Étude thÉorique de la structure et de la dynamique

RICHARD DAIGLE

ÉTUDE THÉORIQUE DE LA

STRUCTURE ET DE LA DYNAMIQUE DE

L’HÉMOGLOBINE TRONQUÉE N

DE Mycobacterium tuberculosis

Thèse présentée

à la Faculté des études supérieures et postdoctorales de l’Université Laval

dans le cadre du programme de doctorat en biochimie

pour l’obtention du grade de Philosophiae doctor (Ph. D.)

DÉPARTEMENT DE BIOCHIMIE, MICROBIOLOGIE et de BIO-INFORMATIQUE

FACULTÉ DES SCIENCES ET DE GÉNIE

UNIVERSITÉ LAVAL

QUÉBEC

2012

© Richard Daigle, 2012

i

Résumé

L’hémoglobine tronquée N de Mycobacterium tuberculosis (TrHbN) protège la respiration

aérobie de Mycobacterium bovis BCG contre l’inhibition causée par le •NO. De plus,

TrHbN catalyse efficacement la dioxygénation du •NO en NO3- (« réaction NOD »,

TrHbN-Fe2+

–O2 + •NO TrHbN-Fe3+

+ NO3-) avec une constante bimoléculaire de

745 µM-1

s-1

à 20°C. Cette haute efficacité, pratiquement limitée que par la vitesse de

diffusion du substrat, a été associée en grande partie à la présence de deux tunnels

hydrophobes visibles dans la structure de tridimensionnelle de TrHbN. L’objectif de cette

thèse s’inscrit dans ce contexte, soit l’étude de la structure et de la dynamique de TrHbN à

l’aide d’outils bio-informatiques, en particulier l’utilisation de simulations de dynamique

moléculaire.

Plusieurs simulations de dynamique moléculaire de TrHbN sous ses formes deoxy,

oxygénée et cyanomet ont été menées. Ces simulations ont permis d’étudier la dynamique

de la chaîne principale, du site actif et en particulier, celle des tunnels. Ces simulations ont

révélé que les tunnels sont dynamiques, davantage complexes que le suggère la structure

cristalline et que ceux-ci prennent place au cœur d’un repliement 2-sur-2 rigide. D’autres

simulations incluant cette fois des molécules de •NO libres ont permis de mettre en

évidence l’utilisation des tunnels de TrHbN par ceux-ci pour diffuser jusqu’au site actif.

Ces simulations ont permis de proposer plusieurs hypothèses sur les routes utilisées et sur

la diffusion des substrats du solvant vers le site actif. Pour valider ces hypothèses et pour

pousser davantage nos connaissances, d’autres simulations couplées à diverses approches

expérimentales ont été employées. D’abord, des simulations de TrHbN sous sa forme

cyanomet couplées à une étude RMN approfondie ont permis de confirmer les résultats de

DM quant à la rigidité du squelette de la protéine. De plus, ces derniers travaux ont révélé i)

des mouvements lents (µs-ms) localisés le long des hélices B et G et ii) que la région pre-A

n’est pas structurée contrairement à ce que suggère la structure cristalline. Enfin, d’autres

simulations et des travaux de cinétiques enzymatiques ont été réalisés sur des mutants avec

tunnel(s) obstrué(s). Ces travaux ont mené à des résultats démontrant que la matrice

ii

enzymatique de TrHbN est très plastique, permettant la diffusion du •NO malgré les

mutations créées. Quoique notre compréhension sur les liens entre la structure et la fonction

de TrHbN soit toujours incomplète, les travaux présentés dans cette thèse constituent un

avancement considérable des connaissances. Plusieurs de nos découvertes mènent à une

meilleure compréhension s’appliquant aux globines en général, aux protéines contenant un

ou plusieurs tunnels et enfin, sur les mécanismes de diffusion des substrats gazeux à

l’intérieur des enzymes.

iii

Abstract

The truncated hemoglobin N from Mycobacterium tuberculosis (TrHbN) protects aerobic

respiration of Mycobacterium bovis BCG cells from the inhibitory effect of •NO. In

addition, TrHbN catalyses the very rapid dioxygenation of •NO into the innocuous NO3-

ions (NOD reaction: TrHbN-Fe2+


+ NO3-) with a bimolecular

rate constant of 745 µM-1

s-1

at 20°C. This high efficiency was largely associated to the

presence of two hydrophobic tunnels visible in the 3D-structure of TrHbN. In this context,

the main goal of this thesis is to study TrHbN structure and dynamics with bioinformatics

tools, especially molecular dynamics simulations.

Several molecular dynamics simulations of TrHbN under its deoxy, oxygenated and

cyanomet forms were conducted. These simulations allowed to study dynamics of TrHbN

backbone, that of the active site and especially, that of the tunnels. As a main result, our

simulations revealed that tunnels are highly dynamics, more complex than anticipated from

the 3D-structure and that they are hosted in a very rigid two-on-two fold. Other

simulations, this time including free •NO molecules, highlighted the use of these tunnels to

reach the buried active site. These simulations allowed to propose many hypotheses

regarding the preferred routes and to propose diffusions mechanisms from the solvent to the

active site. In order to validate our hypotheses and to push further our knowledge on

TrHbN, other simulations coupled with some experimental approaches were performed.

First, simulations on TrHbN under its cyanomet form coupled with a detailed NMR

confirmed that the backbone of the protein is ridig. In addition, this work revealed i) the

presence of µs-ms motions localized along B and G helices and ii) that the pre-A region is

not structured in contrast to the alpha helice seen in the crystal structure. Finally, other

simulations along with kinetics characterizations of obstructed tunnel mutants were

conducted. As a main result, the latter work revealed that TrHbN core is quite plastic,

allowing substrate diffusion despite the presence blocking mutations. Our comprehension

on TrHbN is still incomplete, however the work presented in this thesis constitutes a

considerable progress. Moreover, the work presented herein contributes to other fields of

iv

research, especially on globins, to tunnel-containing proteins and finally, to gaseous

substrates diffusion inside proteins.

v

Avant-Propos

Cette thèse renferme quatre articles publiés dans des journaux scientifiques révisés par les

pairs et pour lesquels je suis premier ou second auteur. Leur incoporation dans cette thèse

de doctorat a été réalisée avec l’appobation des journaux scientifiques respectifs. De plus,

suit ces articles, un dernier chapitre sous la forme d’un manuscrit à être soumis pour

révision par les pairs dans une revue à définir. La contribution de chaque auteur dans la

réalisation de ces travaux ainsi que dans la rédaction est décrite ci-après.

CHAPITRE 5

Ce chapitre présente un article portant sur le rôle des résidus distaux en positions B10 et

E11 et sur l’implication d’une molécule d’eau près du fer dans la forme déoxygénée de

TrHbN. Cet article a été publié sous la référence :

Yannick H. Ouellet, Richard Daigle, Patrick Lagüe, David Dantsker, Mario Milani,

Martino Bolognesi, Joel M. Friedman, and Michel Guertin. Ligand Binding to Truncated

Hemoglobin N from Mycobacterium tuberculosisi is Strongly Modulated by the Interplay

Between the Distal Heme Pocket Residues and Internal Water, 2008, The Journal of

Biological Chemistry, vol. 283, no. 40, pp. 27270 –27278.

J’ai personnellement réalisé les travaux de dynamique moléculaire décrits dans cet article.

Pour ce qui est des travaux en laboratoire, ceux-ci ont été en majeure partie réalisés dans le

laboratoire du Dr. Michel Guertin et effectués par le Dr. Yannick Ouellet. L’équipe du Dr.

Friedman a contribué en effectuant des cinétiques de recombinaison dans des conditions de

haute viscosité. L’équipe du Dr. Martino Bolognesi a contribué à l’interprétation des

résultats via leur expertise sur la structure de TrHbN et celles des mutants à l’étude qu’ils

ont eux-mêmes résolues. J’ai participé en partie à la rédaction de l’article sous la

supervision de Dr. Michel Guertin et Dr. Patrick Lagüe, notamment sur les sections portant

sur mes travaux. La majeure partie de l’article a été rédigée par le Dr. Yannick Ouellet.

CHAPITRE 6

vi

Ce chapitre présente et discute les résultats de simulations de dynamique moléculaire de

TrHbN sous ses formes oxygénées et déoxygénées. La dynamique des tunnels de TrHbN a

été le principal aspect traité dans cet article. Il a été publié sous la référence :

Richard Daigle, Michel Guertin, and Patrick Lagüe, Structural characterization of the

tunnels of Mycobacterium tuberculosis truncated hemoglobin N from molecular dynamics

simulations, 2009, PROTEINS Structure function and bioinformatics, volume 75, pages

735-747.

J’ai effectué la totalité des travaux présentés dans cet article. J’ai rédigé l’article sous la

supervision des Dr. Michel Guertin et Dr. Patrick Lagüe.

CHAPITRE 7

Ce chapitre présente l’étude du rôle des tunnels de TrHbN, sous sa forme oxygénée, dans la

diffusion du •NO entre le site actif et le solvant. Il a été publié sous la référence :

Richard Daigle, Julie-Anne Rousseau, Michel Guertin, and Patrick Lagüe, Theoretical

Investigations of Nitric Oxide Channeling in Mycobacterium tuberculosis Truncated

Hemoglobin N, 2009, Biophysical Journal, volume 97, pages 2967–2977

J’ai réalisé la grande partie des travaux présentés dans cet article. Julie-Anne Rousseau, à

ce moment étudiante au baccalauréat en bio-informatique de l’Université Laval et stagiaire

au laboratoire du Dr. Patrick Lagüe, m’a assisté pour les calculs d’échantillonnage implicite

de ligands. Elle a en particulier aidé au démarrage de ces calculs, à leur analyse et à la

génération de figures pour l’article.

vii

CHAPITRE 8

Ce chapitre présente l’étude de la dynamique de TrHbN sous sa forme cyanomet par

résonance magnétique nucléaire et par simulations de dynamique moléculaire. Ce chapitre

est présenté sous la forme d’un manuscrit soumis et récemment accepté pour publication

dans la revue Biochemistry.

Pierre-Yves Savard, Richard Daigle, Sébastien Morin, Anne Sebilo, Fanny Meindre,

Patrick Lagüe, Stéphane M. Gagné and Michel Guertin, Structure and Dynamics of

Mycobacterium tuberculosis Truncated Hemoglobin N: Insights from NMR Spectroscopy

and Molecular Dynamics Simulations. Biochemistry 2011, volume 50, pages 11121-11130

J’ai personnellement réalisé tous les travaux de dynamique moléculaire présentés dans ce

manuscrit. Les travaux de résonance magnétique nucléaire (RMN) ont été pris en charge

par l’équipe du Dr. Stéphane Gagné. Le Dr. Pierre-Yves Savard a effectué l’ensemble des

travaux portant de l’attribution de spectres RMN et l’étude de la dynamique de la chaîne

principale de TrHbN. Ce dernier a été assisté par l’étudiante stagiaire Fanny Meindre. Le

Dr. Sébastien Morin a réalisé les travaux d’échanges d’amides dans la chaîne principale. La

construction du mutant ΔpreA, la synthèse et purification d’enzymes recombinantes et les

cinétiques enzymatiques ont été réalisées au laboratoire du Dr. Michel Guertin. Ces

derniers travaux ont été réalisés par Anne Sebilo. La rédaction de l’article a été

principalement réalisée par Pierre-Yves Savard, Sébastien Morin et par moi. Cette

rédaction s’est faite sous la supervision de Dr. Michel Guertin, Dr. Patrick Lagüe et de Dr.

Stéphane Gagné.

viii

CHAPITRE 9

Ce chapitre, rédigé en anglais, est présentée sous la forme d’un manuscrit à être

éventuellement soumis. Il contient une étude du rôle des tunnels de TrHbN par la

combinaison d’approches expérimentales et théoriques.

Richard Daigle , Anne Sebilo , Sylvain Lanouette, Julie-Anne Rousseau, Patrick Lagüe,

and Michel Guertin, Experimental and Theoretical Investigations Reveal that

Mycobacterium tuberculosis Truncated Hemoglobin N contains Multiple Diffusion Routes

to Sustain Rapid Gaseous Ligand Entry and Exit. Manuscrit en préparation.

Ces travaux ont été initialement entrepris par Sylvain Lanouette alors étudiant à la maîtrise

sous la supervision de Dr. Michel Guertin et la codirection de Dr. Patrick Lagüe. Nous

avons ensuite repris plusieurs expériences et fait murir plusieurs hypothèses découlant des

travaux de Sylvain Lanouette. J’ai personnellement réalisé la grande majorité des travaux

de dynamique moléculaire et procédé à leur analyse. Tout comme pour l’article présenté au

chapitre 7, Julie-Anne Rousseau a participé pour les calculs d’échantillonnage implicite de

ligands. La partie expérimentale a été menée au laboratoire de Dr. Michel Guertin et les

travaux ont été réalisés par la professionnelle de recherche Anne Sébilo. Moi et Michel

Guertin avons rédigé en grande partie ce manuscrit avec le support de Dr. Patrick Lagüe.

Sylvain Lanouette a également participé à la rédaction.

ix

Pour ma femme, ma fille et ceux et celles qui

m’ont toujours soutenu

x

Table des matières

Résumé ..................................................................................................................................... i

Abstract ................................................................................................................................. iii

Avant-Propos .......................................................................................................................... v

Table des matières .................................................................................................................. x

Liste des tableaux et tables .................................................................................................. xiv

Liste des figures .................................................................................................................... xv

Liste des abréviations ......................................................................................................... xvii

Chapitre 1 Introduction ........................................................................................................... 1

1.1. Contexte biologique de TrHbN ............................................................................... 1

1.1.1. Mécanismes d’infection de Mycobacterium tuberculosis .............................. 1

1.1.2. Mécanismes de résistance de Mtb ................................................................... 2

1.2. TrHbN et la superfamille des globines ................................................................... 3

1.2.1. Définition du terme globine ............................................................................ 3

1.2.2. L’hème ............................................................................................................ 4

1.2.3. Phylogénie des globines .................................................................................. 7

1.2.4. Fonctions des globines .................................................................................. 10

1.2.5. La réaction •NO-dioxygénase ....................................................................... 11

1.2.6. La cavité distale ............................................................................................ 16

1.2.7. Le repliement globine ................................................................................... 19

1.2.8. Cavités internes et tunnels ............................................................................ 22

1.3. Relation structure-fonction chez TrHbN de Mtb .................................................. 28

1.4. Organisation de la présente thèse .......................................................................... 32

Chapitre 2 Étude de la diffusion interne des substrats ......................................................... 34

2.1. Cinétiques enzymatiques de mutants .................................................................... 37

2.1.1. Cinétiques de recombinaison ........................................................................ 37

2.1.2. Cinétiques de liaison ..................................................................................... 40

2.1.3. Exemples d’applications ............................................................................... 41

2.2. Diffraction des rayons X à températures cryogéniques ........................................ 42

2.3. Cristallographie de Laue résolue en temps réel .................................................... 42

2.4. Simulations de dynamique moléculaire ................................................................ 45

Chapitre 3 Méthodologie ..................................................................................................... 47

3.1. La dynamique moléculaire .................................................................................... 47

3.1.1. Histoire de la dynamique moléculaire .......................................................... 47

3.1.2. Principes de base ........................................................................................... 49

3.1.2.1. L’énergie interne ....................................................................................... 51

3.1.2.2. L’énergie externe ...................................................................................... 52

3.1.2.3. Les méthodes de troncations ..................................................................... 53

3.1.2.4. La méthode « Particle Mesh Ewald » ....................................................... 54

3.1.2.5. Conditions périodiques aux frontières ...................................................... 54

3.1.3. Production de la trajectoire ........................................................................... 55

3.2. Simulation des protéines par dynamique moléculaire .......................................... 58

3.2.1. Préparation du système ................................................................................. 58

3.2.2. Initiation et équilibration .............................................................................. 59

3.2.3. Production et analyse de la trajectoire .......................................................... 60

xi

3.3. Étude de la dynamique des tunnels ....................................................................... 60

3.3.1. CAVER ......................................................................................................... 61

3.3.2. Échantillonnage implicite de ligand ............................................................. 62

3.3.3. L’échantillonnage amélioré de ligands ......................................................... 64

Chapitre 4 Objectifs du projet de recherche ......................................................................... 66

4.1.1. Objectif général ............................................................................................. 66

4.1.2. Objectifs spécifiques ..................................................................................... 66

Chapitre 5 Ligand Binding to Hemoglobin N from Mycobacterium tuberculosis is Strongly

Modulated by the Interplay between the Distal Heme Pocket Residues and Internal Water

.............................................................................................................................................. 68

5.1. Résumé .................................................................................................................. 68

5.2. Abstract ................................................................................................................. 69

5.3. Introduction ........................................................................................................... 69

5.4. Experimental procedures ...................................................................................... 71

5.4.1. Mutagenesis, expression and purification ..................................................... 71

5.4.2. Geminate and solvent phase recombination experiments ............................. 72

5.4.3. Molecular dynamics simulations .................................................................. 73

Systems setup ............................................................................................................ 74

5.5. Results and discussion .......................................................................................... 75

5.5.1. Kinetic data indicate that Tyr(B10) mainly contributes to the kinetic barrier

to ligand binding to TrHbN(Fe2+). – (i) O2 and CO binding to TrHbN ...................... 75

5.5.2. O2 and CO binding to TrHbN mutants ......................................................... 75

5.5.3. Geminate and solvent phase recombination ................................................. 77

5.5.4. Molecular dynamics simulations suggest that water may constitute the main

kinetic barrier to ligand binding to TrHbN(Fe2+

) ......................................................... 81

5.6. Conclusions ........................................................................................................... 83

5.7. Footnotes ............................................................................................................... 84

5.8. References ............................................................................................................. 86

Table 5.1 Kinetics constants for the reactions of TrHbN and its mutants with O2 and CO. 89

Chapitre 6 Structural characterization of the tunnels of Mycobacterium tuberculosis

truncated hemoglobin N from molecular dynamics simulations .......................................... 97

6.1. Résumé .................................................................................................................. 97

6.2. Abstract ................................................................................................................. 98

6.3. Introduction ........................................................................................................... 99

6.4. Methods .............................................................................................................. 101

6.5. Results and discussion ........................................................................................ 104

6.5.1. Active site configurations ........................................................................... 104

6.5.2. MD simulations of oxygenated TrHbN ...................................................... 104

6.5.3. MD simulations of the ferrous unliganded TrHbN ..................................... 105

6.5.4. Gln58(E11) and Phe62(E15) dynamics are linked ..................................... 105

6.5.5. Characterization of cavities and tunnels ..................................................... 107

6.6. Concluding remarks ............................................................................................ 113

6.7. Acknowledgments .............................................................................................. 114

6.8. References ........................................................................................................... 114

