74

Phe138 CδHs 7.05 7.08 7.11 7.12 7.12 CεHs 6.94 7.00 7.11 7.06 7.06 CζH 7.02 7.08 7.11 7.16 7.15

Tyr146 CδHs 7.20 7.41 7.48

Interestingly, most of the hyperfine shifted proton signals in the horse CNMb (see

Fig 3.8 and Tables 3.3 and 3.4) are remarkably shifted upon addition of 10 atm of Xenon

overpressure in the tube. The same behaviour is observed, although less pronounced, in

the 1H NMR signals for the pig CNMb. This evidence is clearer when 1H NMR chemical

shift variation induced by Xenon binding into cyano-metmyoglobins from pig and horse

is plotted as a function of the concentration of Xenon in the NMR tube.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 0.5 1 1.5 2

nXe/Vl (M)

∆δ∆δ ∆δ∆δ

1 H (

pp

m)

5CH3 Ho

5CH3 Pig

CαH Val68 Ho

CαH Val68 Pig

CδH

3 Ile99 Ho

CδH

3 Ile99 Pig

Figure 3.9: Total Xenon-induced proton NMR chemical shift for CNMb (~1 mM) from horse and pig at

room temperature as a function of the total number of moles of Xenon in the NMR tube divided by the

volume of solution (nXe/Vl). The open symbols pertain to the pig CNMb while the full symbols to the

horse CNMb. 5-CH3 - Circle; CδH3 Ile99 - rombus; CαH Val68 – square.


75

Fig. 3.9 shows NMR chemical shift variation of protons signals assigned to some

residues lining the active site and/or belonging to the heme methyl groups of horse and

pig CNMbs upon addition of Xenon (1 to 10 atm overpressure of Xenon). The data

plotted in Fig. 3.9 are extracted either by a series of mono-dimensional 1H NMR spectra

and/or by a series of two-dimensional NOESY spectra.

Observation of the proton chemical shift variation scheme as a function of the

concentration of Xenon in tubes containing solutions of CNMbs from pig and horse

allows making further interesting considerations. Xenon-induced 1H chemical shift

changes are less significant in absolute value in pig CNMb than in horse CNMb.

Furthermore, this variation is linear for pig-CNMb while in horse-CNMb ∆δ first rapidly

increases at low concentration of Xe and then asymptotically reach a saturation value for

higher concentrations. The overall change of proton NMR chemical shift (∆δ 1H) upon

Xenon addition depends on each particular proton considered and reaches for the heme

methyl 8-CH3 of horse CNMb the significant value of 1,32 ppm, which is the maximum

value observed.

In order to more clearly describe what is observed, it is useful to briefly discuss

the relevance of each different contribution to the experimental chemical shifts in

myoglobins.

NMR shifts in paramagnetic molecules are, as has been explained above, the sum

of diamagnetic, pseudocontact (dipolar) and Fermi contact contributions.

These three contributions, however, have very different significance in

characterizing observed proton chemical shifts in the paramagnetic myoglobins studied

here.

For what concerns the diamagnetic part, 1H NMR experiments on the diamagnetic

horse COMb (spectra not shown) yielded small variations (~ 0.05 ppm) in the δdia of the

His93 and porphyrin protons in the presence of 10 atm of Xenon. This observation is in

agreement with a relatively small effect of Xe on 1H NMR chemical shift ∆δdia (< 0.1

ppm) generally observed in diamagnetic proteins40,41. The only exception to our

knowledge concerns the studies on T4 lysozyme where a 1H chemical shift change of

approximately 0.2 ppm was observed in the protons assigned to some side chain

residues42,43.


76

Such values are however considerably smaller than the shift observed in the

present study concerning cyano-metmyoglobins and therefore Xenon-induced variation

of the diamagnetic contribution to the observed proton chemical shift variation are not

sufficient to explain experimental results.

Thus, this suggests that the observed 1H chemical shift changes upon Xenon

addition should be considered to be merely influenced by the hyperfine interaction

between the proton and the unpaired electron of the heme iron and ascribed mostly to

dipolar and contact contributions. On the other hand, hyperfine shifts are determined by

very short-ranged interactions, and it could be reasonably assumed that the part of protein

> 7.5 Å far from the iron center is, for all practical purposes, diamagnetic. Hyperfine

shifted resonances in a region roughly enclosed within a sphere of about 7 Å radius and

centred at the iron ion, are very sensitive probes for the environment of the active site.

Since in this region of the protein reside most of the residues that have been considered

essential for the functionality of the protein44, i.e. those aminoacids that are mainly

implied in the ligand binding process, knowledge of the structure of this region and

characterization of possible ligand-induced conformational distortions would be very

relevant because they would provide new important insights into the processes of ligand

diffusion and binding in hemoproteins.

In order to more clearly understand the origin of Xenon-induced 1H chemical shift

variation in myoglobins, and to further characterize molecular details of Xe-induced

distortions of the protein side-chains at the proximal binding site, it is useful to consider

which structural features could reliably influence the observed 1H chemical shifts.

Equation [3.7] correlates the dipolar chemical shift experienced by each proton

with its polar coordinates with respect to the iron-centred coordinate system in which the

magnetic susceptibility tensor χ is diagonal, i.e. with orientation of the tensor itself. The

pseudocontact shift δdip affects both ligated and nonligated residues according to equation

[3.7] and is strongly characterized by the anisotropic magnetic susceptibility tensor χ.

This latter dependence is very important: the analysis of various CNMb point

mutants38 as well as of model compounds45 has in fact demonstrated that it is in turn

strictly related to structural features of the heme cavity In this regard, some NMR studies

of myoglobin mutants38 aimed to the determination of the orientation of the magnetic


77

axes to understand the relationship between the region of the protein where the mutation

is located and the changes in orientation of the tensor χ. It was demonstrated that point

mutations in the distal side can have effect on the orientation of the Fe-ligand axis (e.g.

Fe-CO, Fe-CN) and consequently on the principal axis of the magnetic susceptibility

tensor (described by the angle β), while mutations that interest the proximal side39

influence only the orientation of the rhombic axes of the tensor (described by the angle

κ=α+γ) , either causing a rotation of the heme or of the proximal histidine.

In particular, the orientation of His93 imidazole ring with respect to the heme iron

has been indicated as a critical factor in determining this perturbation in heme proteins:

π-bonding between proximal histidine and Fe-heme porphyrin has been proposed as a

likely reason for this rhombic perturbation observed in myoglobins. Any change in the

observed proton chemical shifts could therefore be traced back to the displacement of the

His93 imidazole plane with respect to the heme. It should also be noted that while a

correlation does exist between axial His plane and rhombic axes, it is still unclear if this

structural modification is the only determinant of the rhombic axes.

Thus, a reliable quantitative assessment of this dependence can potentially

provide valuable structural information of local distortions in CNMbs.

Moreover, any consideration regarding Fe-1H dipolar interactions in paramagnetic

proteins should additionally take into account that observed proton signals experience

sometimes both contact and dipolar shifts simultaneously. Evaluating both contributions,

and quantitatively separating them, appears as an important issue in order to achieve a

correct interpretation of NMR results.

It has been shown in previous studies concerning low-spin

ferrihemoproteins39,46,47 that the heme Hmeso protons exhibit both contact and dipolar

shifts; it was also pointed out that the measured asymmetry in the meso-H hyperfine

patterns, calculated from the equation

[ ]δγβα δδδδδ −+−=∆2

1meso [3.13]


78

is quantitatively predicted by the rhombic dipolar shifts, being the contact contributions

to their observed chemical shift very similar to each other. Thus, calculation of the

observed ∆δmeso would lead to better understand possible rhombic (in-plane) magnetic

axes displacements related to local structural rearrangements attributable to the presence

of Xenon.

The experimental ∆δmeso of 1.99 ppm that we can calculate for Horse-CNMb in

absence of Xenon [see Table 3.3] is in relatively good agreement with that previously

found in wild type sperm-whale CNMbs (∆δmeso =1.8 ppm)39. The ∆δmeso decreases to 1.3

ppm in horse CNMb after adding 10atm of Xenon gas, thus suggesting that the rhombic

axes have rotated clockwise relative to the heme, implying that κ is to some extent

increased by the presence of Xenon in the proximal cavity of the protein. These

observations suggest that proximal His93 has rotated counterclockwise with respect to

the stationary heme. This hypotesis can be further substantiated by NOE measurements,

as it is discussed in the following. Additional quantitative information can be easily

obtained: as previously suggested, a plot39 of calculated ∆δmeso versus κ indicates a

gradient of 2.5 ppm per 10 degrees of rotation; this correlation allows achieving a

quantitative rough estimate of the relative rotation of the proximal histidine with respect

to the heme group. The calculation suggests a counterclockwise rotation of His93 of

about 2.8 degrees (viewed from the proximal side).


79

3.1.4. - NOE measurements used as a tool to further assess His93 rotation relative to

heme.

A small perturbation of the protein structure such as the rotation of proximal

histidine by less than 10 degrees relative to the heme is considered a local phenomenon

and it was suggested not to influence the remaining structure of the heme cavity. Such

rotation in turn can be easily monitored in low-spin cyano myoglobins by following the

inter-residue dipolar contacts between the rotating residue (His93) and backbone protons

on the E and F helices. Moreover, looking at steady-state NOEs upon saturating either the

heme protons or the His93 protons, allows discriminating between a rotation of the heme

and the rotation of the His93, both leading to a relative displacement of one of them with

respect to the other.

-2,4-2-1,6-1,2-0,8-0,40

1H (ppm)

Val

68 C

αHT

hr67

CγH

7-α'

Ala

71 C

βH

8-CH3

8-CH3 + Xe

(a)

Figure 3.10 (a): NOE spectra of horse CNMb solutions obtained irradiating the 8-CH3 signal. The

spectrum below is relative to the solution of the protein without Xenon and the upper spectrum refers to the

CNMb solution pressurized with 10 atm of Xenon. Arrows indicate correspondent signals in the two

spectra, shifted by the presence of Xenon within the protein cavities.


80

891011121314151H (ppm)

+28%

-13% -23%

His

93 N

pH

Leu8

9 C

αH

His

93 C

βHS

er92

NH

NδδδδH93

NδδδδH93 + Xe 8-

CH

3

(b)

Figure 3.10 (b): NOE spectra of horse CNMb solutions obtained irradiating the His93(F8) NδH proton

signal. The spectrum below is relative to the solution of the protein without Xenon and the upper spectrum

refers to the CNMb solution pressurized with 10 atm of Xenon. Percentages represent changes in the signal

intensities. Peak assignments are reported above most relevant signals.

Figure 3.10 (a) shows that steady-state NOEs resulting from saturating 8-CH3 in

samples pressurized with 10 atm of Xenon gas are essentially the same as for degassed

solutions, indicating an unchanged position of the heme with respect to the protein matrix

in the heme cavity. The absence of significant changes has been confirmed by

deconvolution of the signals (not shown). The absence of significant changes has been

confirmed decomposing the signals into individual gaussians by means of the OriginPro

7 program. Figure 3.10 (b) compares steady-state NOEs without and with 10 atm of

Xenon, obtained upon saturating His93(F8) NδH. The spectrum of myoglobin pressurized

with Xenon shows some new signals and various modifications in signal intensities with


81

respect to the degassed myoglobin sample. In particular new signals appear at 11.03 ppm

and 12.10 ppm, assigned to Ser92 NH and 8-CH3, respectively. Decreases in the

intensities of the signals pertaining to Leu89 CαH (∼-23%) and to His93 CβH (∼-13%) are

observed in the spectrum of the Mb-Xe complex with respect to that of the degassed

solution. Moreover, the signal assigned to His93 NpH has an increased NOE (∼+28%)

after Xenon addition.

Considering the R-6 dependence of the NOE, simple calculations lead to the

conclusion that His93 NδH is displaced by ∼4% further from Leu89 CαH by the Xenon

residing in the proximal cavity (Xe1). Given the His93 NδH - Leu89 CαH distance in

XRD structure of horse Mb (pdb code 1YMB) is 4.25 Å, a 23% decrease in signal

intensity in NOE spectra is indicative of an increase of this distance to approximately

4.43 Å. In the same way it can be calculated a decrease of the distance His93(F8) NδH –

His93 NpH of ∼4% (analogous calculations based on XRD crystallographyc structures

suggest a distance change from 2.99 Å before Xe pressurization to 2.87 Å after Xe

addition). These observations are in very reasonable agreement with the new signals

visible in the NOE spectrum of pressurized samples and they can be easily understood by

taking a closer look to the structure of the proximal cavity and to the positions of the

residues just mentioned. Fig 3.11(a) and 3.11(b) will certainly help to this end.


82

a) b)

Figure 3.11: Proximal side of myoglobins complexed with Xenon. Heme group is shown in licorice style,

Xenon is a yellow ball. (a) Atoms belonging to His93, Ser92 and Leu89 are depicted as balls and sticks and

(b) relative positions of heme group, proximal histidine and Xe1 are reported. Carbon atoms coloured cyan,

Nitrogens in blue and Oxygen in red.

In Figure 3.11, the yellow ball is Xenon in the proximal binding site (Xe1). Figure

on the left [Fig 3.11(a)] shows a particular of the proximal cavity of myoglobins where

residues His93, Leu89 and Ser92 are evidenced by ball and stick representation while

Heme group is represented in licorice style. In particular, atoms Leu89 CαH, Ser92 NpH

and His93NδH are marked in the figure. Figure on the right [Fig 3.11(b)] shows a detail

of the proximal heme cavity: His93 atoms NδH, CβH and NpH are marked. Analysis of

these figures confirms what suggested by NOE measurements obtained by irradiating

His93 NδH, i.e. that the effect of Xenon in the Xe1 cavity on axial histidine - (His93 , F8)

is most likely to displace it toward the δ-meso-H heme side. Whether this displacement is

a translation of the whole F8 residue or a counterclockwise rotation (viewed from the

proximal side) of the imidazole ring about the His93 NεH-Fe bond is uncertain but the

latter hypothesis is the most reasonable, given the strong Fe-His coordination bond.

Further confirmation of the tilt/rotation of His93 ring toward the heme δ-meso-H comes

from the observation of NOE enhancement of the proton signal belonging to 8-CH3


83

(12.5ppm) after pressurization with Xenon, which is not present in the absence of Xenon

[see Fig 3.10(b)].

A method that allows to characterize even small differences in the proximal

region of myoglobins from different species and to determine changes induced by

exogenous ligands would represent a very important tool to study the biological function

of the protein. It is in fact recognized that proximal histidine has an important role in

modulating the reactivity of heme proteins: it has been demonstrated by means of a

number of techniques that a displacement of proximal His upon ligand dissociation, also

referred to as doming vibrational mode, plays a fundamental role in characterizing the

reactivity of this type of biological molecules. Proximal histidine acts as a trigger for

structural changes leading to cooperative transition in hemoglobin48. It has also been

shown that CO binding kinetics are modified in proximal mutants of myoglobins with

respect to wild type39,40. It has been also suggested in early studies that complexation

with Xenon and other small molecules such as cyclopropane49 influence the binding

affinity of carbon monoxide in myoglobins. Another important aspect to be underlined is

the relevance of the interaction between His93 and Ser92 residues. It has been shown that

the hydrogen bond linking the hydroxyl group of Ser92 to the NεH of His93 has an

important role in the protein activity. It was observed in this regard that solution studies

yield different results with respect to crystal structures likely due to the existence of some

substates51 in the structure of the proximal region of myoglobin.


84

3.1.5. - Thermodynamics of Xenon binding to cyano-metmyoglobins from Xenon-

induced 1H NMR chemical shift variations.

For each assigned peak of the horse and pig CNMb spectra, the 1H chemical shift

variation, ∆δ, as a function of Xenon concentration is fitted by the two-site model4:

[ ]

[ ] max1δδ ∆

+=∆

XeK

XeK [3.14]

Here K is an equilibrium binding constant, [Xe] the concentration of Xenon in the

buffer4, and ∆δmax the chemical shift difference between the complexed and free protein

(degassed solution).

It is interesting to note that equation [3.14] resembles Langmuir adsorption

equation commonly used in gas adsorption on the surface of solid materials.

Only the fitting results having correlation function F larger than 0.99 and standard error

of the calculated K smaller than 10% were considered.

In the horse CNMb, 34 out of 70 1H chemical shift variations associated to assigned

proton signals are well described by the two-site thermodynamic model: 21 proton

signals belong to the porphyrin ring and the other 13 to several residues around the heme

region. The remaining 36 1H signals shift only negligibly or show a correlation function

F lower than 0.90. Table 3.5 contains some of the K values extracted from the fitting

analysis.


85

Table 3.5: Values of the binding constant K for the residues of horse CNMb extracted from the fitting of the 1H signals.

Resonances Protons K (M -1) r

1-CH3 126±9 0.9983 3-CH3 96±9 0.9976 5-CH3 127±10 0.9987 8-CH3 90±3 0.9998 2-Hα 80±7 0.9996

2-Hβc 106±10 0.9995

2-Hβt 89±8 0.9989

4-Hβc 85±7 0.9986

4-Hβt 94±9 0.9993

6-Hα 89±3 0.9997

6-Hα’ 93±7 0.9992

6-Hβ 137±13 0.9986

6-Hβ’ 158±11 0.9995

7-Hα 154±15 0.9976

7-Hα’ 87±7 0.9973

7-Hβ 81±8 0.9934

7-Hβ’ 88±8 0.9961

Hα 119±10 0.9986

Hβ 103±10 0.9991

Heme

Hδ 80±10 0.9967 CαH 80±6 0.9988

CβH 110±7 0.9991

CβH’ 112±10 0.9991

NpH 85±4 0.9992

NδH 147±11 0.9995

His93

CεH 94±8 0.9986 CδH3 116±9 0.9978

CγH’ 133±12 0.9938 Ile99

CγH 80±7 0.9910 CγH 159±10 0.9981

Ile104 Cδ2H3 145±9 0.9991

Ile107 CδH3 94±9 0.9938 Val68 CαH 98±9 0.9982


86

The 8-CH3 (∆δ=1.32 ppm) and Hα (∆δ=1.81 ppm) protons of the porphyrin ring exhibit

the maximum chemical shift variations.

The binding constant K, as derived from the individual fitting of the 34 signals,

ranges between 80 M-1 and 150 M-1, while fitting all chemical shift variations on the

whole (software package Origin from Microcal) leads to K=109 ±7 M-1. ∆δmax lies

between 2.1 and -0.69 ppm, revealing non-negligible proton chemical shift variations

when Xenon is bound to the protein.

Negligible shifts or linear trends characterize the 56 assigned proton signals in the

pig CNMb when Xenon is added up to 10 atm overpressure in the tube. 24 signals have a

shift larger than 0.05 ppm, 19 belonging to the porphyrin ring and the others to the CδH3

and CγH group of Ile99, to CαH and Cγ2H3 of Val68 and to CδH of Val64. The maximum

∆δ is associated with the 8-CH3 (∆δ=0.45 ppm) and Hα (∆δ=0.37 ppm) protons of the

porphyrin ring [see Table 3.3]. The chemical shift variation of these signals changes

linearly upon Xenon addition. Using equation [3.14] to fit these chemical shift variations

we estimate a binding constant K lower than 5 M-1 for pig CNMb samples, while K=109

M-1 is obtained for the horse CNMb.

The use of a single K value in a linear regime can be justified by the 129Xe spin

lattice relaxation rate, which clearly indicates the presence of Xe in the proximal cavity.


87

3.1.6. - Myoglobins: CONCLUSIONS

The present work summarizes the results of our efforts addressed to the

challenging issue of characterizing the structural properties of the cavities in different

paramagnetic myoglobins in solution. The application of 129Xe NMR to the investigation

of biomolecules has proven to be very useful in providing essential information on their

structure and dynamics involved in the binding process.

The analysis of paramagnetic proteins in solution is time consuming and requires extreme

care in the preparation of the sample. Moreover, the presence of a paramagnetic ion

hinders the application of laser-polarized Xenon and the hyperfine couplings between the

unpaired electron of the metal and the neighboring atoms significantly influence the

NMR results. We have shown here that these strong interactions, far from representing a

drawback of the technique, can be exploited to study molecular details the Xe-protein

interactions in solution. The study combines the NMR analysis of the proton nuclei lining

the cavities within the host protein and the 129Xe that probes all the available

environments in a fast exchange condition.

The analysis of the 129Xe chemical shift and the spin lattice relaxation time

monitors features of Xe-Mb interactions. The Xe binding characteristics are found to be

strongly influenced by the interior structure and by the hydrophobicity of the proximal

(Xe1) cavity as evidenced by the comparison of the data obtained for pig (K=48 M-1;

δin=55 ppm; rFe-Xe=6.8 Å) and sheep (K=36 M-1; δin=51 ppm) CNMbs with those of

horse (K=158 M-1; δin=-18 ppm; rFe-Xe=5.3 Å ) and rabbit (K=131 M-1; δin=-22 ppm )

CNMbs. The agreement between 129Xe and 1H NMR findings indicates that the residue

substitution in the proximal cavity of Mbs causes a different distribution of Xenon inside

the protein. In myoglobins from horse and rabbit, Xe1 is the main binding site, while in

pig and sheep Mbs there is evidence of the presence of Xe in the Xe1 cavity and most

likely in Xe3, as competing binding site of Xe1. This suggested assignment is based on a

previous study where a second xenon binding site was detected in an NMR study25 in

solution and it was believed to be located in a cavity close to the surface of the protein

In particular, the major novelty brought by the present work concerns the detailed

discussion of of the 129Xe/1H - NMR method used to describe the cavities close to the


88

active site of paramagnetic myoglobins, which are supposed to play a fundamental role in

the protein functionality.