Chapitre 7 Theoretical Investigations of Nitric Oxide Channeling in Mycobacterium

tuberculosis Truncated Hemoglobin N ............................................................................... 131

7.1. Résumé ................................................................................................................ 131

xii

7.2. Abstract ............................................................................................................... 132

7.3. Introduction ......................................................................................................... 133

7.4. Methods .............................................................................................................. 135

7.4.1. Analysis ...................................................................................................... 135

7.4.2. Implicit ligand sampling ............................................................................. 136

7.5. Results and discussion ........................................................................................ 137

7.5.1. •NO enters the protein matrix using the ST, LT, and EHT ........................ 138

7.5.2. Diffusion through tunnels ........................................................................... 138

7.5.3. Ligand binding affinities ............................................................................. 143

7.5.4. Comparison of the different paths .............................................................. 144

7.6. Conclusion .......................................................................................................... 146

7.7. Supporting material ............................................................................................. 148

7.8. References ........................................................................................................... 148

Chapitre 8 Structure and Dynamics of Mycobacterium tuberculosis Truncated Hemoglobin

N: Insights from NMR Spectroscopy and Molecular Dynamics Simulations .................... 161

8.1. Résumé ................................................................................................................ 161

8.2. Abstract ............................................................................................................... 162

8.3. Introduction ......................................................................................................... 163

8.4. Material and methods .......................................................................................... 165

8.4.1. NMR ........................................................................................................... 165

8.4.2. Molecular dynamics simulations ................................................................ 166

8.5. Results ................................................................................................................. 168

8.5.1. NMR ........................................................................................................... 168

8.5.2. Molecular dynamics simulation and comparison with NMR results .......... 173

8.6. Discussion ........................................................................................................... 174

8.7. Conclusion .......................................................................................................... 178

8.8. Acknowledgments .............................................................................................. 179

8.9. Supporting information ....................................................................................... 179

8.10. References ........................................................................................................... 180

Chapitre 9 Experimental and Theoretical Investigations Reveal that Mycobacterium

tuberculosis Truncated Hemoglobin N contains Multiple Diffusion Routes to Sustain Rapid

Gaseous Ligand Entry and Exit. ......................................................................................... 193

9.1. Résumé ................................................................................................................ 193

9.2. Abstract ............................................................................................................... 194

9.3. Introduction ......................................................................................................... 195

9.4. Experimental procedures .................................................................................... 197

9.5. Results ................................................................................................................. 200

9.5.1. NOD reaction of TrHbN. ............................................................................ 200

9.5.2. Mutants with obstructed tunnel entrance(s) show ns geminate rebinding of

the •NO…. .................................................................................................................. 202

9.5.3. MD simulations emphasize the importance of side-chain flexibility on ligand

diffusion.. .................................................................................................................... 203

9.6. Conclusions ......................................................................................................... 208

9.7. References ........................................................................................................... 209

9.8. Footnotes ............................................................................................................. 213

Chapitre 10 Discussion ...................................................................................................... 223

10.1. Structure et dynamique de la poche distale de l’hème ........................................ 223

xiii

10.2. Structure et dynamique des tunnels de TrHbN ................................................... 225

10.3. Rôle des tunnels dans la diffusion des substrats entre le solvant et le site actif . 226

10.4. Perspective de recherche sur les relations structure-fonction des tunnels de

TrHbN.. ........................................................................................................................... 229

10.5. Routes de diffusions multiples et pertinence fonctionnelle ................................ 229

10.6. Comparaisons de TrHbN avec d’autres protéines liant des gaz et perspectives . 230

10.7. Localisation de TrHbN au niveau des membranes – nouvelles perpectives de

recherche ......................................................................................................................... 232

10.8. Conclusion .......................................................................................................... 234

Bibliographie ...................................................................................................................... 235

Annexe 1 ............................................................................................................................. 246

Matériel supplémentaire du chapitre 6 ................................................................................ 246

Annexe 2 ............................................................................................................................. 250


Annexe 3 ............................................................................................................................. 257


Annexe 4 ............................................................................................................................. 280


xiv

Liste des tableaux et tables

CHAPITRE 1

Tableau 1-1 Activité NOD chez certaines globines .............................................................. 15

Tableau 1-2 Résidus de la poche distale de certaines globines en lien avec les constantes de

liaison et de dissociation de l’O2. .................................................................................. 18

Tableau 1-3 Structures tridimensionnelles résolues chez les hémoglobines tronquées ........ 21

Tableau 1-4 Volume total interne des cavités de quelques globines .................................... 23

Tableau 1-5 Tunnels observés chez diverses protéines ........................................................ 27

CHAPITRE 5

Table 5.1 Kinetics constants for the reactions of TrHbN and its mutants with O2 and CO. 89

Table 5.2 Average minimum interatomic distances between non-hydrogen atoms and the

heme iron. ..................................................................................................................... 90

Table 5.3 . Cavity formation frequency and volume over the iron atom. ............................. 91

Table 6.1 Interatomic Distances Between Relevant Atoms From Trajectories .................. 119

Table 6.2 Distribution of the Different Rotameric Species Encountered During Simulations

for Q(E11) and F(E15) ................................................................................................ 120

Table 6.3 Tunnels Physical Properties ................................................................................ 121

Table 7.1 Calculated affinities for NO and solvent-excluded volume at tunnel entrances

detected for TrHbN and multiple polar mutant. ......................................................... 152

Table 7.2 Rotamers observed for two residues upon the absence or presence of •NO

molecule in specific cavities. ...................................................................................... 153

Table 8.1 Average R1 and R2 relaxation rates (s-1

) and {1H}-

15N NOEs at 500, 600, and 800

MHz. ........................................................................................................................... 185

Table 9.1 Kinetics constants for the NOD reactions of TrHbN and its tunnel mutants. .... 214

Table 9.2 Number of MD snapshots showing a tunnel open at its entrance from simulations

of TrHbN and the triple mutant. ................................................................................. 215

Table 9.3 Exit and entry events observed in LES simulations of the wild type and triple

mutant. ........................................................................................................................ 216

xv

Liste des figures

Figure 1.1. Structure de l’hème (protoporphyrine IX).. .......................................................... 6

Figure 1.2 Évolution et distribution des globines dans les règnes du vivant.. ........................ 9

Figure 1.3 Cycle de la réaction NOD catalysée par les globines.. ........................................ 12

Figure 1.4 Mécanisme réactionnel de la réaction NOD impliquant la rupture homolytique

du lien O-O.. ................................................................................................................. 13

Figure 1.5 Mécanisme concerté pour la réaction NOD… .................................................... 13

Figure 1.6 Comparaison des repliements globines 2-sur-2 et 3-sur-3. ................................. 20

Figure 1.7 Cavités observées dans la structure de quelques globines. ................................. 25

Figure 1.8 Structure tertiaire de TrHbN.. .............................................................................. 26

Figure 1.9 Site actif de TrHbN.. ........................................................................................... 29

Figure 1.10 Cavité distale de TrHbN sous sa forme oxygénée. ........................................... 31

Figure 1.11 Conformations alternatives de la Phe62(E15).. ................................................. 33

Figure 2.1 Modèles de diffusion des substrats chez la Mb.. ................................................. 36

Figure 2.2 Cinétiques de recombinaison hypothétiques pour une globine donnée sous sa

forme sauvage et pour un mutant. ................................................................................. 39

Figure 2.3 Échelles de temps des différents mouvements se produisant dans les protéines..

...................................................................................................................................... 44

Figure 3.1 Mouvements internes dans les molécules. .......................................................... 52

Figure 3.2. Conditions périodiques aux frontières d’un système en 2D.. ............................. 55

Figure 3.3 Algorithme de Verlet pour l’intégration de l’équation de mouvement. .............. 57

Figure 3.4 Schéma des grandes étapes de la dynamique moléculaire .................................. 58

Figure 5.1 View of the distal heme pocket and the tunnels of cyanomet-TrHbN chain B

under xenon pressure (PDB 1S56). ............................................................................... 92

Figure 5.2 Equilibrium absorption spectra of TrHbN(Fe3+

-H2O), TrHbN(Fe3+

-NO) and

TrHbN Tyr(B10)Leu/Gln(E11)Val(Fe3+

) mutant at pH 7.5. ....................................... 93

Figure 5.3 The time courses of O2 and CO recombination to TrHbN. ................................ 94

Figure 5.4 Kinetic traces showing the recombination of CO subsequent to nanosecond

photodissociation of the CO saturated derivatives of wild-type TrHbN and its distal

mutants.. ........................................................................................................................ 95

Figure 5.5 Kinetic traces illustrating the absorbance changes following photodissociation of

TrHbN(Fe3+

-NO) and Mb(Fe3+

-NO) at 23 °C.. ............................................................ 96

Figure 6.1 TrHbN structure (PDB entry 1IDR, subunit A). ............................................... 122

Figure 6.2 Active site configurations from typical MD frames for oxy-TrHbN and deoxy-

TrHbN. ........................................................................................................................ 123

Figure 6.3 Phe62(E15) χ1 as function of χ2 for oxy-TrHbN and deoxy-TrHbN…. .......... 124

Figure 6.4 Gln58(E11) and Phe62(E15) χ1 dihedral angle as function of simulation time for

oxy-TrHbN and deoxy-TrHbN ................................................................................... 125

Figure 6.5 Different snapshots of cavities in TrHbN. ......................................................... 126

Figure 6.6 Backbone 1H-

15N order parameters as function of residue sequence number

calculated from trajectories of A-TrHbN and B-TrHbN in oxy-TrHbN and deoxy-

TrHbN.. ....................................................................................................................... 127

Figure 6.7 Representation of the long , short, EH, LEH and BE tunnels. .......................... 128

Figure 6.8 Profiles generated for each tunnel leading from distal heme pocket to the bulk

solvent. ........................................................................................................................ 129

xvi

Figure 6.9 Averaged minimum radius (bottleneck) of the long and EH tunnels according to

the Phe62(E15) χ2 dihedral. ........................................................................................ 130

Figure 7.1 TrHbN structure (PDB entry 1IDR, subunit A). ............................................... 154

Figure 7.2 Time of contact between •NO molecules and TrHbN relative to the MD

simulation time.. ......................................................................................................... 156

Figure 7.3 Representative solvent-excluded volume formed over the ST entrance. ......... 157

Figure 7.4 Density probability of •NO derived from explicit MD simulations and implicit

ligand PMF for •NO inside TrHbN calculated from MD frames having the Phe62(E15)

in the M state and T state.. .......................................................................................... 158

Figure 7.5 PMF profiles for •NO diffusion in ST regardless and as function of different

Ile119(H11) rotamers.. ................................................................................................ 159

Figure 7.6 Phe62(E15) χ1 and χ2 dihedral angles as function of the simulation time. ...... 160

Figure 8.1 Structure of TrHbN displaying the four tunnels: Long tunnel (LT), Short tunnel

(ST), EH tunnel, and BE tunnel. ................................................................................. 186

Figure 8.2 Assigned 1H-

15N HSQC spectrum of TrHbN cyanomet ................................... 187

Figure 8.3 NMR raw relaxation data (R1, R2, R2/R1, NOE) at 500, 600, and 800 MHz.;

Model-free parameters (S2, Rex, and τe) for TrHbN cyanomet; Comparision of S

2

parameters obtained either from NMR or MD simulations; NMR amide exchange

data: amide exchanges rates (kex) at pH 7.5and 8.5, and free energy for the opening of

the protecting structure (ΔGHX) at pH 7.5. Molecular Dynamics data: Average

backbone ASA and Backbone amide hydrogen bond occupancy; Secondary structure

of TrHbN calculated by NMR, MD, or taken from the X-ray structure PDB 1S61B.

.................................................................................................................................... 189

Figure 9.1 TrHbN tunnel system. ....................................................................................... 217

Figure 9.2 Reaction of TrHbNFeII(O2) with one equivalent of •NO at 5 ºC, pH 9.5. ........ 218

Figure 9.3 Reaction of the FeII(O2) form of TrHbN and tunnel mutants LT/ST and

LT/ST/EHT with one equivalent of •NO at 5 ºC, pH 9.5. .......................................... 219

Figure 9.4 Kinetic traces illustrating the absorbance changes after photo-dissociation of

TrHbNFeIII

(•NO), ST-FeIII

(•NO), LT/ST-FeIII

(•NO), and LT/ST/EHT-FeIII

(•NO) at

23 °C. .......................................................................................................................... 220

Figure 9.5 PMF profiles for •NO diffusion in the different tunnels for the TrHbN and

mutant. ........................................................................................................................ 221

Figure 9.6 Typical closed and open tunnels at the surface observed for the triple mutant

protein.. ....................................................................................................................... 222

xvii

Liste des abréviations

A Alanine

Ala Alanine

Arg Arginine

Asp Aspartate

BCG Bacille de Calmette-Guerin

D Aspartate

DHP Distal heme pocket

DM Dynamique moléculaire

E Glutamate

ÉIL Échantillonnage implicite de ligand

F Phénylalanine

Gln Glutamine

Glu Glutamate

K Kelvin

H Histidine

Hb Hémoglobine

His Histidine

I Isoleucine

Ile Isoleucine

L Leucine

Leu Leucine

OMS Organisation Mondiale de la Santé

Phe Phénylalanine

P Proline

ms Milliseconde

Mtb Mycobacterium tuberculosis

ns nanoseconde

PFM Potentiel de force moyen

ps picoseconde

xviii

Q Glutamine

R Arginine

RMN Résonance magnétique nucléaire

S Sérine

Ser Sérine

T Thréonine

Thr Thréonine

TrHb Hémoglobine tronquée

TrHbN Hémoglobine tronquée N

Trp Tryptophane

Tyr Tyrosine

µs microseconde

V Valine

Val Valine

VIH Virus de l’immunodéficience humaine

W Tryptophane

Y Tyrosine

1

1.

Chapitre 1

Introduction

1.1. Contexte biologique de TrHbN

Selon le plus récent rapport de l’Organisation mondiale de la santé (OMS), près d’une

personne sur trois sur la planète est porteuse du germe responsable de la tuberculose, c’est-

à-dire la bactérie Mycobacterium tuberculosis (Mtb) [1]. Chez la plupart des personnes

infectées, Mtb est contraint à un stade de dormance par le système immunitaire. Environ

10 % des personnes infectées développeront une pathologie au cours de leur vie et environ

2 millions de personnes en meurent chaque année [1]. Le nombre de nouveaux cas de

tuberculose déclarés chaque année est stable, soit 8.9 millions en 2004 [2] et 9.4 millions en

2008 [1] suivant le rythme d’accroissement de la population mondiale [1]. Parmi les

personnes décédant de la tuberculose, près du tiers sont également porteuses du virus de

l’immunodéficience humaine (VIH). La co-infection Mtb-VIH augmente de 20 à 40 fois le

risque de développement de la tuberculose [1]. Enfin, environ 5% des nouveaux cas de

tuberculose sont causés par une souche Mtb multirésistante aux traitements de première

ligne (antibiotiques rifampicine et isoniazide) ce qui amène une nouvelle inquiétude [3, 4].

1.1.1. Mécanismes d’infection de Mycobacterium tuberculosis

La tuberculose est une maladie complexe et les relations hôte-pathogène sont encore

aujourd’hui partiellement comprises. L’infection se produit généralement chez les

personnes exposées à répétitions à des gens infectés par Mtb et générant des expectorations

via leur toux [5]. Les bacilles sont ainsi transmis par l’inhalation d`aérosols causés par la

toux de personnes malades [5]. Chez les personnes en santé, les bacilles inhalés sont

2

rapidement pris en charge par le système immunitaire de l’hôte [5]. Cette réponse

immunitaire débute par le recrutement de macrophages alvéolaires qui phagocytent les

bacilles. Un premier mécanisme de résistance de Mtb survient alors par l’arrêt de la

maturation du phagosome et par l’évitement de la fusion avec la membrane lysosomale [6-

8]. À ce stade, Mtb prolifère à l’intérieur du macrophage [5]. En réponse, les macrophages

infectés sécrètent des cytokines recrutant d’autres cellules de l’immunité [9-11]. Ce second

mécanisme de défense permet d’empêcher la dissémination de Mtb à tout le poumon et aux

autres tissus. Ces régions isolées portent le nom de granulomes. Il est proposé que les

conditions régnant à l’intérieur des granulomes soient hostiles à Mtb. L’hypoxie, la

présence de molécules réactives dérivées de l’oxygène (ion superoxyde) et de l’azote

(oxyde nitrique, peroxynitrite) et la présence d’acides gras libres seraient les conditions

auxquelles Mtb aurait à faire face [9, 12-15]. À l’intérieur de cet environnement hostile,

Mtb survit, mais est latent. Ce stade de l’infection peut perdurer pendant de nombreuses

années et est habituellement asymptomatique [5]. Durant cette longue période, une lutte

entre le système immunitaire et Mtb a lieu au cœur des granulomes [16]. Avec le temps, les

granulomes peuvent se calcifier emprisonnant les bacilles de manière durable. Dans

d’autres cas, les granulomes se liquéfient, phénomène également appelé la caséation, ce qui

permet à nouveau la prolifération bactérienne dans le poumon [5]. L’infection peut

également s’étendre à d’autres tissus et on parle alors de tuberculose disséminée. Ces

derniers cas se produisent généralement chez les personnes faibles, âgées, stressées, mal

nourries, immunodéprimées ou infectées par le VIH [5].

1.1.2. Mécanismes de résistance de Mtb

Parmi les stratégies d’actions contre la bactérie Mtb émise par l’OMS, l’étude des

mécanismes de résistance et de latence de Mtb a été ciblée [17]. La résistance de Mtb face

aux espèces réactives de l’oxygène et de l’azote est particulièrement visée [9, 12-15]. La

production d’oxyde nitrique (•NO) et la faible concentration d’oxygène moléculaire

disponible empêchent la prolifération de Mtb et favorisent l’entrée en phase

3

stationnaire [15]. Le •NO compromet la respiration aérobie en inhibant des enzymes clefs

du cycle de Krebs telles que l’aconitase et les cytochromes oxydases respiratoires

terminales [18, 19]. Le •NO peut également réagir avec l’ion superoxyde (O2-) et former

d’autres espèces réactives et toxiques, dont l’oxyde nitreux (•NO2), le trioxyde d’azote

(N2O3) et l’ion peroxynitrite (OONO-) [20, 21]. La capacité de Mtb à survivre durant

plusieurs années sous ces conditions ne peut se faire sans la présence d’un mécanisme

endogène de résistance contre le •NO.

L’hémoglobine tronquée TrHbN exprimée par Mtb a été ciblée comme un acteur

potentiellement important dans cette défense contre le •NO. Il a été démontré que TrHbN

est capable de protéger la respiration aérobie de Mycobacterium tuberculosis BCG contre

l’effet inhibiteur du •NO. De plus, TrHbN catalyse très efficacement la conversion du •NO

en nitrate avec une constante bimoléculaire de 745 µM-1

s-1

à 23 °C (réaction

•NO-dioxygénase ou NOD) [22]. À elles seules, ces propriétés biochimiques justifient

pleinement l’intérêt scientifique porté vers TrHbN

1.2. TrHbN et la superfamille des globines

1.2.1. Définition du terme globine

Par définition, les globines constituent une superfamille de protéines dont la structure

tridimensionnelle, plutôt globulaire, renferme un cofacteur appelé l’hème. Certaines d’entre

elles sont des modèles largement étudiés en biochimie telle que la myoglobine (Mb) et

l’hémoglobine tétramérique mammalienne (Hb) dont les fonctions principales sont le

stockage et le transport de l’O2 respectivement. La structure tridimensionnelle des globines

adopte un arrangement structural typique communément appelé « repliement globine ».

Celui-ci est caractérisé par une série d’hélices α organisées dans l’espace afin de contenir

l’hème et le placer dans un environnement adéquat pour assurer la fonction de la globine.

Cet environnement permet notamment la liaison réversible de l’oxygène moléculaire avec

le fer sous sa forme ferreuse et prévient l’oxydation du fer [23].

4

1.2.2. L’hème

L’hème est une molécule chimique composée d’un atome de fer prenant place au centre

d’un large anneau organique hétérocyclique que l’on nomme porphyrine. Des groupements

chimiques (méthyle, vinyle, propionate) bordent la porphyrine. Ces groupements chimiques

peuvent varier donnant lieu à différents types d’hèmes. Le type d’hème retrouvé chez les

globines est l’hème b, également connu sous le nom protoporphyrine IX (Figure 1.1). Ce

type d’hème contient trois groupements méthyles, deux groupements vinyles et deux

propionates.

Le fer de l’hème se retrouve principalement sous deux états d’oxydation, soit ferreux (Fe2+

)

et ferrique (Fe3+

). Lors de certaines réactions d’oxydoréduction, dont la réaction NOD, le

fer peut se retrouver transitoirement sous une forme ferryle (Fe4+

) [24]. En fonction de

l’état d’oxydation, l’hème peut se lier à diverses molécules gazeuses diatomiques et

catalyser des réactions d’oxydoréduction. Sous sa forme Fe2+

, l’hème peut se lier à l’O2, au

CO et au •NO. L’hème sous sa forme Fe3+

peut se lier avec l’eau, le •NO, l’imidazole, l’ion

cyanure (CN-), l’ion hydroxyde (OH

-) et d’autres anions (COO

-, NO2

-, N3

-, Cl

-, F

-). Le fer

peut aussi se présenter sous différents types de coordination. Lorsque le fer n’est chélaté

que par les quatre azotes de l’anneau porphyrique, le fer est dit tétracoordonné (4C).

Lorsqu’en plus l’histidine proximale (His(F8)) est liée au fer, l’hème est alors dit

pentacoordonné (5C). Sous sa forme 5C, une globine est communément appelée « deoxy »

ou « met » selon si le fer est Fe2+

ou Fe3+

, respectivement. Finalement, lorsqu’un ligand est

lié au fer sur la face distale, l’hème est hexacoordonné (6C).

L’état d’oxydation du fer ferrique est un facteur majeur influençant la liaison des ligands. À

un pH près de la neutralité, l’hème ferrique a une charge nette de +1 (+3 pour le fer et -2

pour quatre pyroles) alors que le fer ferreux est neutre (+2 pour le fer, et -2 pour les azotes

des pyroles). L’hème ferrique tend naturellement à réagir plus fortement avec les ligands

anioniques étant donné l’attraction électrostatique [25]. L’hème ferreux quant-à-lui interagit

préférentiellement avec des ligands gazeux neutres (O2, CO, •NO) et ceci est imputable à

5

l’augmentation du nombre d’interactions dans les orbitales d : le fer ferreux (d6) a un

électron additionnel dans une orbitale d comparativement à l’hème ferrique (d5). Cet

électron additionnel permet de doubler l’interaction entre l’orbitale dxx/dxy du fer et

l’orbitale anti-liante p du carbone (CO) ou de l’azote (•NO) [25]. Ceci jouerait un rôle

important dans l’interaction entre le fer et les ligands gazeux. La force du ligand avec le fer

ferreux varie (•NO > CO > O2) et ceci s’explique par certaines différences au niveau des

interactions dans les orbitales. Il est proposé que pour l’O2, l’hybridation soit de type sp2

avec l’hème alors que le CO présenterait une hybridation de type sp [25]. Pour l’O2, la

présence de deux électrons dans des orbitales anti-liantes diminue la force du lien avec le

fer. Par conséquent, le lien Fe-C est plus court que le lien Fe-O. Le cas du •NO (Fe-N-O)

est particulier. L’azote peut se lier selon une hydribation sp ou sp2. Le •NO présente un

nombre impair d’électrons conférant une polarité à la molécule lui permettant de se lier à

un hème ferrique. Le •NO peut partager 2 ou 3 électrons faisant de lui un ligand plus fort

que le CO. Dans l’espace, l’arrangement structural varie également. Fe-C-N est linéaire

alors que l’arrangement Fe-O-O est penchée (angle ~120º) [25]. L’arrangement structural

Fe-N-O peut se présenter selon les deux types configurations.

Il est important de noter que l’O2 liée au fer hémique présente typiquement un caractère

superoxyde (Fe3+

O2-) [24]. Ce caractère serait en particulier important dans la catalyse de la

réaction de dioxygénation du •NO (réaction NOD) tel que proposé initialement par Doyle et

Hoekstra [26].

6

Figure 1.1. Structure de l’hème (protoporphyrine IX). Le fer, les azotes, les oxygènes les

carbones et les hydrogènes sont respectivement colorés en rouge brique, bleu, rouge, gris et

blanc. Les coordonnées sont celles de l’hème contenu dans l’hémoglobine tronquée N de

Mycobacterium tuberculosis (Accession PDB 1IDR, chaîne A). La figure a été produite à

l’aide de PyMOL [27].

7

1.2.3. Phylogénie des globines

Les globines se retrouvent dans tous les règnes du vivant. Le modèle phylogénétique de

l’évolution des globines est présenté à la Figure 1.2 [28]. L’origine des globines est très

ancienne. Les premières seraient apparues il y a environ 3.5 milliards d’années. Il est

proposé que les globines aient évolué à partir d’un gène de globine ancestral [28]. Cette

globine ancestrale serait apparue peu après l’apparition de la vie à une époque où la

concentration atmosphérique en O2 était très faible (< 0.8x10-3

atm), soit avant l’avènement

de la photosynthèse. À cette époque, sa fonction aurait été la détection, la séquestration et la

détoxification de l’O2 [28]. L’évolution de cet ancêtre aurait donné lieu à deux types de

repliements globines, les globines 2-sur-2 et les globines 3-sur-3. Les caractéristiques de

ces repliements sont traitées à la section 1.2.7.

La distribution actuelle des globines place les hémoglobines 2-sur-2 chez les

archaebactéries, les bactéries, les plantes et les eucaryotes unicellulaires. Les

flavohémoglobines se retrouvent chez les archaebactéries et les bactéries. Enfin, les

hémoglobines 3-sur-3 (à simple domaine) se retrouvent chez les bactéries et les

eucaryotes [28]. Lors de la découverte des premières globines 2-sur-2, ces globines étaient

appelées « hémoglobines tronquées » (TrHbs) puisqu’elles présentaient d’importantes

délétions par rapport aux globines dites classiques dites « 3-sur-3 » [28]. La terminologie

« 2-sur-2 » est apparue ensuite et celle-ci s’appuie sur l’arrangement structural en

3-dimensions. La terminologie « hémoglobine tronquée » tend à disparaître au profit de

l’autre appellation. Néanmoins, cette thèse utilisera les deux appellations sans préférence.

Les TrHbs sont subdivisées selon trois groupes distincts : le groupe I (ou N), le groupe II

(ou O) et le groupe III (ou P) [28-30]. De manière étonnante, il existe moins de 20 %

d’identité entre ces trois groupes. L’identité est par contre plus élevée à l’intérieur d’un

même groupe et peut atteindre près de 80 %. Les déterminants structuraux sont orthologues

à l’intérieur du même groupe et sont paralogues d’un groupe à l’autre [31]. Le groupe III

présente le plus haut niveau de conservation. Le groupe II serait la forme ancestrale alors

que les groupes I et III seraient le résultat d’évènements de réplications et de transferts

8

génétiques [31]. La présence des TrHbs des trois groupes chez certains organismes appuie

cette hypothèse. Les TrHbs des différents groupes auraient donc évolué en donnant lieu à

différentes fonctions. Parmi les fonctions suggérées des TrHbs, il y a la détoxification du

•NO et de l’O2, le stockage de substrat, source d’O2 en réponse à l’hypoxie et la détection

de ligand (O2/•NO) [28, 32].

9

Figure 1.2 Évolution et distribution des globines dans les règnes du vivant. Figure adaptée

de Vinogradov et al. [28] avec l’accord du journal.

10

1.2.4. Fonctions des globines

D’une globine à une autre, les fonctions peuvent diverger [32]. Parmi les fonctions

possibles, notons le stockage, le transport et la détection et la détoxification de l’oxygène

moléculaire (O2). À ceci s’ajoute la catalyse de réactions d’oxydo-reduction telles que

l’oxydation de l’oxyde nitrique en nitrate (réaction NOD), la réduction du nitrite, la

nitrosylation de l’O2 et la réduction du •NO [32]. Une même globine peut assurer plusieurs

fonctions [32]. Par exemple, en plus de son rôle de transport de l’O2, l’Hb catalyse la

réaction NOD (section 1.2.5) ce qui permet de réguler la pression artérielle [33]. Quant à la

Mb, en plus de sa fonction de stockage de l’O2, elle jouerait un rôle important dans le

contrôle du niveau intracellulaire en nitrite (NO2-) et •NO. Pour ce faire, en présence d’O2

et de •NO, le Mb abaisserait la concentration en •NO en catalysant la réaction NOD. À

l’inverse, en condition pauvre en oxygène, la Mb sous sa forme deoxy réagit avec le NO2-

pour produire du •NO [34, 35]. En causant la vasodilatation, cette dernière réaction

préviendrait les dommages aux tissus causés lors d’une période d’ischémie prolongée.

La fonction d’une globine donnée est grandement dictée par la réactivité de son hème.

Cette réactivité varie d’une globine à une autre. La nature et le positionnement des acides

aminés bordant l’hème influencent cette réactivité. La nature des acides aminées

permettront ou non certaines interactions directes avec le ligand tel que des ponts

hydrogènes. Ces interactions sont susceptibles de stabiliser ou non le ligand. De plus, la

taille de ces résidus causera plus ou moins d’encombrement stérique modulant l’accès au

site actif. D’autres facteurs peuvent également l’influencer dont le repliement

tridimensionnel et l’allostérie. La présence de tunnel(s) dans la matrice de certaines

globines [36-41] et le phénomène de coopérativité observée pour l’hémoglobine

tétramérique de mammifère [42] sont des bons exemples. Certaines globines sont formées

d’un seul domaine globine alors que d’autres sont chimériques. Ces dernières contiennent

un domaine globine en N-terminal alors que le domaine en C-terminal peut être soit une

réductase (FAD) (les flavohémoglobines) ou encore un domaine de régulation de génique

(globines couplées à un senseur) [32].

11

1.2.5. La réaction •NO-dioxygénase

La réaction •NO-dioxygénase, également appelée « NOD », est une réaction

d’oxydoréduction dont l’équation chimique globale est donnée par :

• →

(Équation 1.1)

En solution, la réaction non enzymatique entre le •NO et l’anion superoxyde (•O2-) est très

rapide, n’étant limitée que par la vitesse de diffusion des molécules (constante

bimoléculaire ~6700 µM-1

s-1

) [43, 44]. Le produit de cette réaction est l’anion peroxynitrite

ONOO-, un oxydant puissant également toxique. La forme déprotonée ONOO

- est stable,

perdurant des jours [45], mais la forme acide ONOOH (pKa=6.8) s’isomérise spontanément

en nitrate à un taux de 1.3 s-1

[44]. À l’intérieur des cellules, ONOO-

peut oxyder les

groupements thiols, hydroxyler les phénylalanines (ONOOH ↔ O=N-O…OH → OH• +

•NO2-, le radical hydroxyle réagissant avec la phénylalanine pour former la m-, o- ou p-

tyrosine), nitrater les tyrosines via le radical •NO2-, cliver l’ADN et oxyder ou nitrater les

guanosines [20, 21, 45, 46].

Étant donné la toxicité du •NO, les niveaux intracellulaires doivent être modulés. La

réaction entre le •NO et l’anion •O2- est très rapide, cependant la concentration

intracellulaire de •O2- est très faible et ne constitue donc pas une voie de détoxification

efficace (~10-10

M) [24]. Par contre, la concentration intracellulaire plus élevée des globines

(flavohémoglobines, hémoglobines et myoglobines) (≥ 10-7

M) permet une régulation plus

efficace des niveaux de •NO [24].

Le cycle de transformation du •NO en nitrate catalysé par les globines est représenté à la

Figure 1.3. Il n’y a toujours pas de consensus quant au mécanisme réactionnel précis

prenant place à l’hème. Deux mécanismes sont proposés. Dans un premier, décrit à la

Figure 1.4, la molécule de •NO se rend d’abord au site actif de la globine sous sa forme

oxygénée. L’O2 lié réagit alors avec la molécule de •NO pour former l’ion ONOO-. Par la

suite, un clivage homolytique survient entre les deux atomes de l’oxygène lié formant un

12

radical dioxyde d’azote (•NO2) et l’hème adopte la forme oxo-ferryl (FeIV

-O). Ces

intermédiaires se réorganisent ensuite pour former l’ion nitrate lié au fer. Une fois formé,

l’ion nitrate, qui a une faible affinité pour le fer, est relâché dans le cytoplasme [47].

Figure 1.3 Cycle de la réaction NOD catalysée par les globines. La forme oxygénée

(Fe2+

O2) réagit avec le •NO pour former une molécule de nitrate. Le nitrate se déplace

ensuite vers le solvant. Une molécule d’eau du solvant peut alors se fixer au Fer3+

. La

forme Fe2+

est régénérée suivant une réduction (réductase, molécule réductrice

cytoplasmique). L’oxygène moléculaire peut se fixer à nouveau au fer et la globine est

prête pour un second tour réactionnel.

Fe2+

O2

Fe3+

NO3- Fe

3+H2O

•N

O

H2O NO3-

Fe2+

O2

H2

O

é

13

Figure 1.4 Mécanisme réactionnel de la réaction NOD impliquant la rupture homolytique

du lien O-O. Les crochets dénotent des intermédiaires réactionnels.

Puisque l’intermédiaire peroxynitrite n’a jamais été observé expérimentalement, un autre

mécanisme a été proposé, soit le mécanisme concerté (Figure 1.5) [24]. Dans ce

mécanisme, la vibration du lien peroxo (O-O) et la contraction du lien Fe-OO-NO

favorisent le déplacement de la paire d’électrons libres sur l’azote vers l’oxygène lié au fer.

Ce déplacement causerait la rupture du lien O-O de concert avec la formation simultanée du

troisième lien O-N. Ainsi, la réaction suivrait le même mécanisme d’isomérisation que

celui de l’acide peroxynitrique (ONOOH) en nitrate en solution [24, 45, 48].

Figure 1.5 Mécanisme concerté pour la réaction NOD. Les crochets dénotent des

intermédiaires réactionnels.

14

De récents calculs de QM/MM ont permis d’étudier ces deux mécanismes réactionnels chez

TrHbN [49]. Ces travaux ont conclu que le mécanisme concerté serait le plus favorable

chez TrHbN. L’autre mécanisme serait également possible mais serait beaucoup plus lent

car limité par la rupture du lien O-O. Dans ce mécanisme, la configuration du site actif de

TrHbN serait cruciale pour le maintien des intermédiaires réactionnels près du fer et éviter

que ceux-ci s’échappent du site actif et diffusent dans la matrice protéique pour réagir avec

la protéine, en particulier les tyrosines souvent retrouvées dans la poche distale. De plus,

cette réaction se ferait très rapidement, soit dans l’ordre de quelques dizaines picosecondes,

faisant vraisemblablement de la diffusion des •NO du solvant vers le site actif comme

l’étape limitante pour la réaction NOD [49].

Comme le montre le Tableau 1-1, la vitesse de la réaction NOD varie beaucoup selon les

globines. Les enzymes catalysant le plus efficacement la réaction NOD se retrouvent chez

les bactéries. En particulier, les flavohémoglobines (FHbs) et l’hémoglobine tronquée

TrHbN présentent des niveaux d’activité très élevés et où les constantes bimoléculaires

atteignent de 745 à 2900 µM-1

s-1

[22, 50, 51]. Ces taux de catalyse s’approchent de la

réaction en solution entre l’ion O2- et •NO, soit 6700 µM

-1s

-1, considérée comme étant

limitée que par la vitesse de diffusion des substrats.

De telles vitesses sont inusitées puisque l’hème des globines se trouve enfoui à l’intérieur

de la matrice protéique. En effet, de telles vitesses suggèrent que la diffusion du •NO entre

le solvant et le site actif se fait très rapidement. Ce phénomène chez TrHbN fait d’ailleurs

l’objet d’une attention particulière dans la présente thèse.