It appears clear from this study that the Xenon population of the proximal cavity,

i.e. Xenon affinity for this region of the protein, is strongly influenced by the structure

and the hydrophobicity of the cavity itself.

It is demonstrated that proton chemical shift variations in these systems are

understandable considering changes in the orientation of the principal magnetic axes of

the CNMb-Xe complexes. Besides complementing 129Xe NMR results with site-specific

information, 1H NMR measurements help to achieve a picture of Xenon-induced local

distortions of the protein. Xenon-induced local distorsions of the protein structure are

spotlighted and associated to a residue located right at the active site. According to the 1H

NMR data, Xenon induces the tilt of the residue His93 relative to the heme plane and

consequently causes an alteration of the magnetic axes. These findings have, in our

opinion, very general involvements, since structural modifications of the cavities and

structural perturbations of the ligand tilt could not only affect the kinetics of ligand

binding, but also determine relative affinities, and consequently the physiological

function of the myoglobin


89

3.2. Copper-containing amine oxidases enzymes: Xenon-induced

reactions

3.2.1. - Lens Esculenta Amine oxidases (LSAO) in solution: 129Xe NMR chemical

shifts

Several Amine Oxidases have been investigated in this work in order to clarify

the interaction of this class of enzymes with Xenon in solution. Our studies first

concentrated on the characterization of one of them, Lens esculenta Amine Oxidase

(LSAO), with the attempt to characterize the main features of Xe-enzyme complexes. In

particular the work has focused on the spectroscopic characterization of Xe-enzyme

interaction and on the challenging issue of understanding the relationship structure-

function in these complex biological systems.

Figure 3.12 shows 129Xe NMR spectra of solutions of LSAO and buffer, both

pressurized with 10 atm of Xenon gas. The spectra were collected 48 hours after Xenon

addition.

-15-10-505101520

ppm

Buffer

LSAO 0.33 mM

Figure 3.12: 129Xe NMR spectra of a solution (Na+-phosphate buffer 1 mM, pH 7.0, 20% D2O) of 0.33 mM

LSAO pressurized with 10 atm of Xenon gas. Shifts refer to the chemical shift of 129Xe dissolved in buffer.


90

The 129Xe NMR signal in the solution containing the LSAO enzyme is shifted

downfield with respect to the resonance of Xenon in buffer, which has been fixed to 0

ppm and used as a reference value of chemical shifts.

The presence of a single resonance in protein solution indicates that Xenon is in

fast exchange, in the NMR time scale, between all the available environments, i.e.

enzyme cavities, enzyme surface and solvent.

Further characterization of the interaction between Xenon and LSAO can be performed

by monitoring the 129Xe chemical shift, spin lattice relaxation time T1 and NMR signal

linewidth. These parameters were measured as a function of both Xenon and protein

concentration.

By increasing LSAO concentration from 0.26 mM to 3.4 mM, while keeping

constant at 8 atm the Xe pressure, the 129Xe NMR resonance is shifted downfield from

0.33 ppm to 2.8 ppm.

Xenon NMR chemical shift are also influenced by Xenon pressure: 129Xe NMR

signal is shifted at 2.5 ppm and 2.9 ppm at 5 atm and 10 atm of Xenon respectively, at

constant protein concentration (0.25 mM LSAO). 129Xe line width, taken as the full width at half maximum (fwhm), increased with

increasing protein concentration (104 Hz and 63 Hz at 0.37 mM and 0.26 mM of LSAO

concentration, respectively) and showed for all the studied concentrations of enzyme

larger fwhm values than the signal of Xenon in buffer solution.

The 129Xe spin lattice relaxation times (T1) in the LSAO and Copper–free LSAO were

measured at 25°C and at the Xenon gas overpressure of 10 atm. The T1 value (3.2 ± 0.5 s)

measured in the LSAO solution (0.3 mM) is longer than the T1 measured in Copper–free

LSAO (18 ± 2 s-1). However, both the T1 values were much smaller than the spin-lattice

relaxation time of Xenon in buffer (T1 ~ 500 s).

All these observations point out that an interaction between Xenon and the LSAO

actually exists which modulates the observed 129Xe NMR parameters: these features were

in fact observed in other protein solutions and resulted from the fast exchange of Xenon

between both specific and non-specific sites of the protein and the buffer 4,8,25,52-54.

The measurement of shorter longitudinal (spin-lattice) relaxation times in LSAO

compared to the T1 observed in Copper-free enzymes suggests that most likely additional


91

relaxation mechanisms associated with the presence of the metal ions characterize Xe-

LSAO interaction. This observation could be further evaluated by explaining the

molecular mechanism of Xe-protein interaction and describing the intermediate species

involved in the enzymatic activity of LSAO.

3.2.2. - Spectral changes in the UV-vis region of LSAO solutions induced by

substrates and Xenon.

600500400300

0.2

0.15

0.1

0.05

0

Wavelength (nm)

Abs

orba

nce

Figure 3.13: UV-vis spectra of native LSAO solution (32 µM of active sites in 1 mM sodium phosphate

buffer, pH 7.0) in anaerobic conditions before () and after (---) addition of 100 µM benzylamine.

Electronic absorption spectra of LSAO show characteristic features of oxidized

Cu-TPQ (resting enzyme) cofactor and of its reduced form, depending on the

experimental conditions. In anaerobiosis, the oxidized form is observed in absence of the

substrate and its typical spetrum is reported in Figure 3.13. A broad absorption band in

the UV region with maximum at 278 nm is ascribed to aromatic amino acids. In addition,

LSAO has a characteristic pink color due to the presence of the oxidized TPQ cofactor,

which has a broad absorption band around 498 nm in the visible spectrum. The extinction


92

coefficients ε at 498 and 278 nm are determined to be 4.5 mM-1 cm-1 and 245 mM-1 cm-1,

respectively, for LSAO55.

The addition of 1 mM benzylamine to a solution containing 16 µM LSAO (32 µM

active sites), in the absence of air, caused the immediate changes in the spectra, which

can be explained considering the reaction path reported Scheme 3.1 below.

The bleaching of the 498 nm absorption band indicates the rapid formation of a

reduced TPQ intermediate and it is in particular associated with the formation of the CuII-

quinolaldimine (IV) (see Scheme 3.1). Oxidation of the bound substrate (followed by

hydrolysis) releases the aldehydic product, leaving the CuII-aminoresorcinol derivative

(V), which has a bound ammonia molecule. This species is still colorless. The CuII-

aminoresorcinol is in equilibrium with the yellow, EPR-detectable CuI-semiquinolamine

radical (VI) , containing the substrate-derived nitrogen covalently bound to the aromatic

ring system56-58, and characterized by absorption bands at 464, 434 and 360 nm.59,60 The

extinction coefficients of reduced LSAO at pH 7 and 298 K are: ε464 = 7.1 mM-1 cm-1,

ε434 = 4.6 mM-1 cm-1.61


93

Scheme 3.1: Molecular reaction pathway so far described for plant Amine Oxidases


94

In our experiments, approximately 30% of the active LSAO enzyme was

converted into the radical, which was determined by ESR measurements of the CuII to

CuI ratio (not shown). The amount of benzaldehyde produced was measured to be 1.0 ±

0.1 mol mol–1 LSAO subunit [see Fig 3.14].

40200

40

20

0

LSAO (nmol active sites)

Pro

duct

(nm

ol)

Figure 3.14: Anaerobic product release. Formation of aldeyde () and ammonia () at different LSAO

concentrations. Slope of the benzaldehyde line is 1 nmol benzaldehyde of LSAO subunits.

It is clear from Figure 3.14 that while aldeyde is a product of the catalytic process

of LSAO in presence of benzylamine in anaerobiosis, ammonia is not formed at this

stage.

Readmission of oxygen rapidly regenerated the oxidized TPQ cofactor with

release of hydrogen peroxide and ammonia.

In order to characterize the effect of Xenon on the spectral features of the enzyme

solution, i.e. on the catalytic process of LSAO, a LSAO solution (1 ml solution; 26 µM

LSAO active sites) has been pressurized with 10 atm of Xenon gas in absence of oxygen

and without any amine substrate. After approximately six hours spectral changes in the


95

UV-vis region appear [see Fig 3.15]: the broad absorption at 498 nm starts to disappear

and simultaneously characteristic absorption bands become detectable suggesting that the

radical CuI-semiquinolamine radical is formed. The reaction is over after 48 h and the

quantitative analysis of absorbances at 464 and 434 nm shows that more than 60% of the

native enzyme was converted into the radical form.

600500400

0.12

0.08

0.04

Wavelength (nm)

Abs

orba

nce

600500400300

0.2

0.1

0

W avelength (nm )

Abs

orba

nce

Figure 3.15: Uv-vis spectra of native LSAO acquired as a function of time after pressurization with 10 atm

of Xenon gas. Arrows indicate formation of the radical species and disappearance of the oxidized TPQ

form. The spectrum of LSAO-Xe complex after 48 hrs is reported in the small panel.

Ammonia production was not detected under 10 atm Xenon gas overpressure in

anaerobic condition. After readmission of Oxygen the absorption spectrum of oxidized

TPQ was rapidly recovered and the formation of approximately 2 mol of ammonia and 2

mol of hydrogen peroxide were detected per mol of LSAO.

As a comparison, two more LSAO samples, prepared with the same procedure as

those treated with Xenon or the substrate, have been respectively kept in presence and

absence of Oxygen and analyzed after 48 hours: their UV-vis spectra do not show


96

modifications of the characteristic absorption at 498 nm suggesting that the observation

of spectral absorptions typical of the reduced Copper-TPQ form can be only ascribed to

the presence of either the amine substrate or Xenon.

Xenon does not cause significant modifications in the secondary or tertiary

structure of LSAO as suggested by measurements on Xenon-treated enzymes by means

of CD spectroscopy, tryptophan fluorescence and fluorescence of the hydrophobic probe

1-anilinonaphtalene 8-sulphonate, which is sensitive to microenvironmental changes of

proteins (spectra not shown). Also X-ray crystallographic studies so far performed on

Xe*AOs complexes have pointed out that only minor structural arrangements of the

protein structure are observable upon Xenon binding.

When a second cycle is performed, i.e. when Xe-treated LSAO is exhaustively

dialyzed and pressurized again with 10 atm of Xenon, neither the radical is observed nor

the production of ammonia and hydrogen peroxide.

On the other hand, Xe-treated LSAO, after being exhaustively dyalized, is able to

react again with an amine substrate, thus forming the radical species (spectra not shown)

in anaerobic conditions and producing after readmission of Oxigen hydrogen peroxide

and ammonia, thus confirming the similarity of native and Xe-treated LSAO enzymes

regarding their activity toward amine substrates.

In order to more clearly describe the catalytic process of AOs and aiming to get

further insight on the role of Copper ion on the redox process, Copper-free enzymes were

also studied. Copper can be removed from native LSAO by treatment with

diethyldithiocarbamate62. In our experiments, the residual copper, measured by atomic

absorption spectroscopy, was measured to be 0.2 ± 0.02% of the original content.

A band at 480 nm chacterizes the visible spectrum of Copper-free LSAO as it is shown in

Figure 3.16.

After treatment with 10 atm Xenon the absorbance relative to the band at 480 nm

of the apoprotein progressively decreased in intensity, clearly indicating the formation of

the colorless aminoquinol, without formation of the characteristic bands of the radical at

464 nm and 434 nm. A new band at around 360 nm instead appeared reaching its

maximum after 24 h [see Fig. 3.16]. Quantitatively analyzing the absorbance at 480 nm,


97

we calculated that about 50% of the apoenzyme remained in its oxidized form after Xe

treatment.

600400

0.3

0.2

0.1

0

Wavelength (nm)

Abs

orba

nce

Figure 3.16: Absorption spectrum of a solution of Copper–free LSAO (40 µM active sites) pressurized

with 8 atm of Xe. It is shown with arrows the bleaching of the absorption band at 480 nm and the appearing

of a new band at 360 nm, which indicates the formation of aminoquinol.

Thus, Copper-free enzyme under 10 atm Xenon, clearly unable to form the radical

species due to the absence of the metal ion necessary to allow the electron transfer, to

some extent forms the aminoquinol species (≈ 50%). Reliably, the absorption band

around 360 nm [see Fig. 3.16] is due to the formation of a neutral form of the product

Schiff base quinolaldimine, similarly to what has been already observed in model

compounds and in bacterial Copper amine oxidases63. This observation, moreover, would

suggest that there is a back reaction between aminoquinol and the allysine residue,

according to the Scheme 3.2 below.


98

-O

OH

N

CH

H

CH2

CH2

CH2

CHCO NH

HO

OH

NH2

HO

OH

N

CH

CH2

CH2

CH2

CHCO NH

(a)(b) (c)

CH O

CH2

CH2

CH2

CHCO NH

Scheme 3.2: Proposed Xe-induced reaction in Copper–free LSAO enzyme. (a) aminoquinol; (b)

quinolaldimine, protonated form; (c) quinolaldimine, neutral form. It is worthnoting that in Copper–free

enzyme the radical is not formed and the quinolaldimine accumulates (λ = 360 nm) over the time of

treatment.

Therefore, quinolaldimine species, well characterized in bacterial amine oxidases

so far, is for the first time isolated and evidenced also in plant enzymes.

In conclusion, while it has been proposed that in some AOs, such as BSAO,

Copper would have just a structural role64,65, mainly interfering in the electrostatic

stabilization of oxygen, the results obtained here rule out this possibility in plant amine

oxidases and in contrast they show that Copper is certainly involved in the amine

oxidation process.


99

3.2.3. - Involvement of a lysine residue in the intra-molecular catalytic mechanism of

LSAO

The observation just described of the activation of LSAO enzyme in the absence

of its substrate led to hypothesize the role of intramolecular reactions induced by the

presence of Xenon atoms inside the hydrophobic cavities of the protein. The side chains

near the active site appear to be likely responsible for the reaction. The candidate residue,

in order to act as intramolecular enzymatic substrate, must clearly mimic the amine

substrate, thus the choice should be restricted to glutamines, asparagines and lysines,

which contain an amine group at the end of their side chain. Between these three amino

acids, lysine appears to be the most eligible because of its relatively long hydrophobic

chain which makes it the the less sterically hindered among the three of them.

LSAO contains 38 lysines in each subunit66. Attempts to identify the lysine

residue converted into allysine under Xenon pressure were carried out by proteolytic

digestion with trypsin and lysyl endopeptidase, followed by HPLC analysis. This

approach, however, revealed to be unsuccessful at a first attempt, as it resulted in very

complicated elution profiles making not easy the unequivocal identification of the

modified lysine.

Moreover, 129Xe NMR experiments cannot provide a more detailed

characterization of the interaction between Xenon and the protein, and the actual location

of a possible cavity or cavities involved remains unknown and would require further

studies that should involve NMR structural determination and signal assignments. These

achievements are however still far from what NMR technique is able to accomplish

nowadays. The dimer of LSAO is in fact formed by more than 1860 residues, which are

way too many to allow a complete structural characterization in solution by NMR, even

with the most advanced instrumentations. The primary limitation in determining protein

structure by NMR is in fact the size of the protein. The size limitation for complete

atomic-resolution structure determination by NMR is currently ~30 kDa, though

backbone assignments and general folds have been described for proteins up to 100 kDa.

Few considerations may help in formulating reliable hypothesis on molecular

mechanisms involved in the catalytic process of AOs.


100

A lysine residue, Lys296, located at the active site of Pea Seedling Amine

Oxidase has been established to be structurally important from X-ray crystallographic

studies of the enzyme.

a) b)

Figure 3.17: Relative positions of Copper, TPQ cofactor and Lysine 296. (a) Top view and (b) side view.

The Cu is represented as an orange sphere, TPQ in red licorice style and lysine is in licorice with Carbon

atoms coloured cyan, Nitrogens in blue and Oxygen in red. Pictures are derived from crystallographic

structure of PSAO (PDB code 1KSI)72

The X–ray crystallography of PSAO has demonstrated67 that Lys296 makes a

hydrogen bond with the phenolic group at C–4 of TPQ [see Figure 3.17]. This residue is

located into domain D4, between β-sheet C5 and α helix H8, close to the entrance to a

channel which has been found to be suitable for the movement of substrate and products

to and from copper/TPQ active site buried in the protein interior. The hydrogen bond

between TPQ and a lysine residue is typical of plant amine oxidase while it has never

been observed in crystals of other AOs.68-71. The role of Lys296 is not yet precisely known

but it has been suggested that it might favor the deprotonation of the TPQ contributing to

the observed differences among the catalytic efficiencies of several AOs72.

The internal surface of the channel above mentioned is lined by residues which become

more hydrophobic as the active site is approached. This would make the channel suitable

for Xenon, which is known to prefer hydrophobic regions73. A direct experimental

evidence that Xenon binds to several sites has recently been obtained in the crystal of

various AOs67. Only one of Xe–binding site is closest to Cu/TPQ center. In pea seedling

AO the Xe atom is bound in a pocket at 7.7 Å from the copper atom and 9.3 Å from TPQ;


101

the nearest neighbor amino acid residues are Ile405, Leu407, Tyr446, Ile601, Leu616 and

Thr618.

Since the identity in amino acid sequences of LSAO and PSAO72,74,75 is about

92% it should be safe to accept that both enzymes have an almost identical structure.

Moreover, PSAO and LSAO are similar in functional properties,60 and this would be

compatible with their structural similarity. As observed in PSAO72 the Copper ion and

TPQ are in close proximity (6 Å) but they are not coordinated. Moreover a small

displacement of the TPQ would be required to facilitate the extremely rapid electron

transfer CuII–aminoquinol/CuI–semiquinolamine, and the TPQ side chain appears to be

sufficiently flexible to accommodate this change. A considerable conformational

flexibility of TPQ is also proposed when an amine substrate attacks at C–5 of TPQ and

H+ abstraction of the active site base Asp300 requires that TPQ rotates by 180 degrees.

Figure 3.18: XRD crystallographyc structure of PSAO Copper in the so called ‘on copper’ conformation67, suggested to be the only productive conformation relative to the catalytic activity of AOs.

In considering the results obtained, we propose a possible reaction pathway that

would explain the results observed. The suggested catalytic process is reported in Scheme

3.3. In native LSAO, under 10 atm Xenon, a movement of either the TPQ cofactor or an

α helix containing a lysine residue may occur making the C-5 of TPQ closer to ε-amino

group of this residue [see Scheme 3.3], and it is well known that TPQ side chain appears

to be sufficiently flexible. The quinoneketimine intermediate (2) occurs between TPQ


102

and ε-amino group of lysine and accumulates during the lag phase. Then, the

quinoneketimine (through intermediate species like the CuII-carbanion and the CuII-

quinolaldimine (3) forms the colorless reduced CuII-aminoquinol derivative (4) which is

in equilibrium to the yellow radical intermediate (5). Quantitative assessment of the

radical species from UV-visible spectra indicates that more than 60% of the Xe-treated

enzyme is in the reduced radical form, suggesting that both monomers in a dimer can

generate the radical. Readmission of oxygen causes the release of ammonia from the

reduced TPQ that is converted to the oxidized form (1’), together with the simoultaneous

formation of hydrogen peroxide. As a consequence, the lysine residue must be converted

into allysine by oxidative deamination.

In order to confirm the proposed Xe-induced molecular mechanism the oxidative

deamination of a lysine residue was monitored by the detection of α-amino-adipic-δ-

semialdehyde-fluoresceinamine derivative (AASF) by HPLC in native Xe-treated LSAO

and Copper-free Xe-treated LSAO.

In Xe-treated LSAO a peak was detected with retention time about 9.7 min and

was identified as AASF by comparison with a Lys homopolymer-fluoresceinamine

derivative as standard. One mol of allysine residue per mol of monomeric enzyme was

detected (not shown). The allysine containing LSAO is unable to regenerate the radical

species under 10 atm Xenon in the absence of the substrate in anaerobiosis. This clearly

demonstrates the relevance of a lysine residue at the active site in the observed Xe-

induced chemical reaction in the plant enzyme oxidases analyzed. It is however very

tough to exactly localize the residue, due to the complexity of the systems.


103

Scheme 3.3: Proposed intramolecular reaction pathway that would explain our experimental results.

Finally, in addressing the usefulness of 129Xe NMR spectroscopy in the

characterization of biological compounds in solution, it must be pointed out that these

systems are generally characterized by complex structures and often by the presence of

more than one specific site for ligands and/or substrates. The nearest neighbor residues of

the bound Xenon atoms in the cavities are predominantly nonpolar side chains, but they

include polar side chains and backbone peptide groups. This, together with the fact that

the observed 129Xe chemical shift is dynamically averaged among different binding sites

and at the same time interacts with the protein surface, makes it difficult to separate the

individual contributions so as to show whether a particular Xenon-binding site is

responsible for the different components observed in the AOs studied in solution.


104

Hyperfine interactions with unpaired electrons in radical species and/or paramagnetic

metal ions could be a further source of information, as long as their effect on 129Xe NMR

observables can be distinguished from other structural or dynamic factors affecting NMR

parameters. These 129Xe NMR outputs cannot provide local information on the host–

guest interaction involved. Experimental evidence of the fast diffusion of Xenon within

AOs clearly opposes the static and average pictures given by single-crystal XRD

structures, which seem to show that Xenon atoms are localized at specific sites.

Moreover, it is worth noting that, as the single crystal XRD results utterly ignore the

fundamental dynamic features involved in the functionality of the biomolecules in

solution, hypotheses on biological activities based on crystal structures should be

considered critically.

X-ray crystallography, on the other hand, does not suffer from the size restrictions

of NMR, with protein size having no direct bearing on the solvability of the protein or

protein complex. This is at least partly why most protein structures have been determined

by X-ray rather than NMR. The limitation of X-ray crystallography is its static nature.

This means that only a single structure can be determined and any protein movement

during data collection results in decreased resolution. Indeed, in many structures there are

segments of the protein that are so disordered they are not contained in the structure.

Aiming to confirm the role of a lysine residue in the enzymatic reaction process of

Amine Oxidases we have investigated the binding of Xenon to AOs purified from various

sources, and our results strongly support the hypothesis that a lysine residue is implicated

in the catalytic mechanism of plant enzymes. In particular, three AOs [pea seedling AO

(PSAO), Euphorbia Characias AO (ELAO) and pig kidney AO (PKAO)] were tested by 129Xe NMR spectroscopy and optical spectroscopy.


105

3.2.4. - 129Xe NMR of PKAO, ELAO and LSAO solutions.

Chemical shifts.

Figure 3.19 shows 129Xe NMR spectra of PKAO, ELAO LSAO solutions

compared to the 129Xe NMR spectrum of Xenon dissolved in buffer solution. A single 129Xe NMR resonance in these solutions indicates that Xenon experiences fast exchange

between all the available environments, i.e. protein cavities, protein surface and solvent:

in this condition the observed NMR parameters are a weighted average of the values of

the same parameters that would be observed in each of all the possible sites in absence of

exchange.

When protein solutions are pressurized in NMR tubes with 10 atm of Xenon gas,

the 129Xe NMR signal in AO samples is shifted downfield as compared to the resonance

of the same amount of Xenon in the buffer, which is used as a reference and set to 0

p.p.m.

Due to intrinsic difficulties of the purification process, only limited and variable

quantities of these enzymes could be succesfully purified. Observed chemical shifts must

be analyzed therefore also taking into account the concentration of each solution

analyzed. Measured chemical shifts in the different protein solutions analyzed are as

follows: ELAO 3.96 p.p.m. per 0.35 mM, corresponding to 11.3 p.p.m.·mM-1; and PKAO

1.4 p.p.m. per 0.15 mM, corresponding to 9.4 p.p.m.·mM-1.


106

-12-8-40481216

ppm

ELAO

LSAO

PKAO

Buffer

Figure 3.19: 129Xe NMR spectra of Xenon. 129Xe (10 atm) spectra in a solution (Na+–phosphate buffer 1

mM, pH 7.0, 20% D2O) containing 0.35 mM ELAO, 0.28 mM LSAO and 0.15 mM PKAO. Shifts refer to

the 129Xe chemical shift dissolved in buffer. Spectra were collected approximately 48 hours after Xe

addition.

Relaxation time T1.

A single value of longitudinal relaxation times (T1) has been sufficient to fit

experimental results, confirming that Xenon atoms are in a fast exchange condition in the

time scale of NMR longitudinal relaxation. The T1 measured for all native enzymes was

found to be much shorter than the T1 value of Xenon in the buffer (ELAO T1 = 4.3 ± 0.5

s; PKAO T1 = 5.5 ± 0.8 s; buffer T1 = ∼500 s). These features, which were also observed

in LSAO and other protein solutions, confirm that there is an interaction between the

dissolved Xenon and the interior of the protein. However, the actual location of the

possible cavity or cavities involved in the binding of Xenon remains unknown.

The relatively high enzyme concentrations (0.25– 0.35 mM) used and the low

ionic strength (1 mM) of the phosphate buffer in the experiments on solutions of Xenon-

enzyme complexes have led to the formation of inactive precipitates. We were therefore

unable to obtain reliable results from bovine serum AO (BSAO).


107

3.2.5. - Spectroscopic features induced by amine substrates and Xenon in several

AOs.

Due to the slightly different amino acid composition and structural conformation

of the three AOs considered, slightly different absorption bands characterize the UV-

visible spectra of BSAO, PKAO and ELAO. BSAO shows an electronic absorption band

at 476 nm (ε476 = 3800 M-1·cm-1)76, PKAO at 490 nm (ε490 = 4000 M-1·cm-1)77, PSAO and

LSAO at 498 nm (ε498 = 4100 M-1·cm-1)78,79, and ELAO at 490 nm (ε490 = 6000 M-1·cm-

1)80.

Spectra of evacuated solutions of native highly purified enzymes and spectral

changes in the UV-visible region observed after the addition of the amine substrate are

reported in Figure 3.20 and Figure 3.21. Immediately after the addition of the substrate,

in the absence of air, the visible absorption band that characterizes spectra of native AOs

disappears, indicating the formation of a reduced TPQ intermediate. However, while in

plant AOs (LSAO, ELAO and PSAO) this reduced TPQ form equilibrates with the

yellow-coloured semiquinolamine radical by transferring one electron to Copper, in

mammalian AOs a different behaviour is observed. In PKAO the transformation of the

radical species can be observed only in the presence of cyanide ions81 [Fig. 3.21(b)]. On

the other hand, BSAO, that does not form the radical species during the normal catalytic

cycle65, stayed in the reduced aminoquinol form.

600500400300

0.2

0.15

0.1

0.05

0

Wavelength (nm)

Abs

orba

nce

A

600500400300

0.2

0.1

0

Wavelength (nm)

Abs

orba

nce

B

Figure 3.20: UV-visible absorption spectra of LSAO and PKAO and spectral changes induced by amine

substrates. (A) 16 µM native LSAO in 1 mM Na+–phosphate buffer, pH 7.0, in anaerobic conditions before

(---) and after () addition of 10 mM putrescine. (B) 19 µM PKAO in 100 mM Tris/HCl buffer, pH 7.2,

before (---) and after () addition of 10 mM cadaverine in anaerobic conditions and in the presence of 100

µM CN–.


108

600500400300

0.3

0.2

0.1

0

Wavelength (nm)

Abs

orba

nce

A

600500400300

0.15

0.1

0.05

0

Wavelength (nm)

Abs

orba

nce

B

Figure 3.21: Absorption spectra of native ELAO and PKAO and spectral changes of their solutions

pressurized with 10 atm Xenon gas. Conditions: (A) 11 µM ELAO and (B) 19 µM PKAO in 1 mM Na+–

phosphate buffer, pH 7.0, in 10 atm Xenon. Spectra of the reduced forms () were recorded after 48 h.

As previously discussed, when a solution containing LSAO (10 mM) is

equilibrated with 10 atm of Xenon gas without a substrate and in absence of air, the

semiquinolamine radical form can be observed after a lag period. Similar behavior was

observed with AOs from pea seedlings and E. characias latex ELAO and from pea

seedling PSAO.

Interestingly, the results we obtained with mammalian proteins were different. For

PKAO81, where the semiquinolamine radical appears in the presence of the substrate and

CN–, bleaching of the 490 nm band started with a marked time lag (∼ 6 h) after addition

of 10 atm of Xenon gas [Fig. 3.21(b)] but that the radical species formed neither in the

presence nor in the absence of Cyanides. As observed in plant enzymes, the absorption

spectrum of oxidized TPQ was recovered after readmission of oxygen, and 1 mol of

ammonia and 1 mol of hydrogen peroxide per mole of active site were detected.

In BSAO, no changes in the spectral features were observed in presence of 10 atm of

Xenon gas, indicating that the TPQ cofactor remained in its oxidized form.


109

Interesting experimental observations come from following reactions of Xe-

treated enzymes with amine substrates, in particular in regard with the reduced catalytic

activity of the enzymes toward specific substrates. Although the general behaviour of Xe-

treated enzymes very closely resembles that of native species, catalytic activity

(described by the kinetic constant kc [sec-1]) of Xenon-treated LSAO towards diamine

putrescine was shown to be about 40% of that of the native LSAO, whereas the kc for

monoamine aromatic benzylamine does not change.

Decreased activity was similarly detected in Xe-treated PSAO and in ELAO. The

catalytic activity of Xenon-treated PKAO towards cadaverine and benzylamine was

shown to be about 20% of that of the native enzyme. Xenon-treated BSAO, which is

unmodified by the effect of Xenon and retains its oxidized form, showed the same

activity as the corresponding native enzyme.

These results are most likely related to two distinct molecular mechanisms of

catalytic activity in enzymes from mammalian and plant sources. We suggest that in the

plant enzyme the ε-amino group of Lys296 may interact with the positive charge of the

amino group of putrescine. This lysine residue could have an important role in conferring

substrate specificity to the enzyme; therefore, the transformation of this lysine into

allysine would have important implications in the catalytic efficiency.


110

TPQ

Asp300 Lys296

H3N+ CH2 CH2 CH2 CH2 NH3

+

Asp300

TPQ

COO-NH3

+

COO- CHOAllys296

H3N+ CH2 CH2 CH2 CH2 NH3

+

Allys296 CHOCOO-

TPQ

Asp300

H3N+ CH2

Figure 3.22: Active site of plant AOs. The model of the active site shows the possible interaction of native

and Xe-treated AOs with two substrates: benzylamine, which represents a substrate with an apolar chain,

and putrescine, with a positively charged amino group. The positively charged ε-amino group of lysine

exerts a repulsive force towards substrates characterized by the presence of a positively charged amino

group, such as putrescine, leading to a lower catalytic efficiency of the enzyme when lysine is transformed

into allysine. Neither lysine nor allysine can interact with the apolar chain of benzylamine, suggesting that

this amino acid residue is responsible for the different substrate specificities.

As stated above, X-ray crystallography of PSAO has demonstrated that a lysine

residue, the Lys296 located in domain D4, forms a hydrogen bond with the phenolic

group of TPQ when in an unproductive on-copper conformation72 while its role in the

off-copper conformation is still unknown.

Currently, new forms of PSAO native protein crystal are available67 in the so-

called ‘off-copper conformation’. In this structure, the O-4 of TPQ is hydrogen bonded to

the hydroxyl group of conserved tyrosinyl residue Tyr286, and the TPQ orientation is in

the active form, with the aspartic active site base residue (Asp300) in an excellent


111

position for abstraction of the Cα proton from the substrate, so that TPQ does not rotate

during the catalytic mechanism.

Figure 3.23: Relative positions of TPQ cofactor, Asp300, Tyr286, Copper atom (orange), and Xenon atom

in crystallographyc structure of PSAO*Xe complex. TPQ cofactor is in the so called ‘off Copper’

conformation.67

Lys296 is conserved in LSAO as well as in ELAO (Lys302), while in BSAO, the

residue corresponding to Lys296 of LSAO is Thr381. However, an arginine is present at

position 382 [Fig. 3.24].

Figure 3.24: Relative position of TPQ cofactor, Copper atom and Arg382 residue in BSAO enzyme from

crystallographic studies


112

The amino acid sequence of PKAO is unknown. However, considering its

significant homology with known reported sequences82 there may be a threonine residue,

as in human kidney AO (Thr369). In this case, the lysine residue that flanks the threonine

(Lys370) could react with TPQ. This would further substantiate the importance of a

lysine residue in the active site for the formation of the radical semiquinolamine species

in plant enzymes and of the aminoquinol in PKAO due to Xenon inclusion. For what

concerns BSAO, no lysine residue can be found close to TPQ and only arginine Arg382

would be able to react with TPQ cofactor such as to mimic the amine substrate. However,

as the arginine residue in BSAO possesses a highly basic guanidine group, it is expected

to be unreactive with TPQ under Xenon pressure.

A nucleophilic residue has been shown with certainty to be involved in the

inhibition mechanism of AOs during the oxidation of 1,4-diamino-2-butyne (DABY)83-87

and other selective AO inhibitors have been tested. It has been demonstrated that the

product of the reaction between plant AOs and DABY is a very reactive aminoallene

given by DABY oxidation. This species reacts with an essential nucleophilic group at the

enzyme active site, forming a covalently bound pyrrole and producing an inactive

enzyme

The possible relevance of a lysine residue has been suggested84,87, but no

experimental evidence has been presented so far.

The structure–function of an enzyme can be successfully studied finding specific

inhibitors and following their effects on the catalytic process. Our interest in the present

study has been the mechanism-based inhibitor 2-butyne-1,4-diamine (DABY), for

basically two reasons: Firstly, the inhibitor has been found to be a suicide substrate for

plant copper AO (Cu-AO) from pea seedlings84 and grass pea85, for mammalian AOs

from pig kidney81 and from beef serum87; secondly, it has been postulated that the

irreversible inhibition of all the enzymes mentioned involves an intermediate

aminoallenic compound that forms covalently bound pyrrole in the reaction with a

nucleophile at the active site.


113

The exact mechanism of inhibition is unclear, and it was only in grass pea AO

that the involved nucleophile was identified as Glu113, a residue corresponding to a

Lys113 in PSAO84. DABY was also shown to be a mechanism-based inactivator for

native LSAO and ELAO. All the Xenon-treated AOs were inactivated by the reaction

with DABY, clearly indicating that the lysine residue involved in the reduction of TPQ

under Xenon pressure is not the nucleophilic residue involved in the DABY inhibition

mechanism; i.e. the reactive turnover product of DABY binds an amino acid residue

without interfering with the TPQ function.

3.2.6. - Copper containing Amine Oxidases: CONCLUSION

As it clearly appears from our results, Xenon is capable of inducing a structural

modification in a number of Amine Oxidases, such that most of them react with one of

their own lysine residues.

As previously reported, changes in active site architecture and charge distribution

seem to be critical during catalysis in AOs. Even relatively subtle conformational

changes at the active site may significantly alterate the biological function of proteins and

enzymes. These experimental observations represent a rare demonstration of Xe-induced

chemical reactions and suggest that using Xenon as a biomolecular probe must be

carefully evaluated for each particular system.

In particular, we have demonstrated, for the first time, the formation of two important

enzyme intermediates:

• the radical species in absence of a substrate in native LSAO, PSAO, ELAO and

PKAO enzymes, and

• the quinolaldimine in the Copper-free LSAO enzyme.

Moreover, hints on the role of Copper ions in the catalytic mechanism of several AOs

have been proposed and substantiated: while in BSAO Copper has a structural role

interfering in the electrostatic stabilization of oxygen, and the CuI-semiquinone state is

off the reaction pathway. The results obtained rule out the possibility that in plant amine


114

oxidase Copper has just a structural role but it is certainly involved in amine oxidation

process and the CuI/semiquinolamine radical represents the highly reactive species with

Oxygen. This confirms the possibility that more than one oxidative reaction pathway is

available in AOs from different sources.

Finally we propose the role of a lysine residue that seems to play an important

role in assisting enzyme catalysis.

The TPQ/semiquinolamine radical represents the highly reactive species with the

Oxygen molecule in the catalytic cycle of plant AOs. Thus, the radical species observed

under 10 atm of Xenon without a substrate in plant AOs, and the fact that in these

systems always a lysine residue was identified at the active site, revealed key aspects of

the structure–function relationship among the various AOs studied. The transformation of

a lysine residue, most likely Lys296, into allysine, together with a residue identified in a

conserved aspartate residue (Asp300), could play an important role in the selectivity of

AOs towards substrates with a positively charged amino group.

In conclusion, although the data reported in the present article may well be valid

generally, the exact location and nature of the observed interactions between Xenon and

the enzymes studied remain somewhat hypothetical due to the size and complexity of

these biological macromolecules which hamper the obtainment of site-specific

information.

Further comparative investigation of the active site in AOs from plants, mammals

and bacteria would be helpful in understanding whether these enzymes, which differ in

structure and action mechanism, follow a similar metabolic pathway.


115

3.3 – Microporous Crystalline Dipeptides

3.3.1 - Variable Temperature continuous flow HP 129Xe NMR: General spectral

features.

In Figure 3.25 are shown, as a typical example of the experimental results, CF HP 129Xe NMR full spectra of VA dipeptides acquired at different temperatures.

Three signals, corresponding to Xenon in three different environments, can be

observed in the high temperature region: the narrow signal at 0 ppm represents the most

shielded Xenon atoms of the gas-stream, which are not adsorbed on the dipeptides

powder; the broader band, partially overlapping the gas-phase signal and slightly shifted

towards lower fields (centered at about 5 ppm at 343 K in VA), can be ascribed to Xenon

adsorbed on the crystal surface and in intercrystallite regions; finally, the signal resulting

from highly deshielded Xenon absorbed inside the nanochannels is shifted to about 120

ppm at 343 K from the gas-phase signal and shows a pronounced axially symmetric

chemical shift anisotropy (CSA) line shape.

Figure 3.25: The experimental continuous-flow hyperpolarized 129Xe NMR spectra of VA dipeptides at

variable temperature.


116

When the temperature is decreased, together with complex changes in the relative

intensities of the signals, modifications of the lineshapes and chemical shifts are observed

for the NMR signals pertaining to the Xenon atoms inside the nanochannels and adsorbed

in intercrystallite regions. More precisely, a broadening, together with its slight shift to

lower fields and a progressive decrease in intensity which leads to its disappearance is

observed for the intercrystallite signal. A global downfield shift and an inversion of the

anisotropy is generally detected for the powder pattern relative to the Xenon inside the

channels; the span of this signal decreases with decreasing temperature and the sign of

the anisotropy is finally inverted at the lowest temperature analyzed (173 K), while an

isotropic line shape can be observed at intermediate temperature.

These general spectral features are common to all the dipeptides studied in this

work and they are in good qualitative agreement with the literature concerning 129Xe

NMR of nanochannels 13,88-96. However, a more detailed analysis of the spectra reveals

subtle but fundamental differences concerning Xenon NMR results in nanochannels

having different structure.

It is worth noting that only the dipeptides AV and VA show the intercrystallite (or

surface) adsorption signal, which in AV is still observable even at 173 K and it overlaps

the anisotropic signal in the temperature range between 243 K and 263 K. AI also shows

intercrystallyte adsorption upon aging of the sample. 129Xe NMR full spectra concerning the other dipeptides, obtained at variable

temperature and continuous flow of hyperpolarized Xenon, are collected in Figure 3.26.

They are in qualitative agreement with other NMR studies on systems where Xenon is

absorbed in nanochannels. At the same time, each dipeptide reveals its own characteristic

spectra, primarily because of the sensitivity of the 129Xe NMR parameters to the

distinctive size and unique geometry of its nanochannels.


117

AI


118

Figure 3.26: Continuous-flow HP 129Xe NMR spectra of AI, AV, LS, VV, IA, IV and VI dipeptides in the

173-343 K temperature range


119

3.3.2. - Temperature dependence of the 129Xe NMR Chemical Shift Anisotropy

(CSA) tensor.

When an atom is placed in a strong magnetic field B0, its nucleus is subjected to

two magnetic fields, B0 and B1.

B1 is a local magnetic field generated by the circulation of the electrons that

surround the nucleus with respect to the direction of B0. The effective field at the nucleus

(Beff) is usually expressed as

Beff = B0 – B1 = B0(1-σ) [3.15]

The shielding σ experienced by a nucleus is the result of how the electron

distribution of the atom to which it belongs is influenced by the environment. The name

shielding basically derives from the fact that generally the magnetic field generated by

electrons is such as to oppose the external principal magnetic field B0.

Depending on the particular electronic and magnetic environment, a particular

shielding value characterizes the observed nuclei and the difference in shielding with

respect to a reference molecule or system is known as chemical shift. Thus, in NMR, the

observable directly related to the shielding is the chemical shift δ.

Although rapid molecular reorientation (especially in solution) often leads to the

observation of only one averaged chemical shift value for a given nucleus in a molecule,

the shielding is anisotropic. Thus, σ is well described by a second rank tensor (3×3

matrix), which expresses the dependency of the shielding on the orientation of the

nucleus with respect to the direction of B0. Three principal components of the tensor,

referred to as σ11, σ22 and σ33, describe the shielding experienced by the detected nuclei

along three different directions with respect to the external magnetic field B0.

In static single crystals97, where all the nuclei have a fixed orientation with respect

to the principal magnetic field, each orientation of the observed nucleus with respect to

B0 can be observed as a narrow line in a NMR spectrum. In polycrystalline samples,

however, all the orientations are possible and the NMR signal broadens, giving rise

sometimes to featureless bands and sometimes to the observation of a so called powder

pattern, where the values of σ11, σ22 and σ33 can be precisely correlated to the


120

singularities of the signal line shape. In a powder pattern, the signals having chemical

shifts relative to each molecular orientation overlap with each other, the intensities being

proportional to the number of atoms/molecules oriented in each particular direction with

respect to B0.

An anisotropic line shape characterizes the NMR signal of the Xenon atoms

trapped within some confined spaces, such as clathrates and a number of one-dimensional

channels 11,13,94,95,98-104.