15

Tableau 1-1 Activité NOD chez certaines globines

Organisme Type de globine k’NOD

µM-1

s-1

Références

Escherichia coli FHb 2400 a [50]

Alcaligenes eeutrophus FHb 2900 a [50]

Bacillus subtilis FHb 860 b [51]

Deinococcus radiodurans FHb 3 b [51]

Mtb TrHbN 745 c [22]

Mtb TrHbO 0.6 c [52]

Ascaris suum Hb 0.07 d [53]

Homme Hb 50 e [54]

Cachalot Mb 34 e [54]

•NO + O2- (en solution) - 6700 [24]

a. 37 °C, pH 7.8, mesure par électrode de la consommation du •NO.

b. Température de la pièce, pH=7, Consommation du •NO suivie par un analyseur de •NO

couplée à un détecteur de chimiluminescence.

c. 23 °C, pH 7.5, spectrométrie en flux arrêté.

d. Température non spécifiée, pH 6.0, Spectrométrie en flux arrêté

e. 20 °C - pH 7.0 (Mb), 7.4 (Hb), Spectrométrie en flux arrêté

16

1.2.6. La cavité distale

Près du fer, on retrouve une cavité définie par les chaînes latérales de résidus occupant des

positions bien précises. C’est dans cet espace disponible que viennent se lier les ligands et

où les réactions redox sont catalysées. Étant donné que les ligands se fixent sur la face

distale de l’hème, cette cavité est appelée la cavité distale de l’hème. La nomenclature

utilisée pour les acides aminés délimitant cette cavité suit celle utilisée chez la Mb de

cachalot [23, 55]. Ces positions sont : B10, CD1, E7, E11, G8. Par exemple, chez la Mb, la

position 7 de l’hélice E est occupée par une histidine et celle-ci est nommée His(E7).

Comme le montre le Tableau 1-2, la nature des acides aminés à ces positions varie d’une

globine à une autre [31].

Chez les TrHbs du groupe I, les positions E7, E11 et B10 sont occupées principalement par

des acides aminés polaires. En position B10 on retrouve avant tout une tyrosine avec

toutefois certaines exceptions comme chez Nostoc commune, Nostoc punctiform,

Synchosystis sp. et Gemmata obscuriglobus où on retrouve un acide aminé apolaire [30].

Aux positions E7 et E11, on observe au moins une glutamine à l’une ou l’autre des

positions. Parfois, une thréonine est observée en position E7. Enfin, le résidu G8 chez le

groupe I est apolaire et de petite taille comparativement au tryptophane observé chez les

groupes II et III.

Chez les TrHbs du groupe II, on observe une corrélation entre la nature du résidu CD1 et

celle du résidu E11 [29]. Alors que toutes les TrHbs des groupes I et III présentent une

phénylalanine en position CD1, un polymorphisme est présent chez le groupe II

(phénylalanine, tyrosine ou histidine). Lorsqu’un acide aminé donneur de pont hydrogène

est présent en position CD1 (Tyr ou His), le résidu en position E11 est toujours apolaire

(leucine ou phénylalanine). Dans le cas inverse Phe(CD1), un acide aminé polaire (sérine

ou glutamine) est observé en position E11. Des analyses Raman ont permis de conclure que

la Tyr(CD1), lorsque présente, est impliquée dans la stabilisation du substrat lié au lieu de

la Tyr(B10) [56]. Toutes les TrHbs du groupe II connues présentent une Tyr(B10) et une

Trp(G8) [30]. L’atome NE1 de l’anneau indole du Trp(G8) est à une distance permettant

17

des interactions avec le substrat lié. Ce résidu est donc susceptible d’affecter la vitesse de

liaison et de dissociation du ligand. Le groupe III est le groupe le plus homogène. Il est

caractérisé par la présence d’une histidine en position E7 complètement conservée à

l’intérieur de ce groupe [30].

Les résidus polaires de la cavité distale peuvent interagir avec le substrat lié et influencer

l’affinité de la globine pour les ligands. Chez TrHbN, la haute affinité pour l’O2 est

attribuable à l’effet stabilisateur de la Tyr33(B10) [22]. En plus de la polarité, la

configuration spatiale des résidus de la poche distale est importante. Par exemple, les

nématodes Cerebratulus lacteus et Ascaris suum présentent des cavités distales où seul le

résidu en position E11 diffère (Tableau 1-2). Chez C. lacteus [57], la constante

d’association est près de 200 fois plus élevée que celle de A. suum [58]. Par contre, le taux

de dissociation de l’O2 est près de 45 000 fois supérieur chez C. lacteus. Ce taux de

dissociation élevée s’explique par la présence de la thréonine en position E11 qui forme un

pont hydrogène avec de la Tyr(B10). Cette interaction empêche la Tyr(B10) de stabiliser

l’O2 lié [57]. Le changement de la Thr(E11) pour une valine permet de réduire près de 1000

fois la vitesse de dissociation de l’O2 [57]. La vitesse d’association plus élevée chez C.

lacteus, s’explique par une autre caractéristique de sa structure tridimensionnelle. En effet,

la structure de C. lacteus renferme un tunnel permettant la diffusion rapide des ligands [57,

59].

18

Tableau 1-2 Résidus de la poche distale de certaines globines en lien avec les constantes de liaison et de dissociation de l’O2.

Protéines Organismes Positions Constantes cinétiques Références

B10 CD1 E7 E11 G8 KO2 (µM-1

s-1

) KO2 (s-1

)

TrHbN Mycobacterium tuberculosis Y F L Q V 25 0.2 [38]

Paramecium caudatum Y F Q T V 30 25 [60]

Chlamydomonas eugametos Y F Q Q V > 10 1.4×10-2

[60]

TrHbO Mycobacterium tuberculosis Y Y A L W 0.1

1.4×10-3

[52]

TrHbP Campylobacter jejuni Y F H I W 0.9 4.1×10-3

[61]

FHb Escherichia coli Y F L I V 38 0.4 [50]

Hb Ascaris suum Y F Q I L 1.5 4.0×10-3

[58]

MiniHb

Cerebratulus lacteus Y F Q T L 240 180 [57]

Cytoglobine Homme F F H V L 300 0.4 [62]

Neuroglobine Homme F F H V V 140 0.8 [63]

Myoglobine Cachalot L F H V I 14 12 [64, 65]

19

1.2.7. Le repliement globine

Les premières structures de globines résolues ont été celles de la Mb de cachalot et de l’Hb

tétramérique humaine. Ces structures ont été déterminées par cristallographie et diffraction

des rayons X par Kendrew [66] et Perutz [67] respectivement. Ces derniers ont d’ailleurs

obtenu le prix Nobel de chimie en 1962 pour leurs travaux. Toutes les globines partagent

dans leur structure une caractéristique commune, soit la présence d’un domaine nommé le

« domaine globine ». Le domaine globine est formé d’une seule chaîne d’acides aminés

comptant généralement de 140 à 160 acides aminés [28]. Il renferme une série d’hélices α

définissant un cœur hydrophobe dans lequel prend place l’hème. Il existe deux types de

repliement globine, le repliement classique dit « 3-sur-3 » et le repliement « 2-sur-2 ».

Même lorsque deux globines présentent une très faible homologie de séquence, il est

possible de superposer leur structure de manière remarquable (Figure 1.6). Les différences

entre les repliements 3-sur-3 et 2-sur-2 sont illustrées à la Figure 1.6.

Les globines 3-sur-3 renferment habituellement sept ou huit hélices α (hélices A à H).

Parmi ces hélices, six composent le repliement 3-sur-3 (hélices A, B, E, F, G, et H). Le

repliement 2-sur-2 est caractéristique des hémoglobines tronquées. Comme mentionné

précédemment, l’appellation « hémoglobine tronquée » tend à disparaître pour faire place à

« hémoglobine 2-sur-2 ». Ce repliement repose sur quatre hélices α (hélices B, E, G et H).

L’hélice F, renfermant l’histidine proximale, est grandement raccourcie (un seul tour pour

les groupes I et III). Quant à l’hélice A, elle est très courte (groupes I et II) voire même

absente (groupe III). L’hélice D est toujours absente. En plus des changements qui

concernent les hélices α, on note la présence d’une longue boucle séparant les hélices E et

F. Chez TrHbN, on remarque la présence d’une hélice supplémentaire caractéristique du

côté N terminal; l’hélice pré-A. Les hémoglobines tronquées portent leur nom du fait que

leur structure présente d’importantes simplifications (raccourcissements et délétions

d’hélices α) par rapport aux globines classiques. Le nombre de globines 2-sur-2 dont la

structure a été résolue est encore aujourd’hui limité. Le Tableau 1-3 résume les différentes

structures de TrHbs disponibles et les différentes méthodes employées. Jusqu’à maintenant,

on note quatre structures du groupe I, cinq pour le groupe II et une pour le groupe III.

20

Figure 1.6 Comparaison des repliements globines 2-sur-2 et 3-sur-3. La structure cristalline

de la forme oxygénée de TrHbN (gauche, Accession PDB 1IDR, chaîne A) a été alignée

avec celle de la Mb oxygénée de cachalot (droite, Accession PDB 1AM6). La figure a été

réalisée à l’aide de PyMOL [27].

21

Tableau 1-3 Structures tridimensionnelles résolues chez les hémoglobines tronquées

Groupes Organismes Formes Méthodes*

Code

PDB Références

I (TrHbN)

Mtb

Fe2+

O2

Fe3+

CN-

Fe3+

CN-

X

X

X + Xe

1IDR

1RTE

1S56

[38]

[68]

[69]

Chlamydomonas

eugametos

Fe3+

CN- X

X + Xe

1DLY

1UVX

[36]

[69]

Paramecium caudatum

Fe3+

CN- X

X + Xe

1DLW

1UVY

[36]

[69]

Synechosystis sp.

bis-His

Fe3+

CN-

RMN

X

1MWB

1RTX

[70]

[62, 71]

II (TrHbO) Mtb

Fe3+

CN- X 1NGK [72]

Bacillus subtilis

Fe3+

CN- X 1UX8 [73]

Geobacillus

stearothermophilus

Fe3+

COO-

+ Fe2+

O2

X 2BKM [74]

Thermobifida fusca

Fe3+

COO- X 2BMM [75]

Agrobacterium

tumefaciens

Fe3+

OH2 X 2XYK [76]

III (TrHbP) Campylobacter jejuni Fe3+

CN- X 2IG3 [77]

* X : Cristallographie et diffraction des rayons X; Xe : sous haute pression de xénon.

22

1.2.8. Cavités internes et tunnels

Les structures des TrHbs contiennent des cavités hydrophobes (Figure 1.7 Cavités

observées dans la structure de quelques globines. Les cavités sont définies pas la zone

grillagée. A: Mtb TrHbN (1RTE, chaîne A) B: Chlamydomonas eugametos TrHbN (PDB :

1DLY), C: Paramecium caudatum TrHbN (PDB : 1DLW), D: Synechocystis sp. TrHbN

(PDB : 1S69), E : Mtb TrHbO (PDB 1NGK, chaîne G), F: Bacillus subtilis TrHbO

(PDB : 1UX8), G : Campylobacter jejuni TrHbP (PDB : 2IG3, chaîne A), H : Myoglobine

cachalot (PDB 1AM6)). Le volume total de ces cavités est nettement supérieur chez les

TrHbs du groupe I [69] (Tableau 1-4). En ce qui a trait aux TrHbs des groupes II et III, les

cavités observées sont de plus petites tailles et plus isolées, telles que celles présentes dans

la structure de la Mb [72, 77, 78].

Initialement, les cavités internes observées à l’intérieur des protéines n’étaient perçues que

comme des emboîtements imparfaits de chaînes latérales internes [79]. Cette description

simpliste s’est grandement bonifiée avec l’avènement de nouvelles méthodes telles que les

simulations de dynamique moléculaire [80] et la cristallographie en temps réel [81-83]. En

particulier, la MD a mis en évidence que les protéines bougent, chaque atome les

composant ayant un caractère diffusionel [80]. Les mouvements intramoléculaires sont

propices au remodelage de ces cavités et la formation de routes transitoires entre elles et

avec la surface de la protéine. Ainsi, il a été proposé que les ligands puissent diffuser de

cavités en cavité [84-90]. Ce phénomène a été non seulement décrit pour la Mb mais

également chez d’autres protéines [37, 91-96].

Chez la Mb, les cavités sont isolées les unes des autres et leur volume est modeste (Tableau

1-4). Il existe d’autres globines, dont les hémoglobines tronquées du groupe I de Mtb ou

encore celle de Chlamydomonas eugametos, où la structure renferme des cavités tellement

importantes qu’elles s’étirent de la surface au site actif [36-41]. On parle alors de tunnel. La

Figure 1.7 illustre les cavités observées chez quelques globines.

23

Tableau 1-4 Volume total interne des cavités de quelques globines

Groupes Organismes Formes Volume *

Å3

Code

PDB Références

I (TrHbN) M. tuberculosis Fe3+

CN- 299 (265) 1RTE [69]

Chlamydomonas

eugametos

Fe3+

CN- 465 (400) 1DLY [36]

Paramecium caudatum Fe3+

CN- 280 (180) 1DLW [36]

Synechosystis sp. Fe3+

CN-

357 1RTX [62, 71]

II (TrHbO) M. tuberculosis Fe3+

CN- 120 1NGK [72]

Bacillus subtilis Fe3+

CN- 124 1UX8 [73]

III (TrHbP) Campylobacter jejuni Fe3+

CN- 62 2IG3 [77]

Mb Myoglobine de cachalot Fe2+

O2 73 1AM6 [97]

* Volume estimé à partir d’une sonde de 1.4 Å de rayon. L’erreur sur la valeur calculée est

inférieure à 1%. La valeur entre parenthèses provient de la référence [69] alors que toutes

les autres ont été calculées. Les différences avec les valeurs publiées peuvent s’expliquer de

différentes façons : rayon de sonde utilisé, frontière de cavité en continuum avec la solvant

(entrée des tunnels), rayon de Van der Waals des différents types d’atomes.

24

La détermination de la structure de TrHbN a été marquée par la présence de deux

tunnels [38] (voir Figure 1.8). Le premier tunnel, appelé le tunnel Long, s’étire du site

distal pour rejoindre la surface de l’enzyme entre les boucles A-B et G-H. L’autre tunnel,

appelé le tunnel Court, s’étire du site distal jusqu’à la surface entre les hélices G et H. Ces

tunnels sont présentés plus en détail à la section 1.3. La présence de cavités s’alignant sur

l’axe du tunnel Long est une caractéristique commune à toutes les structures du groupe I

résolues jusqu`à maintenant [69] (Figure 1.7). Les fonctions qui ont été initialement

proposées pour ces tunnels seraient de permettre la diffusion sélective et le stockage de

substrats apolaires [38, 69].

Au moment de la publication de la structure de TrHbN, la présence des tunnels était

inusitée car ceux-ci n’étaient observés que chez des enzymes multimériques [98, 99]. Le

Tableau 1-5 présente une liste de différentes enzymes comportant un ou plusieurs tunnel(s).

La fonction associée à ces tunnels est toujours liée à la diffusion des substrats ou

d’intermédiaires réactionnels. La première structure d’enzyme révélant un tunnel fut la

tryptophane synthase, un α2β2-tétramère [100]. La sous-unité α catalyse le clivage de

l’indole-3-glycérol phosphate produisant un intermédiaire indole et une molécule de

glycérol-3-phosphate. La sous-unité β contient un site actif catalysant la condensation de

l’indole sur la L-sérine. Ces deux sites actifs, séparés de 25 Å, sont reliés par un tunnel

permettant la diffusion de la molécule d’indole de la sous-unité α vers la sous-unité β. Au

cours des années suivantes, d’autres structures de complexes enzymatiques, notamment

certaines synthases d’acides aminés, ont également révélé de tels tunnels [101-107]. Chez

ces protéines, le tunnel permet d’acheminer des intermédiaires réactionnels d’un site actif à

un autre et très souvent, il s’agit d’une molécule d’ammoniac [102, 104, 105, 107]. En plus

d’assurer la diffusion entre les sites actifs, ces tunnels évitent la perte d’un intermédiaire

réactionnel pouvant être instable et/ou toxique pour la cellule. Il existe d’autres protéines

chez lesquelles des tunnels relient la surface au site actif tel que certaines hydrogénases et

lipoxygénases [37, 91, 108-114] et quelques globines [36-41]. Le chapitre 2 est dédié à

l’étude de la diffusion des substrats à l’intérieur des enzymes.

25

Figure 1.7 Cavités observées dans la structure de quelques globines. Les cavités sont

définies pas la zone grillagée. A: Mtb TrHbN (1RTE, chaîne A) B: Chlamydomonas

eugametos TrHbN (PDB : 1DLY), C: Paramecium caudatum TrHbN (PDB : 1DLW),

D: Synechocystis sp. TrHbN (PDB : 1S69), E : Mtb TrHbO (PDB 1NGK, chaîne G),

F: Bacillus subtilis TrHbO (PDB : 1UX8), G : Campylobacter jejuni TrHbP (PDB : 2IG3,

chaîne A), H : Myoglobine cachalot (PDB 1AM6)

26

Figure 1.8 Structure tertiaire de TrHbN. Les tunnels, les sites de liaison du xénon et l’hélice

pré-A sont identifiés. L’espace des tunnels est représenté par la surface orange. Les hélices

B, E, G et H sont colorées respectivement en bleu, vert, jaune et mauve. Les tunnels ont été

calculés avec le programme CAVER [115] et la figure a été produite avec PyMOL [27].

27

Tableau 1-5 Tunnels observés chez diverses protéines

Protéine Organisme Diffusion Méthodes # PDB Référence(s)

TrHbN Mtb O2 / •NO X* 1IDR/1RTE [38]

Protoglobin Methanosarcina acetivorans O2 / •NO X 2VEB [40]

Neuroglobine Rongeur O2 / •NO X 1Q1F [40, 41]

MiniHb Cerebratulus lacteus O2 / •NO X 1KR7 [39, 116]

Oxydase Cu-amine Hanseluna polymorpha O2 / H2O2 X 1EKM [108]

12/15 Lipoxygénase Soja O2 X 2SBL/1YGE [110, 112]

Cholesterol oxydase Brevibacterium sterolicum O2 X 1I19 [113, 114]

Cytochrome c oxydase Bovin O2 X / DM 1OCC [37]

Cycloxygenase Ovisaries O2 DM 1PRH [111]

ACS-CODH † Moorella thermoacetica CO X 1JJY [101]

Hydrogenase Bactéries diverses O2 / H2 X Plusieurs [109]

Catalase Proteusmicrabilis O2 / H2O2 X 1M85 [91]

Tryptophane syntase Salmonella typhimurium Indole X 1BKS [100]

Carbamoyl phosphate syntase Escherichia coli NH3 / Carbamate X 1JDB [102]

GPA‡ Escherichia coli NH3 X 1ECF [103]

Asparagine synthétase Escherichia coli NH3 X 1CT9 [104]

Glutamate syntase Azospirillum brasilense NH3 X 1EA0 [105]

IGPS § Saccaromyces cerevisiae NH3 X 1JVN [106]

Glucosamine 6-phosphate

syntase

Escherichia coli NH3 X 1JXA [107]

* X : Cristallographie et diffraction des rayons X

† Acetyl-CoA synthase / CO déshydrogénase

‡ Phosphoribosylpyrophosphate amidotransferase

§ Imidazole glycerol phosphate syntase

28

1.3. Relation structure-fonction chez TrHbN de Mtb

L’hémoglobine tronquée N de Mtb est une globine ayant une très haute affinité pour l’O2.

En effet, cette affinité (Kd=7 nM) [117] est plus de 100 fois supérieure à celle de la Mb

(857 nM) [118]. Cette différence est principalement due au faible taux de dissociation de

l’O2 lié (koff = 0.2 s-1

) [117]. Il est proposé que cette haute affinité soit cruciale pour

permettre la détoxification efficace du •NO sous une faible tension en O2 tel qu’il

prévaudrait au sein du granulome [9]. Les structures tridimensionnelles de la forme

oxygénée (Fe+2

O2) et cyanomet (Fe3+

CN-) de TrHbN ont été déterminées par

cristallographie et diffraction des rayons X [38, 68]. Bien que la maille cristalline de ces

structures contienne deux molécules de TrHbN, des expériences de tamisage moléculaire,

de résonance magnétique nucléaire et de diffusion dynamique de la lumière soutiennent que

TrHbN est monomérique en solution (données non publiées).

La Figure 1.9 montre l’arrangement structural du site actif de TrHbN. L’angle formé par les

atomes Fe-O-O est penché par ≈ 120°, l’oxygène distal pointant en direction du résidu

Val94(G8). Les deux atomes d’oxygène sont équidistants de l’atome d’oxygène du

groupement phénol de la tyrosine en position 10 de l’hélice B (Tyr33(B10)). Le

groupement amide de la glutamine en position 11 de l’hélice E, la Gln58(E11), est

également à une distance permettant la formation de pont hydrogène avec la Tyr33(B10) et

le substrat lié. Les autres résidus complétant la cavité distale sont la Phe46(CD1), la

Leu54(E7) et la Val94(G8). Le mutant Tyr33(B10)Phe présente une constante de

dissociation de l’O2 près de 150 fois supérieure à celle de l’enzyme sauvage, suggérant que

la tyrosine stabilise l’O2 lié [22]. De récentes études en spectroscopie de résonance Raman

ont confirmé les interactions entre la Tyr33(B10) et l’O2 lié [56].

29

Figure 1.9 Site actif de TrHbN. Configuration de la cavité distale polaire de TrHbN sous sa

forme oxygénée. L’anneau porphyrique, l’O2 lié, la Tyr33(B10), la Gln58(E11), la

Phe62(E15) et l’His81(F8) proximale sont représentés en bâtons et boules. Les hélices α

sont colorées en bleu (hélice B), vert (hélice E) et mauve (hélice H). La Phe62(E15) montre

deux conformations dans la structure cristalline. Une seule d’entre elles est représentée

pour la clarté de l’image. Pour la même raison, l’hélice G n’apparaît pas. Les coordonnées

des atomes sont celles de la structure cristalline de TrHbN sous sa forme oxygénée

(Accession PDB 1IDR, chaîne A). La figure a été réalisée à l’aide de PyMOL [27].

30

Comme mentionné précédemment, la structure de TrHbN montre deux tunnels

hydrophobes reliant la surface et le site actif de l’enzyme. Ces tunnels sont nommés

« tunnel Court » et « tunnel Long » (Figure 1.8). Le tunnel Court se termine à environ 13 Å

du site actif à une position centrale entre les hélices G et H. Il est formé par les résidus

hydrophobes compris dans les hélices G (Phe91(G5), Val94(G8), Ala95(G9), Leu98(G12)

et H (Leu116(H8), Ile119(H11) et Ala120(H12)). L’entrée du tunnel Court se retrouve à la

base d’un large entonnoir hydrophobe. Pour ce qui est du tunnel Long, sa longueur fait

environ 20 Å. L’entrée du tunnel Long est définie par deux segments de la chaîne

principale formant les boucles A-B et G-H. Les résidus qui définissent l’intérieur du tunnel

Long sont également hydrophobes et sont issus des quatre hélices α formant le repliement

2-sur-2 (Ile19(A15), Ala24(B1), Ile25(B2), Val28(B5), Val29(B6), Phe62(E15),

Leu66(E19), Leu98(G12), Leu102(G16), Ala105(G19) et Ile115(H7)).

La structure obtenue à partir des patrons de diffraction de cristaux de TrHbN sous haute

pression de xénon a permis d’identifier cinq sites de liaison au xénon (Figure 1.8) [69].

Deux de ces sites se retrouvent dans le tunnel Court (Xe3 et Xe4), deux se trouvent dans le

tunnel Long (Xe1 et Xe5) et un dernier est positionné au point de rencontre des deux

tunnels (Xe2). Ces sites de liaison du xénon suggèrent que les tunnels de TrHbN jouent un

rôle dans la diffusion des substrats et des produits à l’intérieur de l’enzyme [69]. L’activité

NOD élevée de TrHbN, n’étant limitée pratiquement que par la diffusion des substrats,

appuie cette hypothèse.

Les tunnels Court et Long se rejoignent près de l’hème, soit à une cavité délimitée par la

chaîne latérale des résidus Phe32(B9), Gln58(E11), Phe62(E15), Leu98(G12) et

Ile119(H11) (Figure 1.10). Cette cavité correspond au site Xe2 et celle-ci communique

avec la cavité distale. La disposition des tunnels de TrHbN permet donc d’acheminer les

substrats directement au site actif de l’enzyme sans qu’aucun changement structural ne soit

requis [38]. En comparaison, la Mb n’a pas de tunnel évident et des fluctuations

structurales importantes sont nécessaires pour permettre un accès au substrat [119] (Figure

2.1).

31

Figure 1.10 Cavité distale de TrHbN sous sa forme oxygénée. L’hème, l’histidine

proximale en position F8, la Phe32(B9) et la Phe62(E15) sont représentés par des bâtons.

Les résidus distaux Gln58(E11) et Val94(G8) de même que l’O2 lié sont identifiés. Les

sites de liaison du xénon Xe5, Xe2 et Xe3 sont identifiés. Les coordonnées des atomes sont

celles de la structure cristalline de TrHbN sous sa forme oxygénée (Accession PDB 1IDR,

chaîne A). La figure a été réalisée à l’aide de PyMOL [27].

32

Il existe une autre caractéristique structurale qui suscite une attention particulière. Il s’agit

de la phénylalanine en position 15 de l’hélice E (Phe62(E15)). Cet acide aminé est

positionné au centre du tunnel Long et présente deux conformations lesquelles se

distinguent par une rotation autour de l’angle de torsion χ1 (Figure 1.11). Ces

conformations sont observées autant chez la forme oxygénée (Accession PDB 1IDR) [38]

que pour la forme cyanomet (Accession PDB 1RTE) [68]. Étant donné cette double

conformation, il a été proposé que ce résidu est mobile et qu’il pourrait jouer un rôle dans

la diffusion des substrats dans le tunnel Long en agissant comme une porte contrôlant le

flux des substrats et des produits [38].

1.4. Organisation de la présente thèse

La présente thèse de doctorat porte sur une étude théorique de la structure et de la

dynamique de l’hémoglobine tronquée N de Mycobacterium tuberculosis. Dans cette étude,

une attention particulière a été portée sur la dynamique des cavités et des tunnels de TrHbN

ainsi que sur le rôle de ceux-ci sur la diffusion des substrats vers le site actif.

Dans un premier temps, le prochain chapitre sera portera sur la diffusion et des substrats à

l’intérieur des protéines. Par après, les outils bio-informatiques employés, principalement la

dynamique moléculaire, seront présentés. Les objectifs de cette thèse ensuite énumérés.

Enfin, les six derniers chapitres suivants présenteront et discuterons tour à tour l’ensemble

des résultats obtenus.

33

Figure 1.11 Conformations alternatives de la Phe62(E15). La Phe62(E15) est représentée

par des bâtons et boules. Les sites de liaison du xénon Xe1, Xe2 et Xe5 sont identifiés.

L’angle χ1 est en « trans » (~180°) pour la conformation dont le cycle pointe vers Xe1.

L’angle χ1 est en « minus » (~ -60°) pour celle où le cycle aromatique pointe vers Xe2. Une

partie de l’hème apparaît sur la droite. Les hélices B, E, G et H sont colorées en bleu, vert,

jaune et mauve respectivement. Le tunnel Long est représenté par la zone grillagée. La

structure est celle de la forme oxygénée (Accession PDB 1IDR, chaîne A). Le tunnel a été

calculé avec le programme CAVER [115]. Les coordonnées des atomes sont celles de la

structure cristalline de TrHbN sous sa forme oxygénée (Accession PDB 1IDR, chaîne A).

La figure a été réalisée à l’aide de PyMOL [27].

34

2.

Chapitre 2

Étude de la diffusion interne des substrats

La structure tertaire de TrHbN présente deux tunnels hydrophobes reliant la surface et le

site actif (Figure 1.8). Il a été proposé que ces derniers jouent un rôle dans la diffusion des

substrats apolaire (O2, •NO) du solvant vers le site actif [38, 69]. Une grande part de des

travaux présentés dans cette thèse de doctorat se concentre sur cette hypothèse. Par

conséquent, ce chapitre porte sur la diffusion des substrats à l’intérieur des protéines et

présente différentes approches pour étudier ce phénomène.

La Mb est le modèle le plus étudié en ce qui a trait à la diffusion des substrats à l’intérieur

des protéines. Cette quête de savoir chez la Mb dure depuis maintenant près d’un demi-

siècle, soit depuis que sa structure ait été élucidée [66]. Ces nombreuses années de

recherches ont permis d’émettre plusieurs concepts généraux sur les relations entre la

structure et la fonction des protéines. Elles ont également motivé le développement de

plusieurs approches expérimentales et théoriques permettant de mieux comprendre la

diffusion des substrats se produisant entre le solvant et le site actif.

Dès l’obtention de la structure de la Mb, une première question a émergé : comment les

substrats gazeux atteignent la cavité distale en apparence isolée du solvant? D’abord, en

l’absence d’indices structurales montrant des routes de diffusions définies, il a été proposé

que les substrats gazeux peuvent accéder toutes les régions internes des protéines, même

celles considérées inaccessibles, grâce à des mouvements des protéines se produisant dans

l’échelle de temps nanoseconde [120]. Cette hypothèse a été par la suite réfutée par

plusieurs groupes de recherche qui ont démontré l’utilisation de routes de diffusion

particulières. Il existe deux grands modèles expliquant ce phénomène. Le premier consiste

au déplacement d’une chaîne latérale, l’histidine en position E7, vers le solvant (modèle de

la porte E7, Figure 2.1) [121]. Ce mouvement engendre la formation d’un tunnel reliant la

surface à la cavité distale permettant l’accès au site actif. Au début des années 80, ce

modèle a été complexifié suite à l’obtention de la structure de la Mb déterminée à partir de

35

cristaux placés sous haute pression de xénon [122]. Cette structure a révélé quatre sites de

liaison au xénon positionnés au cœur de la Mb (Figure 2.1). Puisque ces sites n’étaient pas

positionnés dans l’axe de la porte E7, un second modèle a été proposé. Ce dernier décrit

une diffusion interne des ligands le long de certaines routes précises ne passant pas par les

cavités de xénon (Figure 2.1) [85, 86, 88, 90, 122]. Afin de faire la lumière sur les

mécanismes de diffusion interne des substrats chez la Mb, de nombreuses méthodes

expérimentales et théoriques ont été développées. Ces travaux ont tantôt favorisé l’un ou

l’autre des modèles, tantôt les deux. Encore aujourd’hui, il n’existe pas de consensus.

Les prochaines sections décrivent les principales approches utilisées pour étudier la

diffusion des substrats chez les globines. Ces approches regroupent des méthodes

théoriques et expérimentales. Les méthodes expérimentales renferment les études

cinétiques de mutants, la diffraction des rayons X à température cryogénique et la

cristallographie de Laue résolue en temps réel. Quant à elles, les approchent théoriques

exploitent principalement les trajectoires issues de simulation de dynamique moléculaire.

36

Figure 2.1 Modèles de diffusion des substrats chez la Mb. Les coordonnées proviennent de

la structure de la Mb déterminée à partir de cristaux traités sous haute pression de xénon

(Accession PDB 1J52). La figure du haut montre la position des sites xénon par rapport à la

structure de la Mb. Les hélices B, E, G et H sont colorées respectivement en bleu, vert,

jaune et violet. Les sites xénon, l’hème, l’histidine E7 sont identifiés dans la figure du bas.

La figure a été réalisée à l’aide de PyMOL [27].

37

2.1. Cinétiques enzymatiques de mutants

Des essais en laboratoire sont nécessaires pour tester les hypothèses émises en ce qui a trait

aux voies de diffusion qui serait empruntées par les molécules de substrats. Le plus

couramment, des résidus prenant place le long des routes de diffusion proposées sont

mutés. Les mutations sont réalisées de manière à obstruer ou à élargir une voie d’accès

ciblée [59, 93, 96, 123-125]. La caractérisation cinétique des mutants créés est ensuite

réalisée pour mesurer l’impact des mutations. Idéalement, les mutations ne doivent

qu’obstruer les routes de diffusion sans en affecter la réactivité du site actif. Deux types de

caractérisations cinétiques sont généralement employés chez les globines: les cinétiques de

recombinaison et les cinétiques de liaison. On peut également suivre d’autres réactions

catalysées telles que la NOD.

2.1.1. Cinétiques de recombinaison

Pour les cinétiques de recombinaison, on exploite une propriété biochimique des protéines

hémiques, soit la photosensibilité du lien fer-ligand [126]. En effet, en soumettant un

échantillon à une brève et intense impulsion laser, on provoque la rupture du lien fer-

ligand. Cette propriété est très utile puisque les espèces liées et non liées présentent des

propriétés optiques différentes. Suivant sa photodissociation, un ligand peut soit diffuser à

l’extérieur de la protéine, soit se recombiner au fer. Le phénomène par lequel un ligand

photodissocié revient se fixer au fer est appelé la recombinaison géminée. Il est donc

possible de suivre dans le temps le retour vers la forme liée. Ce phénomène peut se

décomposer en différentes phases en fonction du temps (Figure 2.2). Lorsqu’un ligand

photodissocié demeure près du fer, celui-ci est susceptible de se recombiner très

rapidement. On parle alors de la phase géminée rapide. Lorsque le ligand s’éloigne

davantage du fer, le ligand est susceptible d’être piégé transitoirement à l’intérieur de

cavités. Le retour du ligand est alors retardé et on peut suivre dans le temps cette phase.

Enfin, une dernière phase concerne les ligands arrivant du solvant. La vitesse de cette phase

38

est fonction de la concentration du substrat à l’extérieur de la protéine. On parle alors de

cinétique bimoléculaire. Pour influencer l’une ou l’autre de ces phases, des mutations

peuvent être créées. Celles-ci sont alors susceptibles d’accélérer ou de retarder certaines

phases. Par exemple, à la figure Figure 2.2 (graphique de droite) montre un cas

hypothétique où une mutation limite la diffusion entre le fer et les cavités internes. Dans ce

cas, une plus grande proportion des ligands se combinent à nouveau au fer. Toutefois, pour

ce qui est des ligands s’étant éloignés davantage et étant contenu dans une ou plusieurs

cavités internes distantes, ceux-ci se recombinent moins fréquemment et cette phase se

trouve ralentie.

39

Figure 2.2 Cinétiques de recombinaison hypothétiques pour une globine donnée sous sa

forme sauvage (gauche) et pour un mutant (droite). Le moment de la photolyse du ligand

est indiqué par la flèche et les phases de recombinaison subséquentes sont identifiées. Chez

le mutant, la diffusion des ligands entre le site actif et les cavités internes est altérée. Dans

cet exemple schématique, la forme dissociée (5C) absorbe davantage que la forme associée

(6C) à la longueur d’onde utilisée. Ainsi, le bris du lien fer-ligand provoqué par l’impulsion

laser engendre une augmentation presque instantanée de l’absorbance. Dans le mutant, une

plus grande proportion des ligands photodissociés demeurent près du fer et se recombinent

rapidement. Pour la fraction des ligands ayant exploré la matrice enzymatique, les phases

subséquentes sont ralenties.

40

2.1.2. Cinétiques de liaison

Les cinétiques de liaison concernent l’étude de la diffusion du solvant vers le site actif du

ligand et la formation du lien avec le fer. Pour ce faire, on fait réagir un ligand gazeux avec

une préparation de la globine sous sa forme 5C (ferreux ou ferrique dépendamment du

ligand) et on mesure les vitesses de formation de la liaison fer-ligand. En portant en

graphique la vitesse de liaison en fonction de la concentration du substrat, on obtient la

constante bimoléculaire. Plus cette constante est élevée, plus rapidement le substrat atteint

le site actif. Dans certains cas, l’association du ligand au fer n’est limitée que par la vitesse

diffusion des molécules. Ceci indique que l’accès au fer est libre d’obstacle stérique et que

le fer est très réactif.

Chez la Mb, les cinétiques de liaison et de recombinaison ont été étudiées chez un grand

nombre de mutants [123, 124]. Les mutations ont été créées pour les résidus distaux et pour

un grand nombre de résidus éloignés du site actif. Certaines mutations ont provoqué des

effets plus importants que d’autres. Parmi celles-ci, le changement de l’histidine en position

E7 (His(E7)) pour un acide aminé de petite taille tel que l’alanine provoque l’accélération

des vitesses de liaison et favorise l’échappement du ligand photodissocié vers le solvant.

Les résultats de ces travaux suggèrent que les substrats transitent majoritairement par une

route contrôlée par la porte His(E7). Or, d’autres mutations éloignées du site actif ont

également provoqué des effets modérés. Ces mutations concernent en particulier les résidus

bordant des cavités qui correspondent aux sites Xe1 et Xe4 (Figure 2.1).

Pour obtenir des interprétations structurales et dynamiques de la diffusion interne des

ligands, d’autres méthodes doivent être employées. Ces méthodes renferment la diffraction

des rayons X à températures cryogéniques, la cristallographie de Laue résolue en temps réel

et les simulations de dynamique moléculaire.

41

2.1.3. Exemples d’applications

Chez les enzymes démontrant un seul tunnel spécifique, la mutagenèse dans le but de

restreindre la diffusion des substrats est généralement efficace afin de mettre en évidence

son rôle dans l’accès au site actif [96, 127-132]. La cytochrome c oxydase en est un bon

exemple [129]. Cette protéine présente dans sa structure un tunnel hydrophobe qui serait

favorable à la diffusion des ligands gazeux. D’autres routes sont également notées mais

celles-ci sont hydrophiles et permettraient la diffusion de l’eau et le transfert de protons.