This feature bears within itself useful information on the dynamic averaging

experienced by the Xenon atoms and on the size and geometry of the environments in

which they are encapsulated, number of other sorbate molecules and/or atoms within the

same cavity as Xe atom. In this respect, the extensive experimental work done on 129Xe

NMR of clathrates11,105,106, studies on a variety of one-dimensional channels 13,88-96, as

well as theoretical calculations 89,107-110, have pointed out the sensitivity of the 129Xe

chemical shift tensor to the shape, symmetry, size and chemical composition of the cages

and channels in which the Xenon is enclosed.

The number of singularities of the line shapes observed in 129Xe NMR spectra for

Xenon adsorbed in polycrystalline nanoporous materials has been found to be strictly

correlated to the aspect ratio (symmetry) of the cross section of the one-dimensional

channel and the changes of the line shapes, i.e. the behavior of the principal components

σ33 σ11 σ22


121

of the chemical shift tensor as a function of loading provide further information on the

arrangement of the Xenon atoms within the channels and on the existence of energy-

favorable sites where the Xenon fit in registery94,103,108,109. A number of theories have

been suggested to account for the anisotropic line shapes in 129Xe spectra of confined Xe,

some of which are briefly analyzed in ref. 107 [see bibliography, Section 3.4].

One of the major contribution to the interpretation of observed anisotropic NMR

line shapes of Xenon in nanochannels has been provided in the recent years by the

combination of ab initio calculations of the 129Xe shielding surfaces107 and Grand

Canonical Monte Carlo simulations in a process based on the so called additive dimer

tensor model110. Such method has been by now successfully applied to different

systems89,108,109, and it has been demonstrated that it is able to provide results reasonably

close to the experimental observations.

The availability of new experimental results concerning similar porous structures,

possibly varying in channel shapes and diameters, would be desirable in order to confirm

the picture drawn by theoretical calculations and clarify the basic rules governing the

observed anisotropy. Even though the relationship between Xe shielding tensors and

pore size and shape is well understood for small pore systems, a general approach to the

actual derivation of useful parameters from NMR spectroscopy has not been

demonstrated. The establishment of a consistent relationship between 129Xe NMR

parameters and properties of sorbent-Xenon systems, both physical (macroscopic) and

structural (molecular level), relies on detailed studies on a series of similar sorbents,

where a property changes in small increments over a wide range111-113.

In the presence of an axially symmetric CSA, only two principal values of the

shielding tensor are generally reported, being two of them coincident. In particular, in the

ideal case of a Xenon atom adsorbed in nanochannels with cylindrical cross-section in the

zero-loading limit, the shielding tensor is axially symmetric, with components σ11 ≤ σ22 =

σ33. Generally, in this condition, it is indicated σ11 = σ and σ22 = σ33 = σ⊥; in this case,

the component of the 129Xe NMR chemical shift tensor parallel to the axis of the channel

(σ) is more deshielded with respect to the other two components perpendicular to the

channel walls (σ⊥), which in turn show the same shielding value107.


122

As the chemical shift δ and the effective field Beff are directly proportional, the

directly observed NMR chemical shifts are often discussed instead of the shieldings. In

particular, the NMR anisotropic signals are completely described by three parameters: the

isotropic chemical shift δiso = (δ11 + δ22 + δ33)/3, the span Ω =δ11 - δ33 and the skew κ =

3(δ22 + δiso)/Ω. In the following, the components δ⊥ and δ relative to the directions

perpendicular and parallel with respect to both B0 and the channels axis are described.

In this thesis we will discuss the experimental chemical shifts values δ and δ⊥

correspondent to the shielding just mentioned.

Different behaviors of δ and δ⊥ as a function of the temperature correspond to

different dipeptides, the dissimilarities obviously being related to each particular

electronic environment sensed by the Xenon atoms inside each nanochannel.

In order to more clearly describe and understand the results just mentioned, it

would be useful to very briefly review at least some of the most important concepts so far

drawn by theoretical studies, in particular those obtained by ab-initio calculations

performed by C.J. Jameson and coworkers regarding how Xe shielding tensor

components are influenced by interatomic and intermolecular interactions with other

atoms/molecules. These results revealed the following general behavior: as the

intermolecular partner gets closer to Xe, the component of the Xe shielding tensor along

σ

σ σiso


123

the direction of approach changes slightly, increasing shielding. At the same time, the

tensor components in the plane perpendicular to the direction of approach become

uniformly less shielded.

In all cases of Xe interacting with an atom or approaching collinearly to a linear

molecule, the components of the Xe shielding tensor in the plane perpendicular to the

direction of approach are uniformly deshielded, whereas the tensor component along the

direction of approach is changed only slightly.

For Xe interacting with a wall of atoms, all components are deshielded relative to

the free isolated Xe atom, with the component parallel to the wall being the most

deshielded. The least deshielding occurs for the component along the line of approach of

the Xe to the wall.

Aiming at simulating the case of two or more Xenon atoms sorbed into a

nanochannel at high loading, ab initio calculation of the Xe shielding tensor in the

presence of of two more Xenon atoms have been performed by Jameson and coworkers,

and results showed to be comparable to experimental studies107,108. The shielding tensor

of the central Xe in a Xe3 trimer is very representative of what is actually observed when

the tensor is dominated by Xe-Xe interactions and Xe-wall interactions are negligible in

highly loaded channels. This work demonstrated that the angle formed by the three Xe

atoms strongly influences the shielding tensor and therefore the resulting observed line

shape of 129Xe NMR signals.

We will make use of these results in understanding the Xe shielding in

nanochannels under full loading.

It is now clear that important qualitative information concerning the interaction

between both the Xenon atoms and the channel walls (Xe-wall) and between neighbors

sorbed Xenon atoms with respect to each other (Xe-Xe) can be gained by analyzing the

chemical shift tensor. In particular details will be given here on how hints on the role of

each particular channel structure in differentiating the adsorption of Xenon and

consequently the observed 129Xe NMR spectra can be found and discussed.

The observed line shapes, moreover, are influenced by the dynamics of the Xenon

atoms sorbed in the pores, which are evident in the zero-loading limit and are correlated


124

with the channel size. These evidences are discussed in some more detail in the

following, and divided in four short paragraphs.

3.3.2.1. Effect of channels loading on the CSA

The behavior of the components of chemical shift anisotropy tensor as a function

of channel loading has been observed and described for different nanochannels so far89-

91,93,94,96,107,111 .

According to the calculations performed by Jameson et al., in a diamagnetic

narrow-bore pipe, the chemical shift anisotropy should change as a function of channel

loading from positive values at zero loading, with the component δ⊥ going from being the

most shielded, to negative values at high loadings, the δ⊥ being in this case the most

deshielded component and the component δ remaining unaffected by the filling of the

channels.

These trends are generally observed for all the dipeptides analyzed and prove the

sensitivity of the line shapes to the loading of a number of different channels.

It is interesting to observe, however, that in situations where Xenon is in channels

with a diameter narrower than its Van der Waals diameter (VI, IV, IA and VV, which

have diameters that range from 4.0 Å for VV to 3.0 Å for VI compared to the Xenon’s

Van der Waals diameter of about 4.3 Å), a significantly large positive anisotropy (~100

ppm for VI at room temperature) characterize the signals at high temperatures (low

loading) and a positive anisotropy is still observed also at the lowest temperature (high

loading), although with reduced span compared to the high-temperature signal (an

isotropic line shape is observed for VV at 173K). This observation is consistent with the

presence of very strong Xe-wall interactions that most likely characterize the Xenon

atoms tightly wrapped by these very narrow, although flexible, channels and with Xe-Xe

interactions being unable to balance these effects. It is also reasonable to believe that in

the case of the narrowest channels, the Xe-Xe interactions could be somewhat hampered

due to the constitution of the pores, that can be thought as formed by an alternation of

wide regions and constrictions in a corrugated channel. This hypothesis is furthermore

discussed and substantiated in the next paragraph.


125

3.3.2.2. Presence of specific sites (niches)

Distinctive features characterize the observed CSA in LS, VI and IV with respect

to the other channels. In the high temperature region, all three show a linear behavior of

each tensor component with decreasing temperature. The narrowest channels VI and IV

show similar behavior to LS in that the line shapes do not change in the high-temperature

range (which is up to about 220 K in VI and IV and about 260 K for LS) showing, in fact,

a constant span: nevertheless, LS differs from VI and IV because, decreasing the

temperature, the whole signal is shifted toward high frequencies in the latter two while it

shifts to lower frequencies in the former. Such a linear dependence of the 129Xe chemical

shift tensor components with average occupancy has been observed94 already in

molecular sieves ALPO-11 and SAPO-11 and ascribed to the presence of an ordered

arrangement of Xenon atoms within the channels driven by the existence of energetically

favorable sites along the channel walls where the adsorbed atoms fit. It was recently

pointed out that when these niches are too close to each other no linear behavior of the

chemical shift tensor components should not be observed109. It is interesting to note,

however, that the linear behavior of the tensor components observed in VI and IV is

interrupted at some intermediate temperature and a non-linear behavior is observed at

lower temperatures. This suggests that also the structure of the channel is most likely

changed by the increase of the Xenon loading inside the channels, going from a

corrugated channel to a more smooth one, where the Xe atoms have chance to get closer

to each other, similarly to what observed in the larger channels.

3.3.2.3. Effect of helicity and diameter of the channels on CSA

In order to explain the influence of the helicity of the channels on the observed

spectra, it is useful to consider again the ideal model of a single Xenon atom in a straight

nanochannel. In this case, the presence Xe-wall interactions are merely reflected in the

deshielding of the parallel component of the Xe chemical shift tensor with respect to its

perpendicular component, this latter being in fact influenced only by interactions along

the axis of the channel. This model suggests that varying the loading, which can be

achieved by increasing the Xenon pressure or decreasing the temperature as well, would


126

cause a modification of only the perpendicular component of the tensor leaving

unchanged the parallel one. As we expected, real systems are more complicated than

ideal models. According to our experiments on dipeptides indeed, δ generally increases

with decreasing temperature up to the lowest temperature (173 K) although it shows

almost a constant value for AI. It is worth noting at this point that, according to previous

helium pycnometry measurements and to XRD single-crystal structures88, AI channels

are characterized by having a diameter very close to the Van der Waals diameter of

Xenon atoms and by a low helicity.

Figure 3.27: HP CF 129Xe NMR spectra of AI dipeptides showing the constancy of the parallel component

of the chemical shift tensor at variable temperatures.

The observation of a constant value for δ as a function of experimental

temperature in AI dipeptides is a further confirmation of the results previously obtained

and moreover this finding complements the previous results suggesting that the Xenon

atoms adsorbed inside the AI channels are aligned along the channel axis in a fairly

straight line. Calculations suggested108 that when angular arrangements of Xenon atoms

in highly loaded channels is characterized by angles in the range 180º-150º (i.e., to a good

approximation, in a straight line) the component of the chemical shift tensor along the

axis of the channel (δ) will hardly change with Xenon occupancy.

This, however, is not expected to strictly apply to helical channels, where the

effects of Xe-wall and Xe-Xe interactions are spread out also along the direction

perpendicular to the axis of the channel, thus influencing the observed δ⊥ and to some

extent the δ as well. It is not straightforward, however, to directly correlate the helicity

of the channels to the shift of the parallel component of the 129Xe NMR signal. It is


127

reasonable to think that the deviation from the constancy of δ observed in all the

dipeptides except for AI can be attributed both to the helicity and to the diameter of the

channels. Larger channels than the Xenon Van der Waals diameter should allow zig-zag

conformations of Xenon atoms even in straight highly loaded channels.

3.3.2.4. Dynamics of Xe in the cross section of the pores and CSA of 129Xe NMR signal

Previous calculation based on Lennard-Jones potentials114,115 and ab initio

calculations107, suggested that when Xe atoms are sorbed in narrow pores of the same

size of Xenon atoms, in the limit of zero loading (high temperature region), the atoms

remain at the center of the channels and that 129Xe NMR signal shifts toward lower fields

as a function of increasing temperature. This expected result has been demonstrated by ab

initio calculations by considering the effect that the temperature-dependent dynamics of

the Xenon atoms inside the nanopores would have on the shielding of Xenon. In

particular, an increase in the span of the anisotropic 129Xe NMR signal of adsorbed

Xenons with increasing temperature due to an increasing deshielding of the parallel

component is expected and actually observed in our experiments on almost all dipeptides.

In other words, it is assumed that at higher temperatures the Xenon atoms are supposed to

get closer to the channel walls107. It can be seen that this effect is more or less marked

depending on the diameter of the channel where the Xe are trapped and it will be

considered again later.

The aforementioned theoretical predictions, however, do not apply to the

experimental results for the channels LS, IV and VI which show, at high temperature, a

linear increase of the isotropic chemical shift when the temperature is decreased. The

effects that characterize the observed CSA of these compounds have been discussed

above.


128

3.3.3. - 129Xe NMR isotropic chemical shifts as a function of temperature

Figure 3.28: Variation of the experimental 129Xe NMR isotropic chemical shift (δiso) as a function of

temperature for the eight dipeptides AV (), VA(), AI(), VV(), LS (), IA (), IV(×) and VI (+).

Solid lines represent the fitting of the experimental points according to the thermodynamic model discussed

in the text. The dotted line is the fit to a straight line relative to the linear behavior of δiso observed in IV in

the high temperature region


129

In Figure 3.28 are reported the variations of the 129Xe NMR isotropic chemical

shifts (δiso) pertaining to the anisotropic signal observed for the eight dipeptides studied

as a function of the experimental temperature. The resulting curves clearly resemble

sigmoid functions and each of them can be roughly divided into regions according to the

scheme reported below [see Fig. 3.29]:

Figure 3.29: Schematic representation of the ideal change of isotropic chemical shift for intrachannel Xe

with temperature at isobaric conditions (assuming that Xe-wall interactions are not temperature dependent,

xenon atoms freely enter 1D cylindrical pores and the pore structure remains rigid upon guest inclusion).

The drawings show the cross section of ideal nanochannels at different loadings corresponding to different

temperature regions

In the high-temperature region (A), the concentration of Xenon atoms into the

channels is expected to be very low, thus 129Xe chemical shifts are expected to be merely

characterized by Xe-wall interactions and to correspond to that of a single Xenon atom in

the empty channels.

In the lowest-temperature region (C), the concentration of Xenon atoms in the

nanochannels reaches a maximal, saturated value; the chemical shift δf is defined by both


130

Xe-wall and Xe-Xe interactions and corresponds to 1D close packing of Xenon atoms

inside the channel, with Xe-Xe interactions dominating the observed shielding.

The low-temperature region (B) represents an intermediate situation, where the

concentration of Xenon atoms in the nanochannel and average chemical shift of the

atoms smoothly grow from their minimal to their maximal values.

The experimentally observed behavior of 129Xe isotropic chemical shifts as a

function of decreasing temperature follow different trends for different dipeptides: four of

them (AV, VA, AI and LS) show in the high temperature region a decreasing value, a

minimum being reached at some temperatures between 260 K and 280 K depending on

the particular dipeptide; two (VV and IA) show almost constant chemical shift in this

region and in the last two (IV and VI) the δiso linearly increase.

The behaviors just described are generally in good agreement with what said

above about the relationship existing between the temperature dependence of the

shielding components of Xe in a narrow-bore pipe at zero loading and the dynamics of

the adsorbed Xenon in the cross-sectional plane of the channels. The guidelines provided

by the theoretical works suggest that in the larger channels of AV, VA, AI and LS, the

effect of thermal averaging among different position of Xenon with respect to the center

of the channels should be more pronounced, since the Xe atoms have more chance to get

closer to the channel walls when the temperature is increased rather than what it is

expected in the narrower channels VV and IA, where instead the Xe are most likely more

constricted in the same position corresponding to both the center of the channel cross-

section and to the closest distance to the channel walls. In this sense, the restricted motion

of Xenon inside the narrow channels of VV and IA can explain the observed invariability

of the isotropic chemical shift with respect to the temperature. As a matter of fact, the

observed trends can be classified according to the channel diameters. As it resulted from

He pycnometry88 the sequence from the largest to the narrowest channel is in fact

AV>VA>LS>AI>VV>IA>IV>VI.

Although we should expect that also in this temperature region the loading

slightly increases when lowering the temperature, the chemical shifts are not significantly

influenced by the loading at this stage, because Xenon atoms are still too far to each other

for the Xe-Xe interaction to be sensed, and we can consider to a good approximation that


131

the observed spectra are dominated by only Xe-wall interactions, i.e., the zero-loading

condition is a good approximation for the analysis of the isotropic chemical shift.

In the low-temperature region (B), the experimental chemical shift steeply

increases with decreasing T, showing a distinctive curvature for each particular dipeptide.

An inflection point is observed at some intermediate temperature in all the plots except

those concerning VI, IV and IA.

By lowering the temperature the Xe concentration inside the channel increases

and the average Xe-Xe distance inside the channels decreases. The dependence of the

observed chemical shift upon loading is usually explained in terms of the contribution

due to collisions between Xe atoms in the micropores. At the highest temperatures

analyzed the Xenon atoms are mainly affected by strong Van der Waals interactions with

the channels walls. This zero-loading condition lasts until the loading inside the channels

is such that the absorbed Xenon atoms are able to sense the presence of their neighbors,

that is, when sorbed Xenon atoms get close enough to each other in the nanopore for

allow Xe-Xe interactions to compete with the Xe-wall interactions to dominate the

shielding tensor. Further decrease of the temperature leads to a change in the curvature,

which is most apparent at the lowest temperatures analyzed. It is reasonable to assume

that Xe-Xe interactions, to which has been attributed the steep increase of the chemical

shift would reach a maximum constant value in the completely filled channels, where the

trapped Xenon atoms have reached the closest positions with respect to each other.

At the lowest temperature, the narrow channels IA, VI and IV show different

behavior: no inflection point or changes in curvature are in fact observed in the plots of

isotropic chemical shifts vs temperature. This further suggests that in such very narrow

channels the Xenon atoms are so tightly wrapped by the channel walls that the average

interatomic distance between Xe atoms remains much larger than that in the larger

nanochannels. Similar behavior has been already observed in other one-dimensional

channels91. The variation of chemical shifts in the low-temperature region can be better

understood if we consider that the isotropic chemical shift is the expression of the

probability of collisions between neighboring Xenon atoms, which is of course directly

related to the channel loading. More precisely, as suggested by ab initio calculations of

the intermolecular shielding tensors for pairs of noble gases107,110, the variation of


132

shielding as a function of loading should be considered as a continuous process, being

this dependent upon the variation of the interatomic distances between the observed

Xenon and its neighbors more than on the population of specific neighboring adsorption

sites. In the particular case of Xenon adsorbed in nanochannels, we expect the 129Xe

chemical shift to be affected at each temperature by the average Xe-Xe distance

experienced by each Xenon atom in the chemical shifts NMR time scale, assuming that

the range of temperature considered is such that the influence of Xe-wall interactions on

the observed signal can be neglected. These considerations imply that a smooth variation

should be observed between the lowest chemical shift value related to Xenon in almost

empty channels and the highest one corresponding to Xenon atoms in almost completely

filled channels. This smooth variation is, in fact, what is actually observed.

Nevertheless, the two limiting chemical shift values of Xenon in empty and

completely filled channels cannot be directly related to observed chemical shifts: the

temperature dependence of Xe-wall interactions, in fact, significantly affects the observed

chemical shift of Xe in the empty channels at high temperatures and the possible

condensation of Xenon below 170 K could most likely hinder the continuous flow

measurements and does not allow to study the highest loadings.

3.3.4. - Thermodynamics of adsorption: the Langmuir model

In order to extract quantitative information on the adsorption of Xenon in these

systems, a thermodynamic model based on the observations discussed above can be

suggested.

The solid lines in Figure 3.28 represent the fitting of the experimental points

according to the model explained in the following.

The Xenon adsorption isotherms previously acquired116 for AV and VA were well

approximated by a Langmuir equation, which has the well known general form

PTK

PTKPT

L

Lr ⋅+

⋅=Θ)(1

)(),( [3.16]


133

where satr ΘΘ=Θ is the relative coverage (a fraction of the total sorption sites occupied

by Xenon atoms, 10 <Θ< rel ), KL(T) the temperature dependent Langmuir constant and

P the experimental Xe pressure at equilibrium. The equation [3.16] applies therefore to

isotherms as well as to isobars.

In order to show the dependence of the relative coverage on temperature, equation

[3.16] should be written in a more general form112,117,118:

RTq

RTq

relst

st

ekpT

ekppT

⋅⋅+

⋅⋅=Θ'

'),(

[3.17]

where k' is the temperature-independent portion of the Langmuir constant and qst is the

isosteric heat of sorption already mentioned previously. It should be noted that the square

root of temperature is always a part of the Langmuir117 constant in equation [3.16] but it

is not evident as long as the equation is applied to sorption isotherms (T=const). Equation

[3.17] is quite general as it may be applied to both sorption isotherms and sorption

isobars. In the experiments of this work the partial pressure of Xenon is maintained

constant (0.01 bar), while temperature is varied. Therefore, the coverage in the

experiment changes isobarically.

The Langmuir theory has proven to be a good model to provide useful

information both on the kinetics of adsorption and desorption of gases on crystal

surfaces118 or on the behavior of thermodynamic functions in microporous adsorbents119.

For constant pressure and variable temperature, the equation [3.16] represents an isobar

and its shape is a sigmoid.

In order to relate the observed 129Xe chemical shift to the relative coverage, we

recall the equation proposed by Fraissard and coworkers111:

MXeXesobs δδδδδ +++= −0 [3.18]


134

where δ0 is the reference (gaseous Xenon at zero pressure), δs arises from interactions

with the inner surface of the pores (Xe-wall), δXe-Xe corresponds to interatomic

interactions between adsorbed Xenon atoms and δM accounts for magnetic fields from

paramagnetic ions if they are present in the porous material. The last term δM is zero in

our case and δ0 is zero by convention. As discussed in the previous section, Xe-wall

interactions expressed by δobs define the observed chemical shift in the high-temperature

region. In this region, the Xe-wall interactions depend on temperature due to the

dynamics of the absorbed species over the cross-section of the nanochannel. Therefore,

the variation of δobs in this region is dependent on temperature and is likely to be

independent of loading.

In the low-temperature region, Xe-Xe interactions start to dominate and the only

coverage-dependent term of equation [3.18], δXe-Xe, becomes a significant contributor to

the chemical shift. The chemical shift observed at any particular temperature can be

considered as a sum of the chemical shift defined by the Xe-walls interactions in the

empty nanochannels (δs) and a fraction of the total chemical shift variation ∆δ = δf - δs

[see Figure 3.29] between the chemical shifts corresponding to the completely filled (δf)

and empty (δs) nanochannels:

relsobs Θ⋅∆+= δδδ [3.19]

Previously, such an approach proved to be valid in fitting experimentally

determined Xenon sorption isotherms for zeolites120.

Combining equations [3.17] and [3.19], we obtain the dependence of the observed

chemical shift on pressure and temperature:


135

RTq

RTq

sobsst

st

ekpT

ekp

⋅⋅+

⋅⋅∆+='

'δδδ [3.20]

Under the experimental conditions of this work, p = 0.01 (relative pressure,

dimensionless) and the dependence δobs(T) becomes an isobar, a sigmoid-shaped curve

defined by four parameters: δs, ∆δ, k' and qst. In order to simplify the task of fitting the

experimental data with equation [3.20], δs was found from the extrapolation of the

chemical shifts in the high-temperature region. The value of δs was assumed to be

constant in the lower-temperature region considered (region (B) in Figure 3.29) as it is

relatively narrow and the dynamics of Xenon species responsible for the deviation of the

observed chemical shifts from δs is expected to be suppressed at higher loadings. Fitting

of experimental data with [3.20] is shown in Figure 3.28 (solid lines) and the derived

values of k', qst, δs and ∆δ= δf - δs are reported in Tables 3.6 and 3.7.

It should be mentioned that fitting was impossible for the narrowest channel of

VI. For IA and IV, it was necessary to fix qst to the values obtained from the signal

intensities in high-temperature region (see section 3.2.1) in order to obtain the other

parameters.


136

Table 3.6: Thermodynamic parameters of Xenon sorption in the studied dipeptides derived from the experimental data for the indicated temperature intervals (K): isosteric heats of sorption qst [kJ/(mol of Xe)], k and k' [dimensionless]

Sample Analysis of signal intensities

(high temperatures, low loadings)

Analysis of chemical shifts

(low temperatures, high loadings)

T-rangea k · 103 qst T-rangea k' · 103 qst

343 – 293 (5) 30.8 ± 0.9 8.21 ± 0.08 AV

293 – 273 (3) 1.1 ± 0.2 16.51 ± 0.59 273 – 173 (11) 4.8 ± 1.4 20.80 ± 0.54

333 – 293 (4) 22.4 ± 1.8 12.16 ± 0.23 VA

293 – 253 (5) 8.4 ± 1.0 14.42 ± 0.28 253 – 173 (9) 2.4 ± 0.7 20.96 ± 0.50

LS 333 – 293 (5) 5.9 ± 1.4 20.07 ± 0.72 253 – 173 (9) 6.5 ± 2.5 20.54 ± 0.71

343 – 323 (3) 1.1 ± 0.4 19.00 ± 1.11 AI

323 – 283 (5) 7.4 ± 1.0 14.01 ± 0.37 273 – 183 (10) 8.0 ± 4.3 20.84 ± 1.02

VV 333 – 283 (6) 9.1± 1.2 20.69 ± 0.36 293 – 173 (10) 8.9 ± 2.8 19.81 ± 0.57

IA 333 – 293 (3) 1.8 ± 0.4 17.67 ± 0.65 273 – 173 (6) 5.5 ± 0.5 17.8b

IV 343 – 253 (5) 4.8 ± 1.0 17.76 ± 0.56 213 – 173 (5) 7.5 ± 1.3 17.8b

VI 293 – 193 (6) 2.6 ± 0.4 17.01 ± 0.37 - - - a The number of experimental points used in the fit is given in brackets. b Fixed values


137

Table 3.7: Isotropic chemical shifts (ppm) of absorbed 129Xe at zero loading (δs) and their total variations

(δ) calculated from the experimental data for dipeptides studied

Calculated chemical shift

variation with temperature Sample

δs δ = δf - δs

AV 93.0 63.6

VA 124.1 50.3

LS 166.0 42.0

AI 134.6 53.8

VV 155.0 53.2

IA 196.3 43.8

IV 231.4 46.0

VI - -


138

3.3.5. - Signal intensities

In Figure 3.30 (a) is reported the temperature dependence of the integrated signal

intensities (I=Ichannel/Igas) obtained for VV, as a typical example of what observed in all

dipeptides. In Fig. 3.30 (b) is shown the logarithmic plot of Ichannel/Igas in CF HP 129Xe

NMR for Xenon inside all the dipeptides nanochannels as a function of the inverse

temperature (1/T).

(a)

(b)

Figure 3.30: a) Signal intensities I = Ich/Igas for VV dipeptide as a function of temperature. b) The same

data presented in the ln(I) – 1000/T coordinates and fitting of the data with a straight line (high-temperature

region)


139

As it clearly appears from the observation of Fig. 3.30 (a), a first exponential

increase of Ichannel/Igas at high temperatures is followed by its decrease at low T, a

maximum being observed at some intermediate temperature, which has a different value

for each dipeptide.

This dependence is a result of two opposite effects. On the one hand, the

decreasing temperature results in higher channel loading because of a favorable sorption

enthalpy. On the other hand, the decreasing temperature has a negative kinetic effect,

slowing down the diffusion of Xenon species inside the nanochannels. The condition of

fast diffusion is a critical factor in hyperpolarized 129Xe NMR studies as it guarantees a

continuous replacement of the Xenon nuclei that relax to a thermally polarized state (and

therefore become undetectable) with new hyperpolarized Xenon.

In the high-temperature region, the intensity of the NMR signal is merely

controlled by the thermodynamics of Xenon adsorption in the dipeptide channel and the

ratio Ichannel/Igas is indeed representative of the relative amount of Xenon absorbed.

In the low-temperature region, the kinetic factors become dominant: reduced

diffusion, which is well described by the temperature-dependent diffusion constant D0,

slows down the process of the replacement of relaxed Xenon with hyperpolarized nuclei,

causing a non negligible loss of signal121. Similar behavior has been observed for sorption

of Xenon in tris(o-phenylenedioxy)cyclotriphosphazene (TPP) nanochannels90 and this

might be expected for other sorbents having their micropore space organized in isolated

narrow channels. In order to extract quantitative information on the sorption process from

the intensities of the HP 129Xe NMR signal, therefore, only the high-temperature region,

where the diffusion of Xenon in the nanochannels is fast, should be considered.

Assuming that Xenon can fill a limited number of discrete sorption sites in the

nanochannels, the sorption process can be described by the following equation:

[ ]solid + Xegas = [Xe]solid, [3.21]

where [ ] represents a single sorption site and [Xe] denotes an atom of the sorbate on the

site. The equilibrium (sorption) constant for the process (K) is the ratio of activities of

adsorbed and gas phase Xenon, the ratio being proportional to experimentally measurable


140

ratio of intensities of 129Xe NMR signals from Xenon atoms absorbed in the

nanochannels and Xenon atoms in the gas phase:

IkIIkK probe

gas

channelprobe ⋅=

⋅= [3.22]

where kprobe is a proportional coefficient, the ratio of volumes of the gas phase and

sorbent solid phase in the coil region of the probe. The coefficient depends on how the

sample was loaded in the probe and its precise value is unknown (the estimated value for

highly packed fine powders was kprobe ~ 1). Sets of experiments conducted on the same

sample show that kprobe remains almost constant as long as the same probe in the same

instrumental configuration is used.

In the high temperature region, the exponential growth of intensities is directly

related to the standard free energy of sorption and can be expressed as follows

RT

GeK

∆−= [3.23]

Considering that the adsorption is always driven by the two opposing effects of

the energetic and the entropic terms, the previous equation [3.23] becomes

RTq

RTq

RS

RTG stst

ekeee ⋅=⋅=°∆∆−

0 [3.24]

where ∆So is the standard entropy change in the process [3.21], k0 is a "pre-exponential

coefficient" (entropy term which does not depend on temperature), and qst is the isosteric

heat of sorption122-124 which refers to a particular loading and equals, with opposite sign,

the isosteric sorption enthalpy of process [3.21].


141

Taking logarithms

TR

q

R

S

RT

G st 1⋅+∆=∆− o

[3.25]

which is a straight line with slope Rqst and intercept R

S°∆− .

Combining [3.22], [3.23] and [3.24] produces the dependence of the integral

intensity of the intrachannel Xenon on temperature,

RTq

RTq

probe

stst

ekek

kTI ⋅=⋅= 0)(

[3.26]

where k = k0/kprobe is a temperature-independent term. In coordinates ln(I) – 1/T the

experimental data may be fitted with a straight line according to the linear equation

TR

qkI st 1)ln()ln( ⋅+= [3.27]

with the intercept ln(k) and slope qst/R.

Plots of experimentally determined ln(Ichannel/Igas) vs 1/T, in fact, show for all

dipeptides a linear region at high temperatures, while a decisive deviation from linearity,

due to the contribution of reduced diffusion, can be observed in the low-temperature

region, indicating that even though thermodynamic equilibrium was established, the

method used is not suitable because only a small percentage of the adsorbed Xenon atoms

can be detected .

The fitting of experimental points in the high-temperature region provides values

for isosteric heats of sorption and k (as an estimate of k0) in the observed temperature

intervals, the values being valid for low loadings. Values obtained by fitting are collected


142

in Table 3.6. The correlation factors for the fitted regions are higher than 0.99 indicating

an excellent agreement of the experimental data with the model applied.

Six dipeptides (VV, LS, IA, AI, VI, IV) show only one linear region before they

deviate from the linear behavior, while two of them have more (AV, VA, AI). Solid lines

in the plots reported below correspond to the fitting of the experimental points with

straight lines, which provide standard entropies and heats of adsorption.

Figure 3.31: Experimentally observed ln(I) – 1000/T dependences for AV, VA, LS, AI, IV and VI

dipeptides and fitting of the data in the high-temperature regions with straight lines


143

At this point, some considerations should be made. The presence of more than

one linear region in ln(Ichannel/Igas) vs 1/T plots is observed only in AI, AV and VA, which

are also the only dipeptides between those analyzed that show intercrystallite adsorption.

The exact nature of intercrystallite regions and its actual effect on the adsorption of

Xenon in the channels was quite undefined at this point but it appeared clear from the

comparison of the spectra that the presence of such signal is always associated to a loss of

signal arising from Xenon atoms inside the channels.

Further attempts to characterize this phenomenon suggested that a continuous

aging process takes place in dipeptides crystals, which will be further described in a

separate paragraph in the following.

It must be pointed out here, however, that while signal intensities can be

influenced by the presence of the intercrystallite adsorption due to aging phenomena, the

chemical shifts do not seem to be influenced by this effect to any extent.

3.3.6. - Aging of dipeptide samples

The experiments conducted in this work were complicated by the observation of

irreversible changes in the samples that resulted in an apparent decrease in microporosity.

In particular, the decrease in intensity of the intrachannel Xenon signal was observed

down to its complete disappearance in some cases. Here the changes are referred to as

aging to emphasize their irreversible nature and the fact that the changes progressed

steadily in time. At least three different sources of aging were identified, as discussed

below.

The first type of aging was emblematic particularly in AV, VA and AI. These

three dipeptides, unlike the others, also showed intercrystallite adsorption, with the

appearance of the intercrystallite signal (especially for AI) occurring after new samples

were subjected to a sorption-desorption cycle. The presence or appearance of the

intercrystallite signal was always associated with some loss in intensity of signal from

Xenon residing inside the nanochannels.

These observations are consistent with a stepwise phase transition occurring at a

certain loading and temperature for each of the dipeptides. Indeed, the AV, VA and AI


144

dipeptides show two linear regions on the ln(I) – 1/T plot [Figure 3.31] that indicates the

probable presence of a phase transition in the dipeptide-Xenon inclusion compounds.

Additional experiments revealed a visually evident disintegration of large single crystals

(VA) that had been subjected to Xenon pressure above a certain value.

Further attempts to the characterization of Xe-dipeptides complexes by means of

X-ray crystallography have been so far unsuccessful due to the low degree of crystal

order caused by these phase transitions. Different experimental approaches to the

preparation of Xe derivatives of dipeptides crystals are currently underway.

Although the transitions are reversible, mechanical stresses in different parts of

the crystal, in the vicinity of the transition temperature, cause extensive cracking and

break the crystals. This explains the growth of the intercrystallite Xe signal, to some

extent at the expense of the signal from the intrachannel Xenon as relatively large crystals

are ground into fine crystals or a partially amorphous phase with a highly developed

surface. Such mechanical changes occur every time a sample passes through the phase

transition point and this way the sample "memorizes" how many sorption-desorption

cycles it has experienced.

The second type of aging was observed particularly for VA but might, at some

point, become a curse for any of the dipeptides under study. As reported earlier88,116, VA

in its microporous hexagonal form has a packing coefficient of 0.537(4) which is too low

for a stable molecular crystal, even taking into account stabilization by hydrogen bonds.

Therefore, the existence of a more stable, dense polymorphic form of VA could not be

excluded (note that it is impossible to prove the non-existence of such polymorph). The

problem with the dense form is of kinetic origin: the microporous form exists as a

metastable polymorph for kinetic reasons but starts to disappear upon contact with the

seeds of a stable dense polymorph once they appear in the surroundings. Extensive

experiments with VA samples eventually generated the stable form of the dipeptide and

an almost complete disappearance of microporosity was observed in some samples after

less than two months of work. Some results concerning 129Xe NMR characterization of

sorption in aged VA crystals are shown in Figure 3.32 below.


145

Figure 3.32: 129Xe NMR spectra of as received VA dipeptides crystals and of the same materials after

approximately 3 months aging of the sample. Experiments were done using CF HP 129Xe at a temperature

of 203K and several (∼10) sorption-desorption cicles.

The polymorphic transition appears to be irreversible and occurs for kinetic

reasons; a guest component (Xenon in our case) does not necessarily participate but, as in

other previously described systems125,126, may act, especially in combination with

temperature changes, as a catalyst. The study of the new polymorph is currently

underway.

The third type of aging is an intrinsic property of all the dipeptides studied. The

organization of micropore space in narrow, mutually isolated (non-interconnected)

nanochannels makes such systems subject to easy "poisoning" by the presence of defects

in their crystal structure or low-volatility impurities. The defects may appear as stacking

faults between crystallite domains127 or as molecular-size irregularities. The impurities

may completely block the nanochannels even in very low concentrations. Noteworthy,


146

"poisoning" of zeolite catalysts due to blockage of micropores is one of the major

problems in industry; chabazite128, mordenite128,129, faujasite130 and ferrierite131 were

reported as being subject to this phenomenon. In peptide microporous materials the

problem is rather worse: the impurities cannot be removed by thermal or chemical

treatment because of the limited stability of these materials and the dipeptides can even

produce such impurities themselves as a result of their gradual degradation.

3.3.7. - Thermodynamics of adsorption in nanochannels. General remarks

(summary).

The thermodynamic description of sorption is a primary fundamental account of a

sorbent material which defines its applicability to a desired separation, purification,

storing or catalytic process. On the other hand, the investigation of the relation between

the thermodynamic parameters of sorption and structural parameters of the sorbent-

sorbate components reveals issues defining the energetics and selectivity of sorption and

contribute to the design of new porous materials. The development of numerous synthetic

zeolites and their industrial applications were accompanied by extensive and detailed

sorption studies documented in the literature132.

The thermodynamic parameters of sorption experimentally derived in this study

are summarized in Table 3.6. In general, the sorption of Xenon in all eight dipeptides is

very favorable energetically, as judged from the qst values of 8-21 kJ/(mol of Xe).

Characteristically, at the high loadings, the heats of sorption are similar, while at low

loadings they vary, with the variation being obviously related to the variation in the

sorption entropy. This result likely indicates that at high loadings and low temperatures in

all the channels there is similar dense packing of Xenon atoms having like environments

and low degrees of freedom. This similarity is observed in spite of substantially different

sizes of the channels and may signify the change in channel structure induced by the

Xenon species absorbed.

In contrast, at low loadings and high temperatures different situations may arise,

with various degrees of freedom of the Xenon species in the channel. This possibility is


147

especially well illustrated by stepwise changes in AV, VA and AI dipeptides, implying

possible phase transitions in these dipeptides. From the comparison of thermodynamic

parameters in the two temperature intervals [see Table 3.6], the enthalpies and

approximate temperatures of the phase transitions are 8.3(6), 2.3(4) and -5(1) kJ/(mol of

Xe) and 298, 298 and 323 K for AV, VA and AI, respectively. The entropy terms change

in order to compensate for the contribution of the enthalpy changes to the free energy.

This observation is known as a compensation effect, it having been observed for sorption

in zeolites133. The physical sense of this phenomenon is that a higher ordering of

adsorbed sorbate species in the channel ensures a better interaction with the surroundings

(favorable enthalpy term) but it also implies some loss of freedom (unfavorable entropy

term), while higher mobility creates the opposite effects.

The phase transitions in the dipeptide-Xenon inclusion compounds are likely to

arise from a not very dramatic but stepwise structural adjustment in the dipeptide matrix,

or packing of Xenon atoms inside the nanochannels. As can be seen from the enthalpy

and entropy changes, there are less efficient interactions but more degrees of freedom in

the higher-temperature modifications of AV and VA. Therefore, cooling might induce a

better fit between a Xenon atom and a sorption site and higher degree of guest ordering in

the nanochannels. Phase transitions of this kind have been observed and studied in other

inclusion compounds and may be induced by either temperature134-139. In case of AI, the

increase in entropy and decrease in qst might be a result of a stepwise increase in the

channel diameter that takes place at a certain loading. Such phase transitions are triggered

at a certain "gate pressure" of guest141-145. Possible structural motifs in the lower-

temperature phases may be seen in recent studies showing that the hexagonal structure of

AV transforms into a superstructure with four symmetrically nonequivalent channels

upon inclusion of 2-propanol/water (120 K)146 and the hexagonal structure of VA distorts

to monoclinic in its inclusion compound with acetonitrile/water (105 K)147. In general,

the observed phase changes in the dipeptides studied characterize them as stimuli-

responsive host materials148,149.


148

The heats of sorption [see Table 3.6] either increase or remain almost unchanged

(within experimental error) at low temperatures and high loadings. While the temperature

dependence of isosteric heats of sorption is generally considered to be weak and may be

neglected, the increase of qst with loading is usually ascribed to the presence of strong

lateral interactions between sorbate molecules within the micropores. In general, the

isosteric heat of sorption is defined by a number of factors and previous studies on

sorption in zeolites suggest that it may be a complex function of loading150-155. Other

studies156 signify that deformations of the sorbent may contribute significantly to the

variation of qst, the factor being important for the flexible sorbent materials of this study.

The thermodynamic values derived in this work may be compared with those of

zeolites and other sorbents. Zeolites present host frameworks with much stronger local

charges distributed over the crystal framework that may imply stronger inductive host-

guest interactions. Another major distinction is the rigidity of the zeolite framework. At

the same time, zeolites may reveal similarity to the materials of this work in the topology

of their micropores. Sorption of Xenon in 1D channels of ferrierite having 4.2 x 5.4 Å

cross-section dimensions158 is characterized by an initial (lowest loading) qst of 31.4

kJ/mol.159,160 Sorption of Xenon in the 1D channels of mordenite (6.5 x 7.0 Å) is

characterized by an initial qst of 35.1 kJ/mol.161 Sorption of Xenon in the 3D system of

channels in Linde 5A zeolite (diameter ~5Å) is accompanied by an initial qst of 22.5

kJ/mol which increases to ~25 kJ/mol at higher loading. Sorption of Xenon in 3D system

of channels in silicalite158 (structural type ZSM-5, 5.3 x 5.6 Å and 5.1 x 5.5 Å) is

accompanied by an initial qst of 26.6 kJ/mol.162 From another study163, the sorption of

argon, methane and sulfur hexafluoride in silicalite is characterized with k0 of 50 10-3, 27

10-3 and 3.7 10-3, respectively. The sorption of Xenon in the 3D system of channels in

faujasite158 (NaX, diameter ~7.4 Å) is accompanied by an initial qst of 19.2 kJ/mol and k0

of 7.4 10-3 (zero filling, 193 K)164. In one study, sorption of Xenon in an organo-clay with

~6 Å micropores was reported with a qst of 14 kJ/mol.165 As can be seen from these

comparisons, the materials used in the present study have thermodynamic characteristics

with respect to the sorption of Xenon that are almost in the same range as for zeolites, in

spite of the differences mentioned above. This may well imply that the sorption process is


149

controlled by van der Waals forces to a good approximation and is defined to the greatest

extent by the fit between the Xenon atom and the host cavity. Similar conclusions for

silicalite were made from molecular dynamics simulations166. It also can be speculated

that the advantage of zeolites as sorbents ascribed to the presence of local charges may be

partially compensated by flexibility of dipeptide matrices that can provide better Xenon-

cavity fits by adjusting the pore structure during the course of sorption.