Une seule substitution Gly→Val créée dans le but de bloquer la route hydrophobe a ralenti

les évènements de combinaison de l’O2 et du CO au site actif par plusieurs ordres de

magnitude [129]. Pour expliquer cet effet drastique, les auteurs ont postulé que la structure

de cette enzyme devait être très rigide près du site muté empêchant que d’autres routes se

forment dans le temps. Ceci donc confirmait en même temps que les routes hydrophiles

sont obstruées par des molécules d’eau.

Chez les protéines dont les structures révèlent plusieurs routes, les conclusions sont moins

claires. Chez la 12/15-lypoxygénase, seule une route parmi les trois prédites à l’aide de

calculs théoriques (échantillonnage implicite de ligands ou ÉIL) a été confirmée

expérimentalement [96]. Cette route confirmée permet à l’oxygène moléculaire d’atteindre

le site actif alors que les deux autres sont bloquées par le substrat, c’est-à-dire l’acide

linoléique. Chez l’amine oxydase à cuivre (AOC), deux chemins majeurs pour la diffusion

de l’O2 ont été identifiés par ÉIL[93]. Ces deux chemins se rejoignent au site actif de

l’enzyme. Des mutants créés dans le but de bloquer ces chemins individuellement a

provoqué de très faibles effets sur le ratio kcat/km signifiant que d’autres chemins

existent [93]. Pour expliquer ces résultats, les auteurs de ces travaux ont émis l’hypothèse

que d’autres routes peuvent se former suite à des mouvements dans la structure de

l’enzyme.

42

2.2. Diffraction des rayons X à températures cryogéniques

En 1994, la diffraction des rayons X à températures cryogéniques a permis les premières

observations expérimentales révélant la diffusion interne des substrats gazeux au cœur de la

Mb [133]. Avec cette méthode, on place d’abord les cristaux à très basses températures (10

à 20 K). Ensuite, le cristal est soumis à une impulsion laser afin de briser le lien fer-ligand.

Le bris du lien confère suffisamment d’énergie cinétique au ligand pour que celui-ci

explore momentanément la matrice protéique. Le ligand est alors susceptible d’être piégé à

l’intérieur d’une cavité interne. La température très basse limite par la suite grandement la

diffusion subséquente du ligand. L’acquisition des données de diffraction est ensuite

réalisée. Chez la Mb, ces travaux ont permis confirmer que le ligand photodissocié occupe

les différentes cavités correspondantes à différents sites xénon [119, 133-135]. Cette

méthode a servi de base pour le développement d’une autre technique beaucoup plus

puissante; la cristallographie de Laue résolue en temps réel.

2.3. Cristallographie de Laue résolue en temps réel

L’arrivée de la cristallographie de Laue résolue en temps réel constitue une avancée

majeure dans la compréhension de la dynamique des protéines et de la diffusion des

ligands [81-83]. Cette technique est devenue disponible au milieu des années 90 notamment

grâce à l’arrivée des synchrotrons de troisième génération. Ces derniers permettent de

générer des impulsions de rayon X très intenses en un temps court, soit environ 100

picosecondes (ps)) [82]. Cette technique permet de suivre dans le temps la position d’un

ligand suite à sa photodissociation à des températures ambiantes. En plus de la diffusion du

ligand, elle permet de caractériser en détail la relaxation de la protéine après la

photodissociation. Les échelles de temps couverts par cette méthode s’échelonnent de

100 ps à quelques millisecondes (ms). Comme avantage majeur, cette technique permet de

couvrir la plupart des différents mouvements se produisant chez les protéines (Figure 2.3).

43

Avec cette technique, le cristal est d’abord soumis à une impulsion laser de 7 ns provoquant

le bris du lien fer-ligand. L’acquisition des patrons de diffraction se fait à des intervalles de

temps prédéterminés après l’impulsion laser. Ces patrons de diffraction peuvent être

enregistrés plusieurs fois pour un même cristal, mais puisque les rayons X endommagent

les cristaux, plus d’un peuvent être nécessaires [82, 136]. Une fois les données analysées,

on obtient un film à l’échelle atomique montrant la diffusion du ligand photodissocié et de

la relaxation de la protéine. Comme découverte importante chez la Mb, il a été montré que

les hélices E et F relaxent rapidement suivant la photodissociation (~1 ns). Les molécules

de CO occupent principalement deux régions, soit la cavité distale et la cavité Xe1 [136].

L’occupation maximale du CO au site Xe1 survient 100 ns après la photodissociation de

concert avec la réorganisation d’une chaîne latérale (Leu89). Ces travaux n’ont pas révélé

d’augmentation de la densité électronique aux sites Xe2, Xe3 et Xe4. En ce qui concerne la

porte E7, ces travaux n’ont pas révélé de mouvement permettant le passage des ligands vers

l’extérieur [83, 136].

44

Figure 2.3 Échelles de temps des différents mouvements se produisant dans les protéines.

Les différentes méthodes permettant pour leur étude sont indiquées. Figure adaptée

de [137].

45

2.4. Simulations de dynamique moléculaire

La dynamique moléculaire (DM) permet de simuler le comportement d’un ensemble

d’atomes et de molécules en fonction du temps [80]. Les échelles de temps couvertes par la

DM s’échelonnent de la femtoseconde à la microseconde (Figure 2.3). Puisque la DM est la

méthode principale des travaux décrits dans cette thèse, le prochain chapitre y est dédié.

La DM a permis des avancées remarquables dans la compréhension de la diffusion des

ligands chez la Mb. D’abord, en 1979, l’utilisation de la porte E7 a été remise en question

par Case et Karplus [138]. Ces derniers s’étaient basés sur des trajectoires de DM et des

minimisations d’énergie montrant que le déplacement de l’His(E7) implique un coût trop

élevé en énergie. Ce faisant, l’ouverture de la porte His(E7) serait peu fréquente et

nécessiterait des fluctuations structurales importantes [138]. Deux autres voies de diffusion

à l’intérieur de la matrice protéique ont été proposées. Des travaux subséquents de DM et la

découverte des sites xénons ont également permis de proposer diverses routes de diffusion

internes [85, 88, 90, 122]. En 1990, Elber et Karplus ont précisé le modèle en montrant que

les ligands gazeux diffusent de cavité en cavité à l’intérieur de la Mb [88]. Ces derniers

avaient alors développé une nouvelle technique de simulation biaisée (« Locally Enhanced

Sampling ») favorisant l’exploration du ligand gazeux au cœur de la protéine. Cette

méthode a été utilisée dans cette thèse et elle est décrite plus en détail au prochain chapitre.

En 2004, le groupe d’Alfredo di Nola a été le premier à décrire les résultats d’une

simulation non biaisée de la diffusion du CO à l’intérieur de la Mb. Cette trajectoire, d’une

durée de 90 ns, mimait la diffusion du CO après sa photolyse. Les résultats de ces travaux

étaient grandement en accord avec les données de cristallographie de Laue résolue en temps

réel. Leurs travaux ont également permis de montrer que la diffusion du CO

s’accompagnait de la réorganisation de certaines chaînes latérales. Enfin, une importante

étude a été récemment publiée par Ruscio et coll [90]. Dans cette étude, un impressionnant

total de 68 simulations de 90 ns chacune ont été réalisées dans le but d’étudier la diffusion

du CO entre le solvant et le site actif. Dans 48 de ces simulations, le CO était dans le

solvant au départ des simulations et 16 évènements d’entrée ont été observés. Neuf

différents points d’entrées à la surface de la Mb ont été identifiés dont la porte E7. Une fois

46

internalisé dans la matrice de la Mb, le CO diffuse de cavité en cavité, en accord avec les

travaux précédents [85, 88]. Une approche similaire a également été menée dans les travaux

de cette thèse pour étudier la diffusion du •NO du solvant vers le site actif de TrHbN. Ces

travaux font l’objet du chapitre 7.

47

3.

Chapitre 3

Méthodologie

Le projet de doctorat présenté a été réalisé à l’aide d’une série d’outils bio-informatiques.

Le principal outil employé a été la dynamique moléculaire. D’autres programmes utiles à

l’étude des cavités et tunnels à l’intérieur de protéines ont également été utilisés. Ce

chapitre est dédié à présenter ces outils.

3.1. La dynamique moléculaire

La dynamique moléculaire est une méthode permettant de simuler numériquement le

comportement d’un ensemble de molécules et d’atomes en fonction du temps [80]. Elle est

employée dans plusieurs domaines de recherche. Parmi ceux-ci, notons la physique, la

science des matériaux et plus particulièrement la biochimie et la biophysique [139]. Avec

la puissance actuelle des ordinateurs, ces simulations permettent d’étudier des phénomènes

s’échelonnant de la femtoseconde (fs) jusqu’à la microseconde (μs). La dynamique

moléculaire (DM) est utilisée afin de répondre à plusieurs questions qui sont partiellement

accessibles, voire même inaccessibles, expérimentalement. En plus d’aider à

l’interprétation de données expérimentales, la DM sert également à émettre et à tester de

nouvelles hypothèses. Étant donné que la DM est la principale méthode utilisée au cours de

cette thèse de doctorat, celle-ci sera exposée plus en détail dans les prochaines sections.

3.1.1. Histoire de la dynamique moléculaire

Dès ses premiers pas, la DM a permis des avancées scientifiques importantes. La

méthodologie a été introduite par Alder et Wainwright à la fin des années 50 [140, 141].

48

Leurs travaux ont permis d’étudier le comportement d’un ensemble de sphères dures et

leurs simulations ont permis de faire des liens avec le comportement des liquides. En 1964,

un bond important a été fait avec la publication des résultats d’une simulation de l’argon

liquide [142]. En effet, ces simulations étaient plus réalistes car les atomes de xénon étaient

représentés à l’aide de sphères pénétrables (potentiel de Lennard-Jones). Les premières

simulations réalistes de l’eau liquide ont suivi 10 ans plus tard [143]. Dans ces travaux, les

molécules d’eau étaient représentées par un modèle à quatre charges. Ce modèle a permis

de reproduire plusieurs propriétés physiques de l’eau telles que la densité et le coefficient

de diffusion à différentes températures. Ce n’est qu’en 1977 que la première simulation

d’une protéine a été publiée par le groupe de Martin Karplus [80]. La protéine simulée était

l’inhibiteur trypsique pancréatique de bovin, une petite protéine monomérique composée de

58 acides aminés. Martin Karplus a qualifié cette protéine comme « l’atome d’hydrogène »

de la DM étant donné sa petite taille, sa grande stabilité et la disponibilité de sa structure

tridimensionnelle à une haute résolution (1.5Å) [144]. La trajectoire produite, d’une durée

de 9.2 ps et en absence de solvant, a permis de montrer que cette protéine bouge. Malgré le

temps de simulation court et le champ de force encore primitif, ces travaux ont permis de

révolutionner la biochimie structurale, remettant en question la conception des protéines en

tant que molécules rigides. Cette découverte a appuyé une nouvelle hypothèse reliant la

flexibilité des macromolécules biologiques à leurs propriétés biochimiques et leur

fonction [139]. Cette hypothèse a été confirmée par différents travaux. Par exemple, il a été

démontré que l’accès des substrats au site actif du lysozyme et de l’alcool déshydrogénase

est contrôlé par le mouvement d’une boucle [145, 146]. De même, la DM a permis de tester

et de proposer divers mécanismes reliant la dynamique de la Mb et la diffusion de l’O2 et

du CO entre le solvant et son site actif [88-90, 92, 138].

Bien que la DM soit apparue dans les années cinquante, son premier véritable essor est

survenu durant les années 80. Plusieurs phénomènes biophysiques ont été étudiés chez

diverses protéines et acides nucléiques. En plus de ces travaux, la DM s’est révélée un outil

précieux pour l’interprétation des paramètres de relaxation en résonance magnétique

nucléaire (RMN) [147, 148]. Également, un des premiers sujets traités par la DM est l’effet

de la température sur la structure et la dynamique des protéines [149-151]. Enfin, la DM

49

fait partie intégrante du raffinement des structures obtenues par cristallographie et

diffraction des rayons X et par RMN. Au cours des 20 dernières années, la DM s’est

grandement développée et popularisée d’une part grâce à l’amélioration des algorithmes de

calculs et d’autre part, à l’amélioration fulgurante du matériel informatique : accroissement

de la puissance des processeurs, parallélisation des ordinateurs, capacité de stockage

grandement augmentée, réseaux de plus en plus rapides. Ces améliorations ont permis

l’utilisation de modèles plus rigoureux, de simuler des systèmes de plus en plus grands et

d’étendre considérablement les temps de simulations [152]. L’applicabilité de la DM a du

même coup augmenté en permettant l’étude de mouvements se produisant sur de plus

grandes échelles de temps et de plus grande amplitude. Au début des années 90, les

systèmes typiques comprenaient quelques milliers d’atomes et les temps de simulations

pouvaient atteindre la centaine de picosecondes (ps). La plupart des simulations étaient

réalisées dans le vide. Dix ans plus tard, les systèmes comprenaient typiquement jusqu’à

quelques dizaines de milliers d’atomes et les temps de simulations pouvaient atteindre

quelques nanosecondes (ns) [152]. Ces simulations employaient également un solvant

explicite, c’est-à-dire en présence de molécules d’eau. Au début de 2010, le temps de

simulation dépasse couramment 100 ns et les systèmes comptent plusieurs dizaines de

milliers d’atomes.

3.1.2. Principes de base

La DM est une application dérivée de la mécanique moléculaire (MM). La MM est apparue

en 1930 [153] mais celle-ci ne s’est véritablement développée qu’à partir de la fin des

années cinquante, avec l’augmentation de la puissance des ordinateurs. La MM se base sur

l’approximation de Born-Oppenheimer [154]. Dans cette dernière, puisque les électrons ont

une vitesse grandement supérieure à celle des noyaux, ceux-ci sont implicitement décrits.

En MM, les atomes sont décrits comme des sphères ayant une masse, un rayon et une

charge. Le rayon est nécessaire au calcul du potentiel de Lennard-Jones alors que la charge

est requise pour le calcul de l’énergie électrostatique. Lorsque des molécules composent le

50

système simulé, des termes additionnels sont inclus pour représenter l’énergie interne. Ces

termes sont les énergies de liaison, d’angle et de torsion (dièdre). Le potentiel d’un système

donné est ainsi calculé à l’aide d’une fonction empirique appelée « champ de force » qui

inclut tous les termes d’énergie. Il existe plusieurs champs de forces, certains étant mieux

adaptés pour simuler certains types de système. Dans ce qui suit, nous nous concentrerons

sur celui qui a été utilisé pour les travaux effectués dans le cadre de cette thèse, soit

CHARMM22 [155]. Ce champ de force a été préféré pour les travaux de cette thèse

puisque celui-ci est optimisé pour la simulation des protéines.

La DM vise à simuler le comportement d’un ensemble d’atomes à une température finie en

fonction du temps. La DM s’appuie sur la seconde loi du mouvement Newton qui relie la

force avec l’accélération d’une particule et sa masse. L’accélération est liée au changement

de vitesse de la particule dans le temps.

(équation 3-1)

La position des atomes est connue au départ de la trajectoire. On utilise le plus souvent les

coordonnées issues d’une structure expérimentale. L’objectif de la DM est de déterminer

quelle sera la position de ces mêmes atomes au temps t + dt. En connaissant la vitesse de

chaque atome, on peut déterminer leur position dans le temps. Cependant, étant donné que

le système contient plusieurs atomes, il faut que dt soit court pour éviter des collisions entre

les atomes ou encore l’adoption de géométries non favorables comme par exemple un lien

covalent trop étiré. Pour les molécules étudiées en DM, le mouvement le plus rapide est la

vibration des liens C-H; pour une bonne intégration des équations du mouvement, il faut

choisir un intervalle de temps de 1/10 de cette fréquence, soit 1fs. La force est calculée en

dérivant une fonction d’énergie potentielle appelée communément champ de force.

( )

(équation 3-2)

51

La fonction d’énergie potentielle utilisée par CHARMM [156] est donnée par :

(équation3-3)

( ) ∑ ( )

∑ ( )

∑

( ( ))

∑

( ) ∑ ( )

∑

(( )

( )

) ∑

Les différents termes contenus dans cette équation sont explicités dans les prochaines

sections.

3.1.2.1. L’énergie interne

L’énergie interne est fonction de la géométrie des molécules composant le système.

Différents types de mouvements affectent cette géométrie (Figure 3.1). L’énergie de liaison

fluctue en fonction du mouvement de vibration entre deux atomes liés. L’énergie d’angle

est fonction du mouvement de cisaillement entre trois atomes liés. Quant à l’énergie

d’angle dièdre, celle-ci est fonction de l’angle décrit par deux plans définis par quatre

atomes liés. Par exemple, pour une phénylalanine, l’angle χ1 correspond à l’angle entre le

plan des atomes N-Cα-Cβ et le plan des atomes Cα-Cβ-Cγ. L’énergie d’angle dièdre impose à

la molécule d’adopter préférentiellement certaines conformations dans l’espace, celles-ci

correspondent à des minimums d’énergie locaux. Pour la chaîne latérale des acides aminés,

les différentes combinaisons favorables d’angles dièdres portent le nom de rotamères [157].

Le champ de force de CHARMM inclut d’autres termes d’énergie pour satisfaire certaines

contraintes structurales. L’énergie d’angle impropre sert à satisfaire différentes contraintes

structurales. Elle sert notamment à préserver la chiralité et la planarité de certaines

molécules. Le terme Urey-Bradley est un potentiel harmonique qui est fonction de la

distance entre atomes séparés par deux liens (interactions 1-3). Les termes d’énergie pour

les liaisons, les angles et les angles impropres se calculent selon un potentiel harmonique

alors que l’énergie d’angle dièdre se calcule à l’aide d’un terme

trigonométrique (équation 3-3).

52

Figure 3.1 Mouvements internes dans les molécules. Les mouvements de vibrations, de

cisaillement et de torsion sont illustrés.

3.1.2.2. L’énergie externe

L’énergie externe concerne les interactions entre atomes non liés et généralement séparés

par plus de 2 ou 3 liens. Ces interactions, appelées « interactions non liées », se

décomposent en l’énergie électrostatique (Coulomb), le potentiel de Lennard-Jones (L-J) et

les liaisons hydrogènes. L’énergie électrostatique entre deux atomes est fonction de la

charge portée par chaque atome. Deux atomes non liés de charges opposées vont s’attirer.