It is interesting to follow the changes in the isosteric heat of sorption with the

diameter of nanochannels in the series studied. As the driving force for the absorption is

the affinity between the guest and the pore walls, pores with dimensions complementary

to the diameter of Xenon might be expected to result in a higher heat of sorption due to

the optimal guest-host contacts106. Figure 3.33 shows qst versus channel diameter (from

He pycnometry88 for the series studied. As becomes apparent from the Figure, the highest

heat of sorption occurs for a channel with a diameter of ~4.4 Å, which is virtually equal

to the diameter of the Xenon atom. It is also interesting that the decrease of the channel

diameter to values lower than 4.4 Å causes only a slight decrease in the sorption energy.

Again, this result must be a consequence of the high flexibility of the dipeptide matrices

where the close contact of the channel with Xe requires some expansion.

In Figure 3.33 are summarized the values derived for enthalpies and entropies of

adsorption as obtained by the analysis of 129Xe NMR isotropic chemical shift (a,c) and

signal intensities (b,d) for all the eight dipeptides studied in this work, plotted as a

function of the channel diameters as previously obtained from He pycnometry analysis88.

It is worth noting that the analysis of the 129Xe NMR signal intensities can provide

reliable results only about the thermodynamic of adsorption in the high temperature

region, due to the effects of reduced diffusivity at temperatures below about 300K while

the fitting of chemical shifts is mostly influenced by the low temperature points, the Xe-

Xe interaction dominating the observed chemical shift only below about 260K.

The knowledge of the physical constants which describe the adsorption

characteristics of an adsorbate-adsorbent system is of fundamental significance in the

design and development of nanoporous materials. Future improvements of the adsorption


150

technology rely on understanding the issues that control the selectivity of the adsorption

process and on elucidating the correlation between energetic and structural parameters.

The analysis of the values reported in Table 3.6 points out the variation of the enthalpies

of adsorption with temperature found in some of the dipeptides studied. VA and AV

show an increase of enthalpy of adsorption with increasing adsorbate loading and/or

decreasing temperature. AI on the other hand shows a first decrease and then an increase

of isosteric heats of adsorption and the results were found to be strongly dependent on the

aging of the sample. While the temperature dependence of isosteric heats of adsorption is

generally considered weak and neglecting this effect is a commonly adopted

approximation124, the increase of qst with increasing loading is sometimes ascribed to the

presence of strong lateral interactions between adsorbate molecules within the micropores

while the decrease of this parameter with decreasing temperature can be ascribed to

heterogeneous guest-host interactions155,167. The results concerning AI dipeptides suggest

that the heterogeneity of adsorption characterize this process at high temperature, while

the effect of Xe-Xe interactions is dominant at low T. A single value in the limits of the

experimental and fitting errors characterizes instead the heats of adsorption of Xenon in

all the other dipeptides studied as a function of the temperature, indicating an almost

perfect balance between the strength of the gas-gas interactions and the heterogeneity of

guest-host interactions. The agreement between the values extracted by the analysis of

the chemical shift and those derived from the fitting of signal intensities confirms the

validity of the thermodynamic model proposed and substantiates the characterization

method. A further comment should be done about the interesting relationship existing

between the enthalpic and entropic terms of adsorption and the diameters of the channels.

As the driving force for the adsorption is the affinity between the guest and the pore

walls, pores with the same size of the guest will have the more favorable energy of

adsorption due to the optimization of guest-host contacts. This is evident from Figure

3.33 a and b, where the qst clearly increase up to the diameter of the Xenon atom (about

4.4 angstrom) and then decrease again for larger diameters.


151

a) b)

Figure 3.33: Heats of sorption qst [kJ/(mol of Xe)] plotted as a function of the pore diameter [Å]

(experimental data from helium pycnometry88. The values for qst correspond to a) high loadings, low

temperatures (from isotropic chemical shifts) and b) low loadings, high temperatures (from signal

intensities). Dashed lines are linear fittings to guide the eye.


152

3.3.8. - Dipeptides: CONCLUSIONS

This study exploits the continuous-flow HP 129Xe NMR technique to extract

comprehensive knowledge on the thermodynamics and molecular-level structural features

of Xenon sorption in a series of microporous dipeptides. In particular, quantitative

thermodynamic parameters, such as isosteric heat of sorption have been determined for

each material with a good accuracy for two compositions corresponding to low and high

loading of the microporous solid with Xenon atoms.

The approach introduced is based on the derivation of sorption isobars under the

dynamic conditions imposed by the continuous-flow HP 129Xe NMR experiments. The

interaction of sorbent with sorbate is ananlyzed with two independent methods based on

the temperature dependence of

1) the intensities

2) isotropic chemical shifts

of the Xenon NMR signal.

The first method is applicable in the high-temperature region, where the loadings

are close to zero. The second method applies to the low-temperature region, where the

loadings are close to maximum. Reasonable agreements between the isosteric heats of

sorption derived by the two methods (the values are expected to be different but not more

than 20-30%) as well as reasonable trends in the values in the series studied confirm the

validity of the thermodynamic model proposed and substantiates the characterization

approach introduced. So far, isobaric determinations have been applied to surface

adsorbents112,118, while the approach illustrated in this study may become a practically

important tool for the express evaluation of a wide range of porous sorbent materials.

As compared to traditional studies based on the determination of a series of

sorption isotherms, the sorption isobar approach introduced is much faster and may be

used, as illustrated well in this study, for characterization of less stable materials or

materials subjected to deterioration progressing in time. Also, the method is much more

applicable to materials with very narrow pores when the equilibrium with the total bulk

of the sorbent is not easy attainable. The possibility to cross-check the data obtained from

signal intensities and from chemical shifts makes it possible to control the adequacy of


153

measurements: the data extracted from the chemical shifts do not depend on the total

amount of the sorbent, while the underestimation of this amount in a sorption isotherm

measurement (for example, due to a non-porous impurity) could essentially ruin the

results. The quantity of the material used was ~70 mg; that was sufficient to collect all

necessary spectra and the material was recovered after the measurements.

Another advantage is the simultaneous availability of information on the

structural properties of the micropore space and sorption events occurring on a

molecular-scale from the spectra. In particular, an estimate of the pore diameter is

instantly available and reasonable assumptions about the pore geometry may be made.

On the contrary to these advantages, one drawback of the approach used is the

lack of information on the absolute value of maximal loading, that is, the capacity of a

sorbent studied with respect to Xenon sorbate. This information should be obtained from

an independent experiment; for example, the value may be calculated from a single

sorption isotherm or estimated from the crystal structure of the sorbent material. For the

dipeptides in the present study, the capacity of AV and VA was calculated as 0.5 mol of

Xe per mol of dipeptide116, while the values for the other studied materials, as estimated

from their crystal structures88 should be in the range 0.33-0.5 mol of Xe per mol of

dipeptide.

Previous work has illustrated the use of the continuous-flow HP 129Xe NMR

technique mostly for qualitative characterization of porous solids88,,102,113,116. Recently

continuous-flow NMR techniques on various nuclei were also used to study catalyzed

chemical reactions in zeolites168-170. This study provides a precedent for the use of

continuous-flow NMR measurements to obtain the comprehensive description of a

sorption process rapidly, providing both quantitative determination of its fundamental

thermodynamic parameters and qualitative understanding of the structural changes and

dynamics accompanying the process on the micro-level.


154

3.4 Bibliography

1. Kyte J & RF Doolittle J. Mol. Biol. (1982), 157, 105-132

2. RF Tilton, ID Kuntz, GA Petsko Biochemistry (1984), 23, 2849-2857

3. Rubin SM, Spence MM, Pines A, Wemmer DE, J Magn Reson (2001), 152, 79-

86

4. Locci, E., Dehouck Y., Casu M., Saba G., Lai A., Luhmer M., Reisse J., Bartik

K. J.Magn. Reson. (2001), 150, 167-174

5. Locci, M. Casu, G. Saba, A. Lai, J. Reisse, K. Batik, ChemPhysChem (2002),

3, 812– 814

6. Rubin, S.M., Spence M.M., Goodson B.M., Wemmer D.E., Pines A. Proc. Natl.

Acad. Sci. USA (2000), 97, 9472-9475

7. Ewing GJ and Maestas S, J Phys Chem (1970), 74, 2341-2344

8. Corda M, Era B, Fais A, Casu M BBA (2004), 1674, 182-192

9. Humphrey W, Dalke A, Schulten K, J. Molec. Graphics (1996), 14, 33-38

10. Ratcliffe, C.. Annu Rep NMR Spectrosc. (1998), 36, 124-208

11. Ripmeester, J.A., Ratcliffe C.I., Tse J.S.. J. Chem. Soc., Faraday Trans. (1988),

84, 3731-3745

12. Bonardet, J., Fraissard J., Gedeon A., Sprinuel-Huet M. Catal. Rev. Sci. Eng.

(1999), 41, 115-225

13. Sozzani, P., Comotti A., Simonutti R., Meersmann T., Logan J.W., Pines A..

Angewandte Chemie (International Edition) (2000), 39, 2695-2698

14. Nguyen, B.D., Xia Z., Yeh D.C., Vyas K., Deaguero H., La Mar G.N.. J. Am.

Chem. Soc. (1999), 121, 208-217

15. Oldfield, T.J., Smerdon S.J., Dauter K., Petratos K., Wilson K.S., Wilkinson

A.J..Biochemistry (1992), 31, 8732-8739

16. Pintacuda, Hohenthanner K, Otting G, Muller N, J Biomol NMR (2003), 27,

115-132

17. Solomon, I. Phys. Rev. (1955), 99, 559-560

18. Bloembergen, N, J. Chem. Phys. (1957), 52, 572-573

19. James, T.L.. Nuclear Magnetic Resonance in Biochemistry. Academic Press,

New York, USA, 1975


155

20. Crull GB, Kennington JW, Garber AR, Ellis PD, Dawson JH, J Biol Chem

(1989), 264, 2649-2655

21. Jacobs RE, Singh J, Vickery LE, Biochemistry (1987), 26, 4541-4545

22. Bertini I, Luchinat C, NMR of Paramagnetic Molecules in Biological Systems,

Benjamin/Cummings Publishing Menlo Park, CA, 1986

23. Banci, L., Bertini, I., and Luchinat, C., Nuclear and Electron Relaxation. VCH;

1991

24. Bertini I., Turano P., Vila A.J., Chem. Rev. (1993), 93, 2833-2932

25. Tilton, R.F.; Kuntz, I.D. Jr.; Biochemistry (1982), 21, 6850-6857

26. Emerson, La Mar, Biochemistry (1990), 29, 1545-1556

27. Wuthrich K, Struct. Bonding (1970), 8, 53-121

28. Shulman RG, Glarum SH, Karplus M J. Mol. Biol. (1971), 57, 93-115

29. La Mar, G. N. in Biological Applications of Magnetic Resonances (Shulman, R.

G., Ed.) pp 305-343, Academic Press, New York; (1979)

30. La Mar, G. N., & Walker, F. A. in The Porphyrins (Dolphin, D., Ed.) Part IV-B,

pp 61-157, Academic Press, New York, (1979)

31. Jesson, J. P. in NMR of Paramagnetic Molecules (La Mar, G. N., Horrocks, W.

D., & Holm, R. H., Eds.) pp 1-52, Academic Press, New York, 1973

32. Horrocks, W. D., Jr. in NMR of Paramagnetic Molecules (La Mar, G. N.,

Horrocks, W. D., Jr., & Holm, R. H., Eds.) pp 127-177, Academic Press, New

York, 1973

33. Horrocks, W. D., Jr., & Greenberg, E. S. Biochim. Biophys. Acta (1973), 322,

38-44

34. Emerson SD, La Mar GN, Biochemistry (1990), 29, 1556-1566

35. MP Byrn, BA Katz, NL Keder, KR Levan, CJ Magurany, KM Miller, JW Pritt,

CE Strouse, J. Am. Chem. Soc., (1983), 105(15), 4916-4922

36. GN La Mar, DB Viscio, KM Smith, WS Caughey, ML Smith J. Am. Chem. Soc.

(1978), 100(26), 8085-8092

37. Bondarenko V, Wang J, Kalish H, Balch AL, La Mar GN, J Biol Inorg Chem

(2005), 10, 283-293


156

38. K Rajarathnam, GN La Mar, ML Chiu, SG Sligar, J. Am. Chem. Soc. (1992),

114(23), 9048-9058

39. Wu Y, Chien YET, Sligar SG, La Mar GN, Biochemistry (1998), 37, 6979-6990



41. Lowery TJ, Doucleff M, Ruiz EJ, Rubin SM, Pines A, Wemmer DE, Protein

Science (2005), 14, 848-855

42. H Desvaux, L Dubois, G Huber, ML Quillin, P Berthault, and BW Matthews J.

Am. Chem. Soc., (2005), 127 (33), 11676 -11683

43. Mulder, F. A. A.; Hon, B.; Muhandiram, D. R.; Dahlquist, F. W.; Kay, L.E.

Biochemistry (2000), 39, 12614-12622

44. Antonini e Brunori Hemoglobin and myoglobin and their reaction with ligands,

1971

45. Traylor TG, Berzinis AP, J. Am. Chem. Soc., (1980), 102, 2844-2846

46. Lee KB, La Mar GN, Manfield KE, Smith KM, Pochapsky TC, Sligar SG,

Biochim Biophys Acta (1993), 1202, 189-199

47. Wu Y, Basti M, Chiancone E, Ascoli F, La Mar GN, (1996), 1298, 261-275

48. DD Klug, MZ Zgierski, JS Tse, Z Liu, JR Kincaid, K Czarnecki, RJ Hemley,

Proc Natl Acad Sci USA (2002), 99, 12526-12530

49. Nunes and B. P. Schoenborn, Mol. Pharmacol. (1973), 9, 835-839

50. RD Stewart, TN Fisher, MJ Hosko, JE Peterson, ED Baretta, HC Dodd; Science

(1972), 176, 295-296

51. Y Shiro, T Iizuka, K Marubayashi, T Ogura, T Kitagawa, S Balasubramanian, S

G. Boxer Biochemistry (1994), 33, 14986-14992

52. Rubin, S.M.; Lee, S.Y.; Ruiz, E.J.; Pines, A.; Wemmer, D.E.; J. Mol. Biol.

(2002), 322, 425–440

53. Rubin, S.M.; Spence, M.M.; Dimitrov, I.E.; Ruiz, E.J.; Pines, A.; Wemmer,

D.E.; J. Am. Chem. Soc. (2001), 123, 8616–8617

54. Rubin, S.M.; Spence, M.M.; Pines, A.; Wemmer, D.E.; J. Magn. Reson. (2001),

152, 79–86

55. Padiglia, A., Medda, R. and Floris, G. Biochem. Int. (1992), 28, 1097-1107


157

56. Pedersen, J.Z., El-Sherbini, S., Finazzi Agro’ , A. and Rotilio, G. Biochemistry

(1992), 31, 8-12

57. Dooley, D.M., McIntire, W.S., McGuirl, M.A., Cote’, C.E. and Bates, J.L. J.

Am. Chem. Soc. (1990), 112, 2782-2789

58. McCracken, J., Peisach, J., Cote, C.E., McGuirl, M.A. and Dooley, D.M. J. Am.

Chem. Soc. (1992), 114, 3715-3720

59. Dooley, D.M., McGuirl, M.A., Brown, D.E., Turowski, P.N., McIntire, W.S.

and Knowles, P.F. Nature (1991), 349, 262-264

60. Medda, R., Padiglia, A., Pedersen, J.Z., Rotilio, G., Finazzi Agro’ , A. and

Floris, G. Biochemistry (1995), 34, 16375-16381

61. Medda, R., Padiglia, A., Bellelli, A., Sarti, P., Santanche’, S., Finazzi Agro’, A.

and Floris, G. Biochem. J. (1998), 332, 431-437

62. A Padiglia, R Medda, JZ Pedersen, A Finazzi Agrò, A Lorrai, B Murgia, G

Floris, J. Biol. Inor. Chem. (1999), 4, 608–613

63. S. Kishishita, T. Okajima, M. Kim, H. Yamaguchi, S. Hirota, S. Suzuki, S.

Kuroda, K. Tanizawa, M. Mure, J. Am. Chem. Soc. (2003), 125, 1041–1055

64. Agostinelli, E.; Morpurgo, L.; Wang, C.; Giartosio, A.; Mondovì, B. Eur. J.

Biochem. (1994), 222, 727–732

65. Q. Su, J.P. Klinman, Biochemistry (1998), 37, 12513–12525

66. A.J. Tipping, M.J. McPherson, J. Biol. Chem. (1995), 270, 16939–16946

67. A.P. Duff, D.M. Trambaiolo, A.E. Cohen, P.J. Ellis, G.A. Juda, E.M. Shepard,

D.B. Langley, D.M. Dooley, H.C. Freeman, J.M. Guss, J. Mol. Biol. (2004),

344, 599–607

68. Lunelli, M.; Di Paolo, M.L.; Biadene, M.; Calderone, V.; Battistutta, R.; Scarpa,

M.; Rigo, A.; Zanotti, G.; J. Mol. Biol. (2005), 346, 991–1004

69. Parson, M.R.; Convery, M.A.; Wilmot, C.M.; Yadav, K.D.S.; Blakely, V.;

Corner, A.S.; Phillips, S.E.V.; McPherson, M.J.; Knowles, P.F.; Structure

(1995), 3, 1171–1184

70. Wilce, M.C.J.; Dooley, D.M.; Freeman, H.C.; Gus,s J.; Matsunami, H.;

McIntire, W.S.; Ruggiero, H.C.; Tanizawa, K.; Yamaguchi, H.; Biochemistry

(1997), 36, 16116–16133


158

71. Li R, Klinman JP, Scott Mathews F, Structure (1998), 6, 293–307

72. Kumar, V.; Dooley, D M.; Freeman, H.C.; Guss, J.M.; Harvey, I.; McGuirl,

M.A.; Wilce, M.C.J.; Zubak, V.M.; Structure (1996), 4, 943–955

73. Reisse, J.; New J. Chem. (1986), 10, 665–672

74. Rossi, A.; Petruzzelli, M.; Finazzi Agrò, A.; FEBS Lett. (1992), 301, 253–257

75. Tipping, A.J.; McPherson, M.J.; J. Biol. Chem. (1995), 270, 16939–16946

76. Turini P, Sabatini S, Befani O, Chimenti F, Casanova C, Riccio PL & Mondovı`

B Anal. Biochem. (1982), 125, 294–298

77. Padiglia A, Medda R, Lorrai A, Paci M, Pedersen JZ, Boffi A, Bellelli A,

Finazzi Agro` A, Floris G Eur J Biochem (2001), 268, 4686–4697

78. McGuirl MA, McCahon CD, McKeown KA, Dooley DM, Plant Physiol (1994),

106, 1205–1211

79. Floris G, Giartosio A, Rinaldi A, Phytochemistry (1983), 22, 1871–1874

80. Padiglia A, Medda R, Lorrai A, Murgia B, Pedersen JZ, Finazzi Agro` A &

Floris G Plant Physiol (1998), 117, 1363–1371

81. Dooley DM, McGuirl MA, Peisach J & McCracken J FEBS Lett (1987), 214,

274–278

82. Novotny WF, Chassande O, Baker M, Lazdunski M & Barbry P J Biol Chem

(1994), 269, 9921–9925

83. Pec P, Frebort I, Eur J Biochem (1992), 209, 661–665;

84. Frebort I, Sebela M, Svendsen I, Hirota S, Endo M, Yamauchi O, Bellelli A,

Lemr K & Pec P Eur J Biochem (2000), 267, 1423–1433

85. Lamplot Z, Sebela M, Malon M, Lenobel R, Lemr K, Havlis J, Pec P, Qiao C &

Sayre LM Eur J Biochem (2004), 271, 4696–4708

86. Padiglia A, Floris G, Longu S, Schinina` ME, Pedersen JZ, Finazzi Agro` A, De

Angelis F & Medda R Biol Chem (2004), 385, 323–329

87. Shepard EM, Smith J, Elmore BO, Kuchar JA, Sayre, LM & Dooley DM Eur J

Biochem (2002), 269, 3645–3658

88. Soldatov DV, Moudrakovski IL, Grachev EV, Ripmeester JA, J. Am. Chem.

Soc. (2006), 128, 6737-6744


159

89. Moudrakovski IL; Soldatov DV; Ripmeester JA; Sears DN; Jameson CJ, Proc

Natl Acad Sci USA (2004), 101, 17924-17929

90. Meersmann T, Logan JW, Simonutti, R Caldarelli S, Comotti A, Sozzani P,

Kaiser LG, Pines A, J. Phys. Chem. A (2000), 104(50), 11665-11670

91. Ueda T, Eguchi T, Nakamura N, Wasylishen RE, J. Phys. Chem. B (2003), 107,

180-185

92. Sears DN; Wasylishen RE; Ueda T; J. Phys. Chem. B (2006), 110, 11120-11127

93. Koskela T, Ylihautala M, Jokisaari J, Micropor. Mesopor. Mater. (2001), 46,

99-110

94. Ripmeester, JA, Ratcliffe CI, J. Phys. Chem. (1995), 99, 619-622

95. Springuel-Huet M.A., Fraissard J. Chem. Phys. Lett. (1989), 154(4), 299

96. Ueda, T.; Kurokawa, K.; Eguchi, T.; Kachi-Terajima, C.; Takamizawa, S. J.

Phys. Chem. (2007), C 111, 1524-1534

97. Terskikh, V.V.; Moudrakovski, I.L.; Du, H.; Ratcliffe, C.I.; Ripmeester, J.A.; J.

Am. Chem. Soc. (2001), 123, 10399-10400

98. C. J. Jameson, A. K. Jameson, R. E. Gerald II, and H. M. Lim, J. Phys. Chem. B

(1997), 101, 8418

99. J. A. Ripmeester and D. W. Davidson, J. Mol. Struct. (1981), 75, 67

100. J. A. Ripmeester, J. Am. Chem. Soc. (1982), 104, 289

101. J. A. Ripmeester and C. I. Ratcliffe, Mater. Res. Soc. Symp. Proc. (1991), 233,

281

102. IL Moudrakovski, A Nossov, S Lang, SR Breeze, C Ratcliffe, B Simard, G

Santyr, and JA Ripmeester, Chem. Mater. (2000), 12, 1181

103. C. J. Jameson and A. C. de Dios, in Nuclear Magnetic Shielding and

Molecular Structure, edited by J. A. Tossell (Kluwer Academic, Dordrecht,

1993), pp. 95–116

104. A. C. de Dios and C. J. Jameson, J. Chem. Phys. (1997), 107, 4253

105. Stueber D., Jameson, C.J. J. Chem Phys. (2004), 120(3), 1560-1571

106. Ripmeester, J.A. Ratcliffe C. I. J. Phys. Chem. (1990), 94(25), 8773-8776

107. Jameson CJ, De Dios AC, J. Chem. Phys. (2002), 116(9), 3805-3821

108. Jameson CJ J. Am. Chem. Soc, (2004), 126, 10450-10456


160

109. Sears, D.N.; Vukovic, L.; Jameson, C.J., J. Chem. Phys. (2006), 125, 114708-

114722

110. Jameson C.J., J. Chem. Phys. (2002), 116, 8912-8929

111. Bonardet J, Fraissard J, Gedeon A, Springuel-Huet M, Catal. Rev. Sci. Eng.

(1999), 41,115-225

112. Terskikh, VV, Mudrakovskii IL, Mastikhin VM, J. Chem. Soc., Faraday

Trans. (1993), 89, 4239-4243

113. Terskikh VV, Moudrakovski IL, Breeze SR, Lang S, Ratcliffe CI, Ripmeester

JA, Sayari A. Langmuir (2002), 18, 5653-5656

114. Ripmeester JA, Ratcliffe CI, J. Phys. Chem. (1990), 94, 7652-7656

115. Cheung TTP, J. Phys. Chem. (1995), 99, 7089-7095

116. Soldatov DV, Moudrakovski IL, Ripmeester JA, Angew. Chem. Int. Ed.

(2004), 43, 6308-6311

117. Langmuir I, J. Am. Chem. Soc. (1918), 40, 1361-1403;

118. Ranke W, Joseph Y, Phys. Chem. Chem. Phys. (2002), 4, 2483-2498

119. Myers, A. L., Colloids and Surfaces A: Physicochem. Eng. Aspects, (2004),

241, 9-14

120. Kato N., Ueda T., Hironori O., Miyakubo K., Eguchi T. Phys. Chem. Chem.

Phys, (2004), 6, 5427-5434

121. Sholl D S, Lee C K, J. Chem. Phys. (2000), 112, 817

122. Sircar, S.; Rao, M.B. (1999) Heat of Adsorption of Pure Gas and

Multicomponent Gas Mixtures on Microporous Adsorbents in Schwartz, J.A.,

Contescu, C., Surfaces of Nanoparticles in Porous Materials, eds. Marcel and

Dekker: New York. Chapter 19, pp 501-528

123. Cao, D.V.; Sircar, S. Adsorption (2001), 7, 73-80

124. Myers, A.L. AIChE Journal (2002), 48(1), 145-160

125. Soldatov DV, Ripmeester JA, Shergina SI, Sokolov IE, Zanina AS, Gromilov

SA, Dyadin YA, J. Am. Chem. Soc. (1999), 121, 4179-4188

126. Soldatov DV, Ripmeester JA, Chem. Eur. J. (2001), 7, 2979-2994

127. Newsam JM, Deem MW, J. Phys. Chem. (1995), 99, 8379-8381

128. Barrer RM, Rees LV, Trans. Faraday Soc. (1954), 50, 989-999


161

129. Mori N.; Nishiyama S.; Tsuruya S.; Masai M., Appl. Catal. (1991), 74, 37-52

130. Flego C.; Kiricsi I.; Parker Jr. WO.; Clerici M.G., Appl. Catal. A (1995), 124,

107-119

131. Van Donk S.; Bus E.; Broersma A.; Bitter J.H.; de Jong K.P., Appl. Catal.

(2002), A 237: 149-159

132. Breck, D.W. Zeolite Molecular Sieves. Structure, Chemistry, and Use. (Wiley,

New York), (1974)

133. Takaishi, T.; Yusa, A.; Amakasu, F. J. Chem. Soc., Faraday Trans. (1971),

67, 3565-3576

134. Soldatov, D.V.; Kolesov, B.A.; Lipkowski, J.; Dyadin, Y.A. J. Struct. Chem.

(1997), 38, 819-828

135. Mayo, S.C.; Welberry, T.R.; Bown, M.; Tarr, A. J. Solid State Chem. (1998),

141, 437-451

136. Soldatov, D.V.; Enright, G.D.; Ripmeester, J.A.; Lipkowski, J.; Ukraintseva,

E.A. J. Supramol. Chem. (2001), 1, 245-251

137. Yeo, L.; Harris, K.D.M.; Kariuki, B.M. J. Solid State Chem. (2001), 156, 16-

25

138. Nishikiori, S.; Takahashi, A.; Ratcliffe, C.I.; Ripmeester, J.A. J. Supramol.

Chem. (2002), 2, 483-496

139. Majda, D.; Makowski, W. Stud. Surf. Sci. Catal. (2005), 158, 1161-1168

140. Takaishi, T.; Tsutsumi, K.; Chubachi, K.; Matsumoto, A. J. Chem. Soc.

Faraday Trans. (1998), 94, 601-608

141. Seki, K. Phys. Chem. Chem. Phys. (2002), 4, 1968-1971

142. Takamizawa, S.; Nakata, E.; Saito, T.; Yokoyama, H. Inorg. Chem. Commun.

(2003), 6, 1326-1328

143. Takamizawa, S.; Saito, T.; Akatsuka, T.; Nakata, E. Inorg. Chem. (2005), 44,

1421-1424, and ref 12f therein

144. Uemura, K.; Kitagawa, S.; Fukui, K.; Saito, K. J. Am. Chem. Soc. (2004), 126,

3817-3828

145. Halder GJ, Kepert CJ, Aust. J. Chem. (2006), 59, 597-604

146. Görbitz CH, Acta Crystallogr. (2002), B58, 849-854


162

147. Görbitz CH, CrystEngComm (2005), 7, 670-673

148. Soldatov DV, J. Inclusion. Phenom. (2004), 48, 3-9

149. Soldatov DV, Stimuli-responsive Supramolecular Solids: Functional Porous

and Inclusion Materials in Urban, M.W., Ed. ACS Symposium Series 912;

American Chemical Society: Washington, Chapter 13, pp 214-231, (2005)

150. Barrer RM, Davies JA, Proc. Roy. Soc. Lond. (1971), A 322, 1-19

151. Thamm H, Zeolites (1987), 7, 341-346

152. Shen D, Bülow M, Micropor. Mesopor. Mater. (1998), 22, 237-249

153. Pulin AL, Fomkin AA, Russ. Chem. Bull. (2004), 53, 1630-1634

154. Fomkin AA, Adsorption (2005), 11, 425-436

155. Chakraborty A, Saha BB, Koyama S, Mg KC, Appl. Phys. Lett. (2006), 89,

171901-171903

156. Jakubov TS, Mainwaring DE, Phys. Chem. Chem. Phys. (2002), 4, 5678-5682

157. Ustinov EA, Do DD, Carbon (2006), 44, 2652-2663

158. Meier, W.M.; Olson, D.H. Atlas of Zeolite Structure Types. (Butterworth-

Heinemann, London), (1992)

159. Takaishi T, Pure Appl. Chem. (1986), 58, 1375-1382

160. Takaishi T, Nonaka K, Okada T, J. Chem. Soc., Faraday Trans.1 (1987), 83,

3317-3329

161. Barrer RM, Peterson DL, Proc. Roy. Soc. Lond. (1964), A 280, 466-485

162. Bülow M, Härtel U, Müller U, Unger KK, Ber. Bunsenges. Phys. Chem.

(1990), 94, 74-76

163. Myers AL, Colloids and Surfaces (2004), A 241, 9-14

164. Kiselev AV, Du PQ, J. Chem. Soc., Faraday Trans. 2 (1981), 77, 1-15

165. Sozzani P, Bracco S, Comotti A, Mauri M, Simonutti R, Valsesia P, Chem.

Commun., (2006), 1921-1923

166. Ermoshin VA, Russ. J. Phys. Chem. (2002), 76, 1127-1131

167. Dunne JA, Mariwala R, Rao M, Sircar S, Gorte RJ, Myers AL, Langmuir

(1996), 12, 5888-5895

168. Wang W, Seiler M, Ivanova II, Sternberg U, Weitkamp J, Hunger M, J. Am.

Chem. Soc. (2002), 124, 7548-7554


163

169. Hunger M, Micropor. Mesopor. Mater. (2005), 82, 241-255

170. Sundaramurthy, V.; Cognec, J.-P.; Thomas, K.; Knott, B.; Engelke, F.;

Fernandez, C. C.R. Chimie, (2006), 9, 459-465

CHAPTER VI – MATERIALS AND METHODS

164

Chapter IV

Materials and methods


165

4.1 - Hyperpolarized Xenon: solving sensitivity problems

Early studies, dated about 1980, first anticipated the usefulness of combining

Xenon as an ideal probe and nuclear magnetic resonance as the spectroscopic technique

of choice for the characterization of void space in porous materials and biomolecules1-4.

Since then, Xenon NMR has been applied to a number of different systems thus

confirming the actual sensitivity of NMR parameters of Xenon to its local physical and

chemical environment.

However, very likely Xenon NMR wouldn’t have been so widespread diffused as

it actually became if it wasn’t for the discovery and development of hyperpolarization

processes that allow enhancing Xenon NMR signal usually by factors of 105-106. The

notorious lack of sensitivity, embedded in the nature of NMR measurements, was in fact

one of the most important drawbacks of this technique and somewhat limited its power of

applicability until when, in 1991, a group of researchers at Berkeley University, led by

Prof. Pines, first applied5 NMR measurements to the characterization of materials

surfaces with hyperpolarized 129Xe.

The possibility of inducing non-equilibrium distribution in electronic spins of

gaseous metal vapors was previously demonstrated6-8 by Prof. Alfred Kastler, who won

the Nobel Prize in Physics in 1966 "for the discovery and development of optical

methods for studying Hertzian resonances in atoms". Kastler’s studies basically showed

that circularly polarized light could be used to pump electronic spin polarization to non-

equilibrium population distributions in metal gases.

Only few decades later it was shown that also nuclear spin polarization of the

noble gases present as buffer gases within the pumping cell could be enhanced by means

of spin exchange induced by collisions with electronically spin-polarized alkali metal

atoms9,10. Happer and coworkers successively further studied and developed the

hyperpolarization process in detail11.

In order to achieve hyperpolarization of Xenon nuclei angular momentum must be

transferred first to electronic and then to nuclear spins thus requiring the simultaneous

presence of different species in the polarizing batch apparatus. In particular, the most

used method for producing optically polarized Xenon is the so-called alkali metal spin

exchange.


166

This method basically consists of two steps: in the first step circularly polarized

laser light polarizes electrons of gaseous alkali metal atoms and in the second step

collisions between unpolarized Xenon atoms and electronically polarized alkali metal

atoms permit spin exchange between electron spins and noble gas nuclear spins. Often

nitrogen is added to the gaseous mixtures in order to quench the fluorescence of the

electronically excited metal atoms that would otherwise depolarize the electron spins.

In more detail the hyperpolarization process can be described as follows:

In the first step, circularly polarized laser light tuned to the D1 transition of

Rubidium atoms (D1 = 794.7 nm) drives their electronic population from mJ = -1/2

sublevel of the ground state into the mJ = +1/2 excited state. Thus electronic spin state of

Rb, 52S1/2, is excited to 52P1/2.

Figure 4.1: Schematic representation of the alkali-metal spin exchange processes.

The population Pel of Rb electrons is determined by the relation

SDOP

OPelP

ρρρ+

= [4.1]

where ρOP is the rate at which Rb is electronically hyperpolarized by means of laser light;

ρSD is the so-called “spin destruction” rate, which basically identifies the spin-relaxation

rate of Rubidium atoms. PRb drops to zero near the cell walls but it permits significant


167

polarization of Xe gas by spin exchange far from the cell walls. Over time (seconds to

hours, depending upon experimental conditions), the nuclear polarization of the noble gas

accumulates, yielding values as high as several tens percent.

In the second step, collision-induced spin exchange is made possible by Fermi-

contact hyperfine interactions between spin polarized electron spins of Rb atoms and

unpolarized Xenon nuclear spins.

The final nuclear polarization Pnu reached in nuclear spins of Xenon after a time t of

optical pumping is given by the equation:

( )[ ]tRb

SE

SE SEePP 010

ρρ

ρρρ +−−⋅⋅

+= [4.2]

where PRb is the electron spin polarization of Rb atoms, ρSE is the rate of spin exchange

between Rb electrons and Xenon nuclei, ρ0 is a factor which takes into account

longitudinal relaxation of noble gas nuclei.

4.1.1. - Continuous-flow measurements.

An apparatus for the production of large quantities of hyperpolarized Xenon was

first proposed by Happer and coworkers12. In the original design, a high power diode

array (130W) coupled to a polarizer produces a circularly polarized light at the proper

wavelength to electronically polarize Rubidium atoms; the gas used was composed by

few hundred torr (∼0.3-0.5atm) of the mixture of Xenon and Nitrogen, while Helium (8-

10 atm) was used as buffer gas. Moreover, in the original experimental setup the

hyperpolarized Xenon was collected into a Dewar of liquid nitrogen and stored. With

careful preparation of storage vessels, hyperpolarized Xe can be maintained for long

periods of time (hours to days). It has been in fact demonstrated, for example, that T1 of 129Xe is ∼3 hours at 77 K (liquid nitrogen) and >100 hours at 4.2 K (liquid helium) when

it is kept under a high magnetic field (> 500 G).


168

The experimental setup used at National Research Council (NRC-SIMS) in

Ottawa, Canada, is based on the just described apparatus and uses a continuous flow of a

gas mixture of Xenon (∼1%), Nitrogen (3%) and Helium (98%) and is therefore

commonly referred to as “continuous flow” device.

Here, two cells are connected by a tube to the gas cylinder where the gas mixture

is kept under pressure: the first cell contains Rb (Rb reservoir) vapors and in the second,

the so called “pumping cell”, spin exchange between Rb and flowing Xe atoms takes

place. Both cells need to be heated to about 150 °C during the polarization process. The

pumping cell is placed in the fringe field (8-20mT) that is present outside the

superconducting magnet, which is exploited in order to split electronic levels of Rb atoms

with different mJ. Two traps are also connected to the gas line in order to avoid presence

of oxygen and water in the pumping cell. At the end of the line, usually just after the coil

region of the NMR probe, the gas mixture is blown off into the atmosphere [see Figure

4.2].

A gas flow of about 0.3-0.5 liters per hour is typically used with the setup just

described. Laser power does not usually need to be too high especially because it could

result in cell overheating and possibly explosions. Additionally, a very high laser power

would be almost unnecessary as reduction in laser power often causes only small

reductions in spin polarization.


169

Figure 4.2: Experimental setup Continuous flow-type for the production of hyperpolarized Xenon installed

at NRC-Canada, SIMS (Ripmeester’s group), Ottawa. Detailed description is given in the text.]

4.1.2. Advantages (and drawbacks) implied in the use of hyperpolarized and

thermally polarized Xenon.

Several important advantages are connected to the use of hyperpolarized gases

beyond sensitivity enhancement.

A very important involvement of developing and improving optical pumping

processes is that when hyperpolarized species are used the polarization of the spins is no

longer dependent on the magnetic field strength they experience, because the

magnetization of the sample is no longer given by the Boltzmann distribution. Therefore

the technique of hyperpolarization allows also measurements of small samples in very

small magnetic fields, which is possibly the major breakthrough of NMR measurements,


170

because of the intrinsic difficulty in the construction of ultra-high field magnets. This, for

instance, has provided the opportunity to develop mobile devices that are able to perform

high resolution Nuclear Magnetic Resonance measurements in the Earth’s magnetic

field13. Mobile, zero-field measurements and squid detectors is therefore greatly favoured

by the use of hyperpolarized species. Moreover, by avoiding the use of large and strong

magnets, the safety, cost, comfort flexibility and portability are improved. The ease of

management is also much enhanced. The use of a low partial pressure of Xenon and high polarizations simultaneously

allows obtaining well resolved spectra in relatively short time and neglecting Xe-Xe

interactions which usually influence spectra obtained by using higher Xe pressures. Thus,

spectra obtained by utilizing this setup are characteristic of the framework or surface of

the systems analyzed14. Polarization transfer by SPINOE is another important tool which

is available only if HP Xenon is used.

It is worthnoting, however, that only thermally polarized Xenon samples the bulk

of materials while fastly relaxing HP atoms can give information of just the surface of a

material. Also, use of hyperpolarized Xenon in systems containing paramagnetic centers

is usually unfruitful because hyperfine couplings between hyperpolarized atoms and

unpaired electrons act as an additional significant relaxation mechanism.


171

4.2 Myoglobins.

Myoglobins (wild-type) were isolated and purified according to the Wittenberg

and Wittenberg method15 with some modifications. Fresh or frozen animal heart tissue

was homogenized in 10 mM Tris–HCl pH 8.0 and differentially precipitated with

ammonium sulfate (65%, 75% and 90%). The precipitated myoglobin was dissolved in

0.02 M Tris–HCl pH 8.2 plus 1 mM EDTA and passed in a column (5±50 cm) of

Sephadex G-100 equilibrated with the same buffer. The solution of myoglobin was

completely purified by passing it on a column (2.5_25 cm) of DEAE-cellulose

equilibrated with 0.02 M Tris–HCl pH 8.4 containing 1 mM EDTA. Pig MMb was

obtained with a very small excess of potassium ferricyanide followed by exhaustive

dialysis against water to remove the excess of salt. The purity of both pig and horse MMb

was checked by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)

under reducing conditions according to the classical procedure16.

The CNMb solutions were prepared from the ferric derivatives (MMb) by adding

the sodium cyanide (Sigma) in large excess over the heme. After the reaction, the excess

cyanide sodium was removed by mild dialysis with H2O/NaOH pH 8.0. The UV-visible

spectra of aqueous CNMb samples, in the absence of Xenon, show the typical features of

a low-spin hemoprotein17 with absorption maxima at 422 and 540 nm.

Xenon gas at natural isotope abundance (purity of 99.99%) was purchased from

SIAD (Italy), Xenon (90.9% 129Xe, 99.95% purity) from Chemical Research 2000, Italy.

Wilmad high-pressure NMR tubes (OD 10 mm and ID 7.1 mm, OD 5 mm and ID 4.2

mm) were used for all measurements.

Before NMR analysis, the Mb solutions were freshly prepared at room

temperature by dissolving the protein in a phosphate buffer 0.01 M, 20% D2O. The exact

protein concentration was determined spectrophotometrically at λmax=280 nm (ε=31000

M-1cm-1). Samples were degassed on a vacuum line without freezing to avoid protein

denaturation. Up to 10 atm of Xenon gas were pressurized into the samples at room

temperature. Before acquiring the NMR spectra, the samples were equilibrated for one

hour.

Carbonmonoxy myoglobins (COMb) were prepared by extensively purging the

solution with CO, followed by reduction with a minimum amount of freshly prepared 1


172

M sodium dithionite solution. Then the solution was eluted by ECONOPAC 10DG

column in a resin to get rid of the sodium dithionite and then again the protein solution

was extensively purged with CO. This solution was lyophilized and then anaerobically

transferred to a dry box filled with argon atmosphere. COMb was diluted in a phosphate

buffer, which was previously degassed using three freeze–pump–thaw cycles. A small

aliquot was then removed to search for protein concentration and oxidation status. The

sample absorbance at 540 and 580 nm for COMb and at 505 and 640 nm for MMb were

used to test the oxidation status. The solution of COMb was then prepared only in the

case of total absence of MMb bands. The COMb solution was prepared in a dry box and

the sample was then transferred to Wilmad high-pressure NMR tubes. Sample was

degassed of argon on a vacuum line without freezing. At the end of each experiment

session, the sample was checked to verify the oxidation status of our protein by UV-vis

absorption spectrum. In the samples, a content of MMb lower than 4% is accepted. 129Xe NMR spectra were performed on a Varian VXR-300 spectrometer at a

resonance frequency of 82.968 MHz. The chemical shift measurements were carried out

at 25.0 ± 0.1°C, using 21 µs pulse (90°), 0.5 s repetition time, and spectral width of 20

kHz. The 129Xe chemical shift of Xenon (1 atm Xenon overpressure) dissolved in

phosphate buffer 0.1 mM, 20% D2O represented our reference. Each experiment run,

consisting of 10 and 12 points, lasted 4 days. Before and after each experimental session,

a buffer solution containing 1 atm of Xenon was run and the 129Xe chemical shift checked

in order to ensure that the measured values were not caused by artefacts.