Ils se repousseront dans le cas contraire. Le potentiel de Lennard-Jones est aussi connu sous

l’appellation « potentiel 6-12 ». Le terme à la puissance 6 est la composante attractive

communément appelée forces de Van der Waals. Le terme à la puissance 12 est la

composante répulsive et celui-ci devient dominant à mesure que la distance devient courte

par rapport à la distance optimale (exclusion de Pauli). En ce qui a trait aux ponts

hydrogènes, les premiers champs de force utilisaient une variante du potentiel de

Lennard-Jones, soit le potentiel 10-12 :

((

)

(

)

) (équation 3-4)

Aujourd’hui, les champs de forces modernes traitent la liaison hydrogène à même l’énergie

électrostatique et le potentiel 6-12. Cette simplification permet de réduire les temps de

calcul sans réduire la précision du champ de force.

53

3.1.2.3. Les méthodes de troncations

Le nombre d’interactions non liées à calculer dans un système contenant plusieurs milliers

d’atomes représente une lourde charge informatique. En effet, la complexité de ce calcul est

de l’ordre Ο(N2) lorsque toutes les paires d’atomes non liés sont considérées. Pour réduire

cette complexité, on fait appel à des méthodes de troncations [158]. Ainsi, seules les paires

d’atomes séparés d’une distance inférieure à une limite fixée sont considérées.

Généralement, les distances de troncation sont de l’ordre de 8 Å à 12 Å [152]. L’utilisation

des méthodes de troncation réduit la complexité du calcul à Ο(N). Le fait d’annuler

abruptement le potentiel d’interaction au-delà de la limite de distance engendre des effets

non désirables sur la validité de la trajectoire [159]. En effet, cette manière de procéder

introduit une discontinuité dans les forces à la distance de troncation empêchant l’énergie

d’être conservée [152]. Pour éviter ce problème, il existe différentes méthodes où l’énergie

est amenée progressivement à zéro [158]. La méthode « Switch » amène le potentiel

d’interaction progressivement à zéro sur un court intervalle de distance prédéterminé,

généralement sur les derniers 2 à 4 Å [158]. Il existe également la méthode « Shift » où le

potentiel d’interaction est progressivement amené à zéro sur l’ensemble de la distance

considérée [158]. Il existe des variantes de ces méthodes où les potentiels sont altérés de

manière à annuler progressivement la force (méthodes « FSwith » et « FShift ») [158]. Les

méthodes de troncation sont convenables pour l’énergie de Van der Waals lors de la

simulation de système employant un solvant explicite car elle devient négligeable au-delà

de la distance de troncation.

Par contre, pour certaines paires d’atomes non liés, le potentiel électrostatique demeure non

négligeable à la distance de troncation. Ce problème peut être évité en employant la

méthode connue sous l’appellation anglaise « Particle Mesh Ewald ». Cette méthode est

présentée à la prochaine section.

54

3.1.2.4. La méthode « Particle Mesh Ewald »

La méthode connue sous l’appellation anglaise « Particle Mesh Ewald » (PME), est une

variante de la sommation d’Ewald [160]. Cette méthode a révolutionné la dynamique

moléculaire en considérant efficacement les interactions électrostatiques à longue distance.

Il a été démontré qu’il est important d’inclure ces interactions dans différents

systèmes [161], notamment pour la simulation des acides nucléiques. Son emploi augmente

par contre légèrement le temps de calcul, sa complexité étant de l’ordre Ο(N•log N).

Cette méthode requiert des systèmes avec conditions périodiques aux frontières (explicité à

la prochaine section, Figure 3.2). La méthode PME s’effectue en subdivisant d’abord le

système en une fine grille tridimensionnelle (résolution ~ 1 Å3). En fonction de la position

des atomes du système, une charge est assignée à chaque point de la grille. L’énergie

électrostatique est calculée en deux sommes. La première somme concerne les interactions

de courte portée qui sont explicitement calculées. Pour cette étape, les interactions à

calculer sont déterminées à partir d’une distance de troncation déterminée. La seconde

somme est réalisée par une transformée de Fourier rapide (FFT) pour obtenir le potentiel

électrostatique en fonction de la distribution des charges sur la grille. Pour chaque point de

la grille, l’énergie est calculée en différenciant numériquement le potentiel. La force

s’exerçant sur chaque atome est calculée en interpolant l’énergie sur la grille.

3.1.2.5. Conditions périodiques aux frontières

De manière courante, les simulations de protéines sont réalisées en présence d’un solvant

explicite, c’est-à-dire en incluant les molécules d’eau dans le système. Le système est alors

contenu à l’intérieur d’une boîte de simulation dont la géométrie permet l’utilisation de

conditions périodiques aux frontières (Figure 3.2). Pour appliquer des conditions

périodiques aux frontières, il suffit de répliquer le système simulé de manière à remplir

complètement l’espace tridimensionnel autour de ce système. Les répliques portent le nom

d’images. Lorsqu’une molécule sort à l’extérieur de la boîte de simulation, la molécule est

remplacée par la molécule correspondante provenant d’une image voisine. Ainsi, le nombre

55

de particules dans le système demeure constant. Les interactions non liées aux frontières

tiennent compte des particules contenues dans les images voisines.

Figure 3.2. Conditions périodiques aux frontières d’un système en 2D. Le système simulé,

au centre, est répliqué 8 fois. Dans un système cubique, 26 images sont requises.

3.1.3. Production de la trajectoire

Comme mentionné précédemment, l’objectif de la DM est de simuler le mouvement des

atomes et des molécules dans le temps. Grâce au champ de force, on peut résoudre

l’équation du mouvement en dérivant la fonction d’énergie potentielle :

( )

(équation 3-5)

Pour connaître la nouvelle position, il faut intégrer la force :

∬ ∬

(équation 3-6)

56

L’intégration numérique se fait sur des intervalles de temps très courts. L’intervalle de

temps (Δt), ou pas de temps, est fixé de manière à ce qu’il soit plus court que la période du

mouvement avec la plus haute fréquence. Pour les systèmes biologiques, la vibration des

liens covalents impliquant un atome d’hydrogène présente la plus courte période de temps

(~10-14

s), donc utilisation d’un temps d’intégration de 10-15

s). Pour atteindre des

simulations plus longues sans augmenter le temps de calcul, il est avantageux de restreindre

ces liens à leur longueur d’équilibre. Ceci permet d’utiliser un pas de temps (Δt) plus long.

Il existe différentes méthodes pour y parvenir, SHAKE étant celle utilisée avec CHARMM

[162]. En utilisant ces approximations, le pas de temps permis est alors typiquement de une

à deux femtoseconde(s) (fs).

Pour obtenir une trajectoire, l’intégration se fait de manière itérative jusqu’à l’obtention de

la durée de simulation désirée. Ce processus peut se faire selon divers algorithmes

informatiques appelés « intégrateurs ». L’intégrateur le plus couramment utilisé est celui de

Verlet (« leapfrog ») [163]. La méthode de Verlet est basée sur une série de Taylor et se

décompose en trois étapes (Figure 3.3). Au départ du calcul, la position de chaque atome

est connue et la force peut être calculée par la fonction d’énergie potentielle (équation 3-5),

et une vélocité de départ est assignée à chaque atome. Ensuite la vitesse au temps t+Δt/2 est

obtenue par :

(

) (

)

( ) (équation 3-7)

On se sert ensuite de la force et des positions au temps t et des vitesses estimées aux temps

t-Δt/2 et t+ Δt/2 pour déterminer la position au temps t+Δt.

( ) ( )

[ (

) (

)]

( )

(équation 3-8)

Le processus est ensuite recommencé un grand nombre de fois jusqu’à la durée de

trajectoire désirée.

57

Figure 3.3 Algorithme de Verlet pour l’intégration de l’équation de mouvement. La

première étape consiste à calculer les forces en fonction de la position de tous les atomes. À

la deuxième étape, les forces et les masses des atomes sont utilisées pour calculer les

vitesses aux temps t ± Δt/2. À partir des vitesses, on peut déterminer la nouvelle position

des atomes au temps t + Δt.

58

3.2. Simulation des protéines par dynamique moléculaire

La simulation d’une protéine par DM se fait selon quatre grandes étapes (Figure 3.4)

Figure 3.4 Schéma des grandes étapes de la dynamique moléculaire

3.2.1. Préparation du système

La première étape consiste à la préparation du système que l’on veut simuler. Le système

idéal est celui qui reflète le plus fidèlement les conditions expérimentales pour lesquelles

nous avons des données (ex. force ionique, pH, concentration du substrat, état de

protonation des acides aminés ionisables). Pour construire ce système, il faut d’abord

obtenir les coordonnées cartésiennes de la protéine. Le plus souvent, on utilise une structure

déterminée expérimentalement, le plus couramment par cristallographie et diffraction des

rayons X. Si les coordonnées expérimentales ne sont pas disponibles, celles-ci peuvent être

parfois obtenues par modelage par homologie ou à l’aide d’outils de prédiction de

structures tridimensionnelles. Dans ce dernier cas, la structure obtenue peut contenir des

erreurs se répercutant sur le calcul de DM et ainsi, sur la validité des données que l’on en

tirera. L’état de protonation des résidus ionisables doit être fixé en fonction de leur pKa

respectif et le pH expérimental. Le plus souvent, les protéines sont simulées en présence

d’un solvant explicite. Dans ce cas, la protéine est placée à l’intérieur d’une boîte d’eau qui

a été préalablement équilibrée à la température de simulation désirée. Les molécules d’eau

se trouvant à moins d’une certaine distance (~2.4 à 2.8 Å) de tout atome de la protéine sont

Préparation

du système

Initiation et

équilibration Production Analyse de la

trajectoire

Dynamique moléculaire

59

éliminées du système. La couche d’eau entre la protéine et les frontières de la boîte de

simulation est d’au moins 10 Å [152]. Des ions sont ajoutés dans la boîte de simulation

afin de neutraliser le système et/ou reproduire la force ionique expérimentale. La neutralité

du système est nécessaire si l’énergie électrostatique est évaluée à l’aide de la méthode

« Particle Mesh Ewald (PME) »[160]. Avant de démarrer le calcul de DM, une

minimisation d’énergie peut être nécessaire. Généralement, la minimisation se fait en

conservant les coordonnées de la protéine fixées et ce faisant, les contacts entre la protéine

et le solvant sont optimisés. Il arrive parfois que la structure de la protéine renferme des

conformations non favorables nécessitant une minimisation d’énergie appliquée localement

(exemple, distance entre deux atomes trop courte).

3.2.2. Initiation et équilibration

La seconde étape consiste à l’initiation du calcul de DM et à l’équilibration. Pour ce faire,

une vitesse initiale est attribuée à chaque atome du système. Ces vitesses sont tirées d’une

distribution gaussienne en accord avec la température désirée. La température de départ est

habituellement plus basse que la température de simulation désirée. Le système est ensuite

réchauffé en augmentant progressivement les vitesses jusqu’à la température désirée.

Lorsque le réchauffement est terminé, la simulation entre en mode « équilibration ». Le

temps de simulation requis à l’équilibration varie d’un système à un autre. Pour vérifier si

le système est à l’équilibre, plusieurs observables physiques et biophysiques sont analysés.

Parmi ceux-ci, il y a l’énergie du système, le volume de la boîte de simulation, la

température et la pression. Pour ce qui est des observables biophysiques, la déviation de la

structure par rapport la structure de départ (« RMSD » ou « root mean squared deviation »)

ou encore la fluctuation de la surface accessible au solvant en fonction du temps peuvent

être suivies.

60

3.2.3. Production et analyse de la trajectoire

Lorsque l’équilibre est atteint, la simulation se poursuit, mais on dit alors que la simulation

entre en mode « production ». C’est cette partie de la trajectoire qui est utilisée pour les

analyses. De façon périodique, les coordonnées du système sont enregistrées pour ces

analyses, généralement une fois à chaque ps.

Le temps de simulation en mode production dépend des phénomènes que l’on veut étudier.

De longues simulations permettront d’étudier des phénomènes plus lents ou encore

observer des changements de conformation moins fréquents (meilleur échantillonnage). La

Figure 2.3 montre les échelles de temps des différents mouvements dans les protéines en

lien avec différentes méthodes utilisées. Dans l’optique de la présente thèse, les temps de

simulations total nécessaire pour étudier la diffusion des molécules de substrat gazeux à

l’intérieur de différentes enzymes varie typiquement entre 10 ns et 6 µs [85, 90, 92, 93, 96].

L’analyse d’une trajectoire donnée débute dès l’initiation du calcul de DM. Plusieurs

éléments doivent être analysés. Certains d’entre eux sont importants pour s’assurer que la

trajectoire en cours est stable. Entre autres, l’évolution dans le temps de l’énergie

potentielle, de la température, du volume et de la pression du système simulé. En ce qui a

trait à la protéine simulée, les observables d’intérêts sont nombreux et variés. Par exemple,

on peut s’intéresser à la dynamique de la chaîne principale, à la conformation dynamique

du site actif de la protéine ou encore à la dynamique essentielle. Pour TrHbN, l’étude de la

dynamique des tunnels présente un intérêt certain. Les principales méthodes utilisées pour

ce type d’étude sont traitées à la section suivante.

3.3. Étude de la dynamique des tunnels

Le rôle des tunnels dans les protéines est lié à la diffusion des substrats et des produits.

L’utilisation de la DM a révélé que la morphologie des tunnels est changeante en fonction

61

du temps et pour cette raison, l’étude de la dynamique des tunnels n’est pas triviale [115].

L’inspection visuelle des trajectoires en utilisant des outils de visualisation moléculaire tel

que PyMOL [27] peut permettre d’identifier une ou plusieurs route(s) de diffusion

potentielle(s). Par contre, pour obtenir une meilleure description de la dynamique de ces

tunnels, d’autres outils doivent être utilisés en complément. Il existe quelques outils bio-

informatiques récents et spécialement conçus à cette fin. Les outils les plus utilisés sont

CAVER [115] (MOLE [164]) et l’échantillonnage implicite de ligands [89]. Il existe

également le programme HOLE, dédié à l’étude des canaux ioniques dans les membranes,

mais celui-ci est plutôt mal adapté pour l’étude des tunnels dans les protéines [165]. Pour

étudier la dynamique des tunnels, ces outils bio-informatiques nécessitent d’abord une

trajectoire de DM.

3.3.1. CAVER

CAVER est un outil qui permet de détecter et de caractériser les routes de diffusion les plus

probables à l’intérieur d’une protéine [115]. Le programme MOLE [164] est une version

dérivée de CAVER mais c’est encore ce dernier qui continu à être préféré et développé.

CAVER utilise la théorie des graphes et plus précisément une variation de l’algorithme de

Dijkstra [166] afin de trouver le ou les meilleur(s) chemin(s) pour atteindre l’extérieur de la

protéine à partir d’une coordonnée interne définie par l’utilisateur. L’algorithme de

CAVER découpe d’abord la protéine selon une grille tridimensionnelle puis calcule en tout

point la distance avec l’atome de la protéine le plus près. Dans une seconde étape,

l’algorithme trouve le(s) tunnel(s) le(s) plus facile(s) permettant d’atteindre l’extérieur de la

protéine. Pour chaque tunnel détecté, CAVER décrit l’amplitude de l’ouverture le long du

tunnel. En analysant plusieurs étapes de DM avec CAVER, on obtient une meilleure

caractérisation du ou des tunnels. Il existe d’autres programmes dédiés au calcul du volume

de cavités (analyse volumétrique) tel que VOIDOO [167] mais CAVER présente deux

avantages majeurs par rapport à ceux-ci. Le premier consiste au fait que CAVER n’utilise

pas une sphère de rayon de sonde fixe ce qui permet de passer à travers les pincements le

long du tunnel. À l’opposé, les outils d’analyses volumétriques ne peuvent détecter qu’une

62

seule cavité isolée. Le second avantage est que CAVER, avant l’analyse même des tunnels,

délimite la surface de la protéine. Cette étape est cruciale pour permettre à CAVER de

déterminer où se termine un tunnel. Chez les outils d’analyses volumétriques, l’absence de

cette fonctionnalité est problématique lorsque la cavité est en continuum est l’extérieur de

la protéine. Le volume calculé est alors largement surestimé.

3.3.2. Échantillonnage implicite de ligand

L’échantillonnage implicite de ligand (ÉIL) est une méthode puissante qui est devenue

disponible au cours des travaux décrits dans cette thèse [89]. L’ÉIL permet de déterminer

les routes de diffusion de molécules gazeuses les plus probables à l`intérieur d’une

protéine. Pour y parvenir, le changement d’énergie libre associé au placement d’une

molécule mono- ou diatomique donnée (ex. O2, CO, Xe) est calculé en tout point à

l’intérieur de la protéine. Cette méthode s’appuie sur l’hypothèse que les molécules de gaz

interagissent faiblement avec la protéine et donc affectent peu la dynamique de la protéine.

Un grand nombre de jeux de coordonnées, tirées de la trajectoire de DM est nécessaire, soit

généralement > 5000 pour assurer une erreur minimale. Une fine grille tridimensionnelle

(résolution ~ 1 Å3) couvrant tout le système est d’abord définie. Dans chacun des cubes de

cette grille, un ligand donné est positionné sous diverses positions et orientations (si ligand

diatomique). Le potentiel de Lennard-Jones est calculé pour chacune de ces conformations

(équation 3-4). Le potentiel de force moyen (PFM), c’est-à-dire le changement d’énergie

libre associé au positionnement du ligand en un point quelconque du système, peut donc

être estimé. Le changement d’énergie libre est une quantité très pertinente, car celle-ci est

directement reliée à la probabilité de trouver un ligand à une position donnée. L’énergie

libre est calculée selon l’équation

63

( ) ∑∑ ( )

(équation 3-9)

où G(r) est le changement d’énergie libre associé au placement d’un ligand donné à la

position r, kB est la constante de Boltzmann, T est la température, N est le nombre d’étapes

de DM utilisées, C est le nombre de conformations du ligand testées (positions et rotations),

ΔEn,k(r) est l’énergie d’interaction entre le ligand et le système à la position r à une

conformation (k) et une étape de DM (n) données. Les calculs sont réalisés par l’outil

volutil disponible sur le site internet du groupe du Pr. Klaus Schulten

(http://www.ks.uiuc.edu/Development/MDTools/volutil). Après ces calculs, une carte

tridimensionnelle du PFM est obtenue. Cette carte peut être facilement interprétée avec

l’outil de modélisation moléculaire VMD [168].

Comme mentionné précédemment, l’énergie d’interaction calculée ne prend en compte que

le potentiel de Lennard-Jones. Ceci peut être problématique lorsque le ligand utilisé a un

dipôle comme le monoxyde de carbone ou l’oxyde nitrique. Malgré cette approximation,

les énergies libres de solvatation obtenue avec cette méthode sont très près des valeurs

expérimentales [89]. L’ÉIL présente une autre limitation, soit la surestimation des valeurs

d’énergie libre élevées [89]. Cette tendance se vérifie par l’équation

( ) [ [ ( ) ]

]

Équation 3-10

où ΔG-(r) est l’erreur inférieure sur le PFM calculé G(r), N est le nombre d’étapes de DM

utilisées, ΔEmin est une valeur d’énergie correspondant à l’énergie d’interaction entre le

ligand gazeux et son environnement dans les conditions les plus favorables. Cette dernière

valeur d’énergie est déterminée à partir d’une simulation d’un ligand explicite dans une

boîte d’eau. Lorsque la valeur du PFM est élevée par rapport à ΔEmin, l’erreur tend à être

http://www.ks.uiuc.edu/Development/MDTools/volutil

64

beaucoup plus importante. Cette erreur doit alors être contrebalancée par un

échantillonnage plus élevé sans quoi la valeur du PFM risque d’être surestimée.

3.3.3. L’échantillonnage amélioré de ligands

L’échantillonnage amélioré de ligand (« Ligand Enhanced Sampling ») (LES) est une

technique de simulation biaisée développée il y a près de 20 ans. Elle a été utilisée pour la

première fois par Elber et Karplus pour l’étude de la diffusion du CO à l’intérieur de la

myoglobine [88]. Il a été démontré très récemment que la diffusion complète du CO entre

la cavité distale et l’hème nécessite plusieurs dizaines de nanosecondes [90, 169]. Au début

des années 90, la puissance des ordinateurs ne permettait pas de produire de telles

trajectoires. La méthode LES utilise une astuce pour favoriser l’exploration du CO à

l’intérieur des protéines. Dans cette technique, le substrat simulé est libre et répliqué un

certain nombre de fois. Chacune des copies du substrat est invisible l’une par rapport à

l’autre. Les interactions non liées entre la protéine et le substrat sont quant à elles

multipliées par un facteur 1/N, où N est le nombre de copies du substrat. Ce faisant, les

molécules de substrats peuvent plus facilement traverser les barrières stériques et donc

explorer plus rapidement l’espace interne accessible à l’intérieur de la protéine. Le grand

nombre de copies augmente du même coup l’échantillonnage. Les routes de diffusion mises

en évidences peuvent ensuite être caractérisées plus en détail.

Contrairement à l’ÉIL, cette méthode permet difficilement de calculer précisément le PFM

pour la diffusion des ligands. Le PFM étant fonction de la probabilité de trouver un ligand

en tout point dans la protéine [89], il faut générer un grand nombre de simulations pour y

parvenir. Sans PFM, il est de plus ardu de prédire quelle sera la ou les route(s) de diffusion

privilégiée(s). Malgré que plusieurs molécules de substrats soient inclues dans le système

lors d’une simulation, les copies de substrats s’influencent indirectement par leur

interaction avec la protéine. Par conséquent, elles visiteront plus fréquemment les mêmes

endroits. De plus, dans ce type de simulation, l’exploration de la matrice protéique est

souvent contrainte artificiellement à une région limitée de l’espace [89]. En somme, la

65

méthode LES est efficace pour rapidement mettre en évidence plusieurs routes de diffusion

possible. Par contre, la carte tridimensionnelle de la diffusion des ligands à l’intérieur de la

matrice protéique ainsi que sa compréhension (PFM) peut ne pas être complète.

66

4.

Chapitre 4

Objectifs du projet de recherche

L’enzyme TrHbN protège la respiration aérobie de Mycobacterium bovis BCG contre

l’effet inhibiteur du •NO et catalyse efficacement la réaction NOD (k`NOD ≈ 745 µM-1

s-1

à

23°C) [22]. La vitesse de cette réaction s’approche même de celles limitées par la diffusion

des substrats [43, 44]. Dans ce contexte, le sujet de cette thèse vise à mieux comprendre les

relations existant entre la structure de TrHbN, sa dynamique et sa fonction. En particulier,

la dynamique du site actif et des tunnels ainsi que la diffusion interne •NO ont étudié en

détail.

4.1.1. Objectif général

L’objectif général de ce projet de doctorat est la caractérisation de la structure et de la

dynamique de TrHbN en solution sous ses formes oxygénée, deoxy et cyanomet en

présence et en absence de substrats libres à l’aide de différentes approches théoriques.

4.1.2. Objectifs spécifiques

1. Étude de la dynamique d’une molécule d’eau au site actif de deoxy-TrHbN et son

impact sur les cinétiques de liaison de substrat gazeux.

1.1. Étudier la persistance et l’espace visité par une molécule d’eau dans la cavité

distale de la forme deoxy de TrHbN et celle des mutants Gln58(E11)Val,

Tyr33(B10)Phe et le double mutant Gln58(E11)Val+Tyr33(B10)Phe.

1.2. Interpréter les résultats en relation avec les données expérimentales.

67

2. Investigation de la diffusion du •NO dans les tunnels de TrHbN.

2.1. Identifier le ou les site(s) d’entrée et/ou de sortie à la surface de TrHbN.

2.2. Étudier la diffusion du •NO dans la matrice de TrHbN.

2.3. Étudier les interactions entre le •NO libre et TrHbN et leurs impacts sur la structure

et la dynamique de TrHbN.

2.4. Étudier par dynamique moléculaire des mutants dont les tunnels ont été bloqués et

interpréter les résultats en relation avec des données de cinétiques enzymatiques.

3. Étude de TrHbN sous sa forme Fe3+

CN- en alliant conjointement la dynamique

moléculaire et la résonance magnétique nucléaire.

3.1. Étudier de la dynamique de la chaîne principale.

3.1.1. Comparer et interpréter les paramètres d’ordre généralisés S2 des liens N-H

du plan peptidique.

3.1.2. Comparer et interpréter les modèles de relaxation des vecteurs N-H du plan

peptidique

3.1.3. Comparer et analyser différents observables issus de trajectoires de DM en

lien avec des données d’échange d’amides du plan peptidique.

Cette thèse de doctorat présente les résultats obtenus en vue de la réalisation de ces

objectifs. Les résultats obtenus pour l’objectif 1 sont détaillés dans le chapitre 5. Les

travaux en lien avec les objectifs 2.1 et 2.2 sont détaillés dans le chapitre 6. Les chapitres 7

et 8 présentent des travaux en lien avec les objectifs 2.3 et 2.4 respectivement. Quant-à-lui,

le chapitre 9 décrit les travaux réalisés dans le cadre de l’objectif 3. Enfin, le chapitre 10

présente une discussion générale sur les travaux réalisés et propose des perspectives de

recerches.

68

5.

Chapitre 5

Ligand Binding to Hemoglobin N from Mycobacterium

tuberculosis is Strongly Modulated by the Interplay between

the Distal Heme Pocket Residues and Internal Water

5.1. Résumé

La survie de Mycobacterium tuberculosis requiert la détoxification du •NO produit par l’hôte. La

forme oxygénée de l’hémoglobine tronquée N de Mycobacterium tuberculosis catalyse

efficacement l’oxydation du •NO en nitrate (constante bimoléculaire de second ordre de k´NOD ≈

745 × 106 M

-1s

-1), soit 15 fois plus que la réaction catalysée par la myoglobine extrait du cœur du

cheval. Nous nous sommes intéressés à déterminer quels sont les aspects de la structure et/ou de

la dynamique de la protéine qui confèrent une telle réactivité. Une première étape consiste à

exposer les éléments contrôlant la liaison des ligands et substrats à l’hème. Nos travaux ont

soulevé des indices selon lesquels la barrière principale à la liaison des ligands à deoxy-TrHbN

consiste au déplacement d’une molécule d’eau présente dans la cavité distale de l’hème, laquelle

étant principalement stabilisée par la Y(B10) tout en demeurant non-coordonnée au fer. Comme

observé dans des mutants apolaires des résidus Tyr(B10)/Gln(E11) chez lesquels cette barrière

cinétique est moindre, la liaison du CO et de l’O2 est très rapide avec des vitesses avoisinant de 1

à 2 × 109 M

-1s

-1. De telles vitesses représentent presque certainement les vitesses de liaison à une

hemoprotéine les plus rapides connues et indiquent que l’atome de fer à l’intérieur de TrHbN est

hautement réactif. Des mesures cinétiques sur le produit photodissocié de la forme •NO de met-

TrHbN, où le •NO et l’eau peuvent être suivis directement, révèlent que la liaison de l’eau est

très rapide (1.49 × 108 s

-1) et est responsable de la faible fraction de recombinaison géminée chez

TrHbN. Des simulations de dynamique moléculaire, réalisée avec TrHbN et quelques mutants du

site distal, indiquent que dans l’absence de la molécule d’eau distale, l’accès du ligand au fer est

libre. Ces simulations montrent aussi que la molécule d’eau est stabilisée tout près du fer par

l’entremise de liaison hydrogènes avec les résidus Tyr(B10) et Gln(E11).

69

5.2. Abstract

The survival of Mycobacterium tuberculosis requires detoxification of host •NO. Oxygenate

Mycobacterium tuberculosis truncated hemoglobin N catalyzes the rapid oxidation of nitric

oxide to innocuous nitrate with a second-order rate constant (k´NOD ≈ 745 x 106 M

-1·s

-1), which is

~15-fold faster than the reaction of horse heart Mb. We ask what aspects of structure and/or

dynamics give rise to this enhanced reactivity. A first step is to expose what controls

ligand/substrate binding to the heme. We present evidence that the main barrier to ligand binding

to deoxy-TrHbN is the displacement of a distal cavity water molecule, which is mainly stabilized

by residue Tyr(B10) but not coordinated to the heme iron. As observed in the

Tyr(B10)/Gln(E11) apolar mutants, once this kinetic barrier is lowered, CO and O2 binding is

very rapid with rates approaching 1-2 x 109 M

-1·s

-1. These large values almost certainly represent

the upper limit for ligand binding to a heme protein and also indicate that the iron atom in

TrHbN is highly reactive. Kinetic measurements on the photoproduct of the •NO derivative of

met-TrHbN, where both the •NO and water can be directly followed, revealed that water

rebinding is quite fast (~ 1.49 x 108 s

-1) and is responsible for the low geminate yield in TrHbN.

Molecular dynamics simulations, performed with TrHbN and its distal mutants, indicated that in

the absence of a distal water molecule, ligand access to the heme iron is not hindered. They also

showed that a water molecule is stabilized next to the heme iron through hydrogen-bonding with

Tyr(B10) and Gln(E11).

5.3. Introduction

•NO plays an important role in host defense against microbial pathogens by inhibiting or

inactivating key enzymes such as the terminal respiratory oxidases (1-5) and the iron/sulfur

protein aconitase (6,7). •NO also combines at near diffusion-limited rate with superoxide

produced by respiring cells to form the highly oxidizing agent peroxynitrite (8,9). •NO-

metabolizing reactions are thus required to defend microbial pathogens against •NO poisoning.

70

In Mycobacterium tuberculosis the glbN gene encodes the truncated hemoglobin N (TrHbN)

(Fig. 5.1). Inactivation of glbN in Mycobacterium bovis BCG impairs the ability of stationary

phase cells to protect aerobic respiration from nitric oxide (•NO) inhibition, suggesting that

TrHbN may protect M. tuberculosis from •NO toxicity in vivo (10). This functional assessment

is supported by the observation that TrHbN catalyzes the rapid oxidation of •NO to nitrate

[TrHbN(Fe2+

-O2) + •NO → TrHbN(Fe3+

) + NO3-], with a second-order rate constant k´NOD ≈ 745

x 106 M

-1·s

-1 (23°C) (10). The nitric oxide dioxygenase (NOD) reaction catalyzed by TrHbN is at

least 15-fold faster (k´NOD ≈ 45 x 106 M

-1·s

-1 at 23°C) than the one recorded for horse heart

myoglobin (Mb) and is almost as efficient as the diffusion-controlled reaction of •NO with free

O2 . A critical issue in this context is what aspects of structure and/or dynamics give rise to this

enhanced reactivity. A first step is to expose what controls ligand/substrate binding to the heme.

Once the ligand/substrate accesses the distal heme pocket (DHP), the issue of reactivity focuses

on local factors such as iron reactivity and steric effects originating within the DHP. Inspection

of Mb and TrHbN structures shows that in Mb the imidazole ring of the proximal His is in an

eclipsed orientation with respect to the pyrrole nitrogen atoms. In contrast, that in TrHbN is in a

staggered geometry, suggesting reduced repulsive interactions between the imidazole ring and

the pyrrole nitrogen atoms and a stronger heme-iron bond (higher iron reactivity). This

assessment is supported by resonance Raman studies of deoxy-TrHbN and Mb also indicating a

stronger Fe-His bond in TrHbN (11). Based on the favorable proximal environment, one would

anticipate faster combination rates and higher geminate yields for TrHbN relative to Mb.

Surprisingly both proteins bind O2 with relatively similar rates and the geminate yields for CO

are both comparably low in the few percent range at ambient conditions (12,13). These

observations reveal that significant distal factors dictate the binding properties of TrHbN.

There are several categories of distal effects that can modulate ligand binding. Steric effects from

the side chains of distal residues can increase the barrier for binding through either static

positioning or relaxation subsequent to ligand dissociation and diffusion. In our earlier work, that

showed a dramatic increase in the geminate yield with increasing solvent viscosity, we

postulated that viscosity dependent relaxations of side chains were responsible for the large

changes in the geminate yield (13). In the present study, this hypothesis is reexamined along with

consideration for another potential distal contribution arising from water occupying the DHP.

71

Ligand binding to ferrous (Fe2+

) and ferric (Fe3+

) Mb requires the displacement of a water

molecule that is hydrogen-bonded to the distal His(E7) residue side chain (14-21). In ferric Mb,

the distal water molecule coordinates as a weak ligand, while in the ferrous derivative it occupies

a site in the DHP, blocking access to the heme iron but, at most, it only transiently interacts with

the ferrous heme iron (22). Kinetic data supports the concept that the distal water molecule

increases the enthalpic contribution to the kinetic barrier by sterically hindering ligand access to

the heme iron (14,17-19,23,24). Static and dynamic steric contributions to such barrier have been

previously shown for various DHP side chains including that of His(E7) in Mb (14,15,17-

19,23,25-29). The His(E7) stabilized DHP water molecule can be viewed as increasing the

effective size of the sterically active His(E7) side chain. In agreement, photolysis experiments on

Mb(Fe2+

-CO) and Mb(Fe3+

-NO) show that substitutions of His(E7) by different apolar residues

resulted in enhanced ligand rebinding rates which are quantitatively related to the lack of

occupancy of the distal water molecule (14-21,30).

In the present work, we examined the ligand binding properties of TrHbN bearing mutations at

residues Tyr(B10) and Gln(E11). Our results indicate that both the main barrier to ligand binding

to deoxy-TrHbN and the origin of the low geminate yield are due to the presence of Tyr(B10)

stabilized water within the DHP at a site that blocks access to the heme iron. Such proposal is

further supported by the observation that in the Tyr(B10)/Gln(E11) double mutants, where side

chain stabilization of the DHP water molecule is not possible, the combination rate becomes very

rapid with rates approaching those measured for diffusion-controlled reactions and the geminate

yield increases by almost two orders of magnitude.

5.4. Experimental procedures

5.4.1. Mutagenesis, expression and purification

Recombinant TrHbN and mutants were expressed and purified as previously described (31).

Flash-photolysis experiments – Laser flash-photolysis studies were carried out using the LKS.60

Spectrometer from Applied Photolysis (Leatherhead, U.K.) at 23°C. Photolysis was initiated by a

72

5 ns pulse of light at 532 nm provided by a Brillant B Nd:YAG laser (QUANTEL S.A., Fr.).

Absorbance changes were measured using the monochromator-filtered light from a 150 W xenon

arc lamp. Passing through the sample, the probe light beam was refocused on the slits (slits

widths at 1 mm) of a second monochromator. Changes in transmitted probe light intensity were

detected by a 1P28 PMT coupled with a HP 54830B DSO digital oscilloscope (Agilent

Technologies Inc., USA) and transferred on a RISC platform PC (Acorn, U.K.) for processing.

An average of at least ten kinetic traces from at least two separate experiments were averaged

and analyzed with the instrument manufacturer software (Applied Photolysis, U.K.) to obtain the

rate constants. Plots of the pseudo first-order rate constants and plots showing absorbance

changes following •NO photolysis were obtained using the KaleidaGraph software (Synergy

Software, USA).

Protein samples for the flash-photolysis experiments were used at concentrations ranging from

1.5 μM to 10 μM and buffered in anaerobic 50 mM potassium phosphate pH 7.5 containing

50 μM EDTA The ferric and deoxy protein samples were prepared in a glovebox as described

previously (31) and put into a gastight quartz cuvette with a 5 mm path length. To obtain the

desired complexes, the deoxy and ferric samples were equilibrated with different concentrations

of either O2, CO or •NO provided by a series 4000 gas mixing system from Environics (Tolland,

CT). Combination rates for CO and O2 were followed at wavelengths ascribed to maxima and

minima in either the TrHb(Fe2+

-CO) or the TrHb(Fe2+

-O2) minus the TrHb(Fe2+

) differencial

spectra. In order to study the extend of water regulation on ligand binding to Mb and TrHbN,

•NO recombination kinetics were followed over a broad timescale (ns - ms) and at specific

wavelengths corresponding to isobestic points between the (Fe3+

-H2O), (Fe3+

-NO) and (Fe3+

)

5-coordinate (5C) species [Fig. 5.2 and ref (21)]. Absorption spectra were recorded before and

after time course measurements to ensure the integrity of the samples.

5.4.2. Geminate and solvent phase recombination experiments

Geminate and solvent phase recombination measurements were carried out using 8 ns 532 nm

pulses at 1 Hz from a Nd:YAG laser (Minilite, Continuum, Santa Clara, CA) as a

73

photodissociation source and a greatly attenuated continuous wave 442 nm probe beam from a

He:Cd laser to monitor time dependent changes in absorption. Details of the apparatus, data

collection and data display can be found in a previous publication and citations therein

(13,26,32). The kinetic traces are displayed on a log-log plot of normalized absorbance

(proportional to the survival probability of the photoproduct) versus time.

Kinetic measurements were typically carried out on solution samples (~ 0.25 - 0.5 mM in heme)

contained in standard 1 mm stoppered cuvettes placed in a custom-built dry N2 purged variable

temperature cuvette holder (-15 to +65°C). The one sol-gel encapsulated sample was prepared as

a thin layer lining the bottom portion of a 10 mm diameter NMR tube as previously

described (33) but with the protocol modified (no added glycerol) in order to minimize the

increase in internal viscosity.

5.4.3. Molecular dynamics simulations

Simulations were performed using CHARMM (34) and the CHARMM22 all-atom