Titration data were obtained by a non-linear fitting procedure by means of the

Kaleidagraph program; reported errors in binding constants and 129Xe chemical shifts

must be intended as fitting errors. 129Xe NMR spin lattice relaxation times were measured using the inversion

recovery method with an acquisition time of 0.5 s and a recycling delay of 5 s. The

average T1 values were obtained by three-parameters non-linear least square fitting

procedure. The reported errors were estimated from fitting errors. The number of scans

recorded varied from spectrum to spectrum to achieve a good signal to noise ratio (6<

S/N < 10 in the spectra with 1 atm of Xenon overpressure, S/N >10 in all the others).


173

We collected 1H NMR spectra on a Varian Unity-Inova spectrometer at a

resonance frequency of 399.948 MHz and at T = 28 + 0.1 °C. The experiments were

carried out on a 5 mm high-pressure tube using 7 µs pulse (90°), 1 s repetition time, and

spectral width of 100 kHz. Chemical shifts in all spectra are referenced to DSS (2,2-

dimethyl-2-silapentane-5-sulfonate) through the residual solvent signal. CNMb solutions

were ~1 mM in 100 mM sodium phosphate, pH 7.5. Each experiment session,

corresponding to 11 points, took 6 days.

The magnitude COSY spectra18 were acquired in D2O over a spectral window of

20000 Hz using 4096 t2 complex points. We performed 128 scans for each block with a

total of 1024 t1 blocks. The acquisition time was 0.102 ms, and the total acquisition time

~20 h. The residual H2O was suppressed by pre-irradiating the water signal for 0.5 s.

Phase-sensitive NOESY spectra19 were collected in H2O. 10% D2O was added for the

lock, over a spectra window of 20000 Hz using 4096 t2 complex points. Each of the 512

t1 increments was sampled by 128 scans. The mixing time was 50 ms with a repetition

delay of 0.102 ms resulting in a total acquisition time of 12 h. The pre-irradiating

treatment of the water signal was carried out for 0.5 s. Phase-sensitive TOCSY spectra20

were acquired in D2O over a spectra window of 20000 Hz using 4096 t2 complex points;

64 scans were collected for each block with a total of 916 t1 blocks. The spin lock time

used was 50 ms with a recycle time of 0.6 s using the MLEV-17 mixing scheme.

Suppression of the residual H2O signal was obtained by pre-irradiating the water signal

for 0.5 s.

Volumes of the cavities and centres of gravity (CGs) were calculated with

VOIDOO21, which is a program for computing molecular volumes and studying cavities

in macromolecules such as proteins. This program uses an approach where the cavity

volume is defined as the volume swept out by a probe sphere rolling over the surfaces.

The contacts between the probe-sphere and the van der Waals’ protein surface delimit the

probe-occupied cavity, which is similar to the Connolly-type surface22. In addition, all

atoms/residues lining the cavities are identified.

The calculated volumes were sensitive to the inputted parameter values chosen for

the grid size and the probe radius parameters. Different probe sizes, from 1.0 to 1.4 Å,

and grids, from 0.3 to 0.5 Å, were tried. At least five cavity calculations were carried out


174

for each molecule with randomly generated orientations in order to minimize

measurement artefacts. The use of a probe radius of 1.2 Å and a grid of 0.35 Å gave the

most reliable results since the volumes were found to be consistent with those reported23

in the analysis of cavities in sperm whale MMb.

4.3 Microporous Dipeptides

Eight dipeptides (AV (Ala-Val), VA (Val-Ala), AI (Ala-Ile), LS (Leu-Ser), VV

(Val-Val), IA (Ile-Ala), IV (Ile-Val) and VI (Val-Ile)) were purchased from Bachem.

Powdered samples for 129Xe NMR analysis were prepared by grinding as-received

materials followed by drying them overnight at 60ºC to remove moisture from the pores.

LS had to be recrystallized and desolvated prior to preparation, as described previously24.

In addition, the samples were purged with a continuous flow of the Xe-N2-He gas mixture

(BOC, Canada, volume composition Xe:N2:He = 1%:3%:96%) at room temperature for

10-15 minutes in the NMR probe before each NMR experiment. About 70 mg of each

dipeptide was sufficient to obtain a good S/N ratio for all spectra.

All 129Xe NMR spectra were obtained on a Bruker AMX300 spectrometer

operating at 83.013 MHz (magnetic field 7.05 T) using a customized probe from Morris

Instruments. The majority of the experiments were performed using a continuous flow of

hyperpolarized Xe gas, as described previously14. The flow rate was monitored using a

Fathom Technologies flow controller (model GR-116 3-A-PV). The gas flow was set to

0.3 L/hr in order to achieve a good signal to noise ratio and kept constant for each

experiment. A Solid-State Spin-Echo sequence was used to acquire all of the data25 with

90º pulse length of 3 µs and 180º pulse of 6 µs τ1 and τ2 delays were chosen as short as

100 µs and a recycle time of 1 s was used.

The continuous flow of hyperpolarized (CF HP) Xe was delivered to the NMR

coil through a 2 mm plastic tubing. The temperature in the probe was controlled using a

Bruker BT1000 temperature controller with an accuracy of 0.1 K. The variable

temperature experiments were performed with decreasing temperature stepwise from 343

K to 173 K. The sample was allowed to equilibrate at each temperature for 10 minutes

before the corresponding spectrum was collected.


175

The observed intensities, chemical shifts and signal anisotropies were constant for

the same dipeptide at a given temperature and they showed reproducible values over a

series of repeated measurements. Also, the temperature dependent changes upon heating

and cooling were reversible.

The reported 129Xe NMR chemical shifts were referenced to Xenon gas, set to 0

ppm. Analysis of the anisotropic lineshapes was performed using a Bruker simulation

module, Topspin 1.3. Integration of the signals was done using the Bruker processing

program XWinNMR. The relative integral intensity of the signal (I) was calculated as the

ratio of intensity from the signal of Xenon residing in the nanochannels (Ich) and the

combined intensity of Xenon in the gas phase and Xenon adsorbed on the external

particle surface and intercrystallite regions (Igas).

4.4 Copper containing Amine Oxidases

Materials - All reagents were of the highest purity degree available. 1,4-

Diaminobutane dihydrochloride (putrescine), glutamate dehydrogenase, NADH,

phenylhydrazine, α–oxoglutarate acid and 3–aminopropionic acid, 1,5-diaminopentane

dihydrochloride (cadaverine), benzylamine hydrochloride and N,N’-bis(3-aminopropyl)-

1,4-butane diamine tetrahydrochloride (spermine) were purchased from Sigma Aldrich

(St. Louis, MO). Xenon chemical shift measurements were made using 92% enriched 129Xe (Chemical Research 2000; Rome, Italy). Wilmad high–pressure NMR tubes

(Buena, NJ; OD 5 mm and ID 7.1 mm, OD 5 mm and ID 2.2 mm) were used for all

measurements.1,4-Diamino-2-butyne (DABY) was synthesized as previously reported26.

.

Enzymes - Only protein of the highest quality was utilized on the basis of the ratio

TPQ/dimer in the range 1.9-2.1. The concentration of the quinone content was

determined by titration with the carbonyl reagent PHY which gives a hydrazone with a

very high extinction coefficient27 at 445 nm (ε445 = 6.4 × 104 M–1 cm–1). An ε498 of 4.1 ×

103 M-1 cm-1 or an ε278 of 2.45 × 105 M-1 cm-1 for the purified enzyme (2 copper ions and

a Mr = 150000) was used to estimate the enzyme concentration27. Catalytic activity, (kc)

defined as mol of substrate used per mol of active sites in 1 s. Amine oxidases from


176

bovine plasma (BSAO; kc = 0.35 s–1 using benzylamine as substrate)28, pig kidney

(PKAO; kc = 4.5 s–1 using cadaverine as substrate)29, pea seedlings (PSAO; kc = 140 s–1

using putrescine as substrate)30, lentil seedlings (LSAO; kc = 155 s–1 using putrescine as

substrate)31 and Euphorbia characias latex (ELAO; kc = 23 s–1 using putrescine as

substrate)32, were each extracted and purified according to previously described

procedures. Copper–free lentil amine oxidase was prepared as previously described33.

Residual copper, measured by atomic absorption spectroscopy, was 0.2 ± 0.02% of the

original content.

Activities of the tested enzymes were performed according to the procedures

reported in the related references. Oxygen uptake was determined with a Clark–type

electrode coupled to a OXYG1 Hansatech oxygraph (Hansatech Instruments Ltd. King’s

Lynn, Norfolk, England). The temperature of reaction chamber was controlled at 37 °C

using a circulating water bath. The solution (1 ml) containing the enzyme in 1 mM Na+–

phosphate buffer, pH 7.0, was maintained for 20 min at constant level of oxygen as

previously reported34,35 and the reaction was started by addition of the related substrate.

The value of KM for AOs using different substrate concentrations at saturating

concentration of oxygen (219 µM), or varying concentrations of oxygen at a saturating

concentration of substrate, was calculated from initial velocity data fitted to the

Michaelis–Menten equation by nonlinear regression and by double reciprocal plots by

Michaelis–Menten analysis in 1 mM Na+–phosphate buffer, pH 7.0. Benzylamine oxidase

activity was measured in 1 mM Na+–phosphate buffer, pH 7.0, by monitoring the

increase in absorbance of UV-light at 250 nm using an ε250 = 12.8 mM-1 cm-1 for

benzaldehyde36. Catalytic centre activity, (kc) is defined as mol of substrate consumed per

mol of active sites × s–1.

Spectroscopic Methods - UV/Vis Experiments. Absorption spectra of LSAO in 1

mM sodium phosphate buffer, pH 7.0, were recorded at 25 °C with an Ultrospec 2100 pro

spectrophotometer (Biochrom Ltd., Cambridge, England). Anaerobic experiments were

made in a Thunberg–type spectrophotometer cuvette (Soffieria Vetro, Sassari, Italy).

Solutions were subjected to several cycles of evacuation followed by flushing with

Argon.


177

Fluorescence Spectra - Fluorescence spectra were obtained using a Perkin–Elmer

LS–3 spectrofluorimeter (Perkin–Elmer Ltd, Buckinghamshire, UK).

129Xe NMR Experiments - Samples of native lentil AO in 1 mM sodium phosphate

buffer, pH 7.0, 20% D2O, were degassed using three freeze–pump–thaw cycles,

pressurized with Xenon gas into 5 mm Wilmad high–pressure tubes (OD 5 mm and ID

7.1 mm, OD 5 mm and ID 2.2 mm; Buena, NJ), and allowed to equilibrate for 48 hours. 129Xe NMR spectra were recorded on a Varian VXR-300 spectrometer (Palo Alto, CA) at

a resonance frequency of 82.968 MHz. Chemical shift measurements were carried out at

25.0 ± 0.1 °C, using 30° pulse lengths (7 ms), 2 s repetition times and spectral width of

20 kHz. The obtained chemical shifts were referred to the 129Xe chemical shift of Xenon

(8 atm overpressure) dissolved in 1 mM sodium phosphate buffer, pH 7.0, 20% D2O,

used as reference standard. The 129Xe NMR spectrum, due to the noisy signals, allows the

determination of each resonance peak with an estimate accuracy of 0.1 ppm. 129Xe NMR

spin lattice relaxation time of native and Copper–free lentil AO were measured using the

inversion recovery method, with an acquisition time of 1 s and a recycling delay of 3T1.

Assays of Products - Benzylamine oxidase activity was measured in 1 mM

sodium phosphate buffer, pH 7.0, by monitoring the increase in absorbance at 250 nm

using an ε250 = 12.8 mM-1 cm-1 for benzaldehyde37. Ammonia production was checked

from the amount of NADH consumed in the presence of glutamate dehydrogenase and

hydrogen peroxide formation was detected with the peroxidase/4-hydroxy-3-

methoxyphenylacetic acid method38.

α-Aminoadipic-δ-semialdehyde (also 2-amino-6-oxo-hexanoic acid or, more

commonly, allysine) residue was derivatized to a decarboxylated fluoresceinamine

(AASF) and determined by high performance liquid chromatography (HPLC)40,41.


178

4.5 Bibliography

1. Bonardet J, Fraissard J, Gedeon A, Springuel-Huet M, Catal. Rev. Sci. Eng.

(1999), 41,115-225

2. J. A. Ripmeester, J. Am. Chem. Soc. (1982), 104, 289

3. Tilton, R.F.; Kuntz, I.D. Jr.; Biochemistry (1982), 21, 6850-6857

4. Ratcliffe, C.. Annu Rep NMR Spectrosc. (1998), 36, 124-208

5. D. Raftery, H.Long, T. Meersmann, PJ Grandinetti, L. Reven and A. Pines Phys

Rev Lett (1991), 66, 584-587

6. J. Brossel, A. Kastler Compt Rend (1949), 229, 1213-1215

7. A Kastler J Phys Radium. (1950), 11, 255-265

8. A Kastler, J Opt Soc Am. (1957), 47, 460-465

9. MR Bouchiat, TR Carver, CM Varnum Phys Rev Lett (1960), 5, 373-375

10. BC Grover, Phys Rev Lett (1978), 40, 391

11. Walker TG and Happer W, Rev Mod Phys (1997), 69, 629-642

12. B Driehuys, GD Cates, E Miron, K Sauer, DK Walter, and W. Happer Appl. Phys.

Lett. (1996), 69, 1668-1670

13. S. Appelt, F. W. Hasing, H. Kuhn, J. Perlo, B. Blumich Phys Rev Lett (2005), 94,

197602-197604

14. Moudrakovski IL, Nossov A, Lang S, Breeze SR, CI Ratcliffe, Simard B, Santyr

G, Ripmeester JA Chem Mater (2000), 12(5), 1181-1183

15. JB Wittenberg, BA Wittenberg, Methods Enzymol. (1981), 76, 29–33

16. U.K. Laemmli, Nature (1970), 227, 680– 685

17. Smith DW, Williams JP, Biochem. J. (1968), 110, 297-301

18. Bax, A.. Two-dimensional nuclear magnetic resonance in liquids. Delft

University Press, Delft, Holland, 1982

19. Jeener J, Meier BH, Bachmann P, Ernst RR, J. Chem. Phys. (1979), 71, 4546-

4553

20. Davis, D.G. and Bax A.. J. Am. Chem. Soc. (1985), 107, 2820-2821

21. Kleywegt, G.J. and Jones T.A.. Acta Cryst. 1994, D50, 178-185

22. Connolly, M.L. J. Mol. Graphics (1993), 11, 139-143


179

23. Tilton RF Jr, Singh UC, Weiner SJ, Connolly ML, Kuntz ID Jr, Kollman PA,

Max N, Case DA, J. Mol. Biol. (1986), 192, 443-456

24. soldatov JACS2006

25. Kunwar A. C., Turner G. L., Oldfield E., J. Magn. Res. (1986), 69, 124-127

26. Peč P, Frébort I Eur J Biochem (1992), 209, 661–665

27. Padiglia, A.; Medda, R.; Floris, G.; Biochem. Int. (1992), 28, 1097-1107

28. Turini P, Sabatini S, Befani O, Chimenti F, Casanova C, Riccio PL & Mondovì B

Anal Biochem (1982), 125, 294–29831

29. Padiglia A, Medda R, Lorrai A, Paci M, Pedersen JZ, Boffi A, Bellelli A, Finazzi

Agrò A & Floris G Eur J Biochem (2001), 268, 4686–4697

30. Padiglia A, Medda R, Lorrai A, Murgia B, Pedersen JZ, Finazzi Agrò A, Floris G,

Plant Physiol (1998), 117, 1363–1371

31. McGuirl MA, McCahon CD, McKeown KA & Dooley DM Plant Physiol (1994),

106, 1205–1211

32. Floris G, Giartosio A & Rinaldi A Phytochemistry (1983), 22, 1871–1874

33. Rinaldi, A.; Giartosio, A.; Floris, G.; Medda, R.; Finazzi Agrò, A.; Biochem.

Biophys. Res. Commun. (1984), 120, 242–249

34. Schwartz B, Green EL, Sanders-Loehr J & Klinman JP Biochemistry (1998), 37,

16591–16600

35. Novotny WF, Chassande O, Baker M, Lazdunski M & Barbry P J Biol Chem

(1994), 269, 9921–9925

36. Mills SA, Klinman JP, J Am Chem Soc (2000), 122, 9897–9904

37. Neuman R, Hevey R, Abeles RH, J. Biol. Chem. (1975), 250, 6362–6367

38. Janes SM, Klinman JP, Biochemistry (1991), 30, 4599–4605

39. Leyton, G.B.; Pathology (1981), 13, 327–333

40. Medda R, Mura A, Longu S, Anedda R, Padiglia A, Casu M, Floris G Biochimie

(2006), 88, 827–835

41. Medda R, Padiglia A, Bellelli A, Sarti P, Santanchè S, Finazzi Agrò A, Floris G

Biochem J (1998), 332, 431–437 (and references therein)

PAPERS PUBLISHED

180

PAPERS PUBLISHED DURING THE DOCTORATE

1. R. Medda, S. Longu, R. Anedda, A. Padiglia, A. Mura, M. Casu, and G. Floris AN UNEXPECTED FORMATION OF THE SPECTROSCOPIC CuI–SEMIQUINONE RADICAL BY

XENON–INDUCED SELF–CATALYSIS OF A COPPER QUINOPROTEIN

Biochimie, 88 (2006) 827-835.

2. Matteo Ceccarelli, Paolo Ruggerone, Roberto Anedda, Mariano Casu, Antonella Fais, Benedetta Era, Maria Carla Sollaino and Marcella Corda

STRUCTURE-FUNCTION RELATIONSHIP IN A VARIANT HAEMOGLOBIN: A COMBINED

COMPUTATIONAL-EXPERIMENTAL APPROACH

BIOPHYSICAL JOURNAL, 91 (2006) 3529-3541

3. A. Mura, F. Pintus, R. Anedda, M. Casu, A. Padiglia, G. Floris and R. Medda IMPORTANT LYSINE RESIDUE IN COPPER/QUINONE CONTAINING AMINE OXIDASES.

FEBS Journal 274(10), (2007) 2585-2595 4. R. Anedda, C. Cannas, M. Casu, A. Musinu, G. Piccaluga

A TWO-STAGE CITRIC ACID - SOL/GEL SYNTHESIS OF ZnO/SiO2 NANOCOMPOSITES:

STUDY OF PRECURSORS AND FINAL PRODUCTS

Journal Nanoparticle Research, in press (2007) 5. Matteo Ceccarelli, Roberto Anedda, Mariano Casu and Paolo Ruggerone

CO ESCAPE FROM MYOGLOBIN WITH METADYNAMICS SIMULATIONS Proteins - Structure, function and bioinformatics, in press (2007) 6. Fais A, Anedda R, Porcu S, Casu M, Ruggerone P, Ceccarelli M, Sollaino M, Galanello R, Corda M

IDENTIFICATION AND MOLECULAR CHARACTERIZATION OF A NEW DOUBLE VARIANT

HEMOGLOBIN (Hb G-Philadelphia/Duarte α268Asn→Lysβ2

62Ala→Pro) Submitted (2007)

7. R. Anedda, D.V. Soldatov, I.L. Moudrakovski, M. Casu, J.A. Ripmeester A NEW APPROACH TO CHARACTERIZE SORPTION IN MATERIALS WITH FLEXIBLE MICROPORES

Submitted (2007)

8. Long-Li Lai, Chun-Han Wu, Roberto Anedda, Kuang-Lieh Lu, Yu-Shen Wen, Dao-Wen Luo, Kung-Lung Cheng, Zhi Yu, Kui Yu, and John A. Ripmeester

UNUSUAL INCORPORATION OF CD(SCN)64- BY THE VIOLOGEN TEMPLATES VIA

THE HOST-GUEST SELF-ASSEMBLY AND SUBSEQUENT REACTION OF ONE OF THE REPRESENTATIVE ON THE SOLID STATE

Submitted (2007)

PAPERS PUBLISHED

181

9. Roberto Anedda, Benedetta Era, Antonella Fais, Matteo Ceccarelli, Marcella Corda, Mariano Casu and Paolo Ruggerone

129Xe AND 1H NUCLEAR MAGNETIC RESONANCE STUDY OF XENON BINDING TO HYDROPHOBIC CAVITIES OF MYOGLOBINS

Submitted (2007)

Università degli Studi di Cagliari - people.unica.it · Università degli Studi di Cagliari...

Documents

Transcript of Università degli Studi di Cagliari - people.unica.it · Università degli Studi di Cagliari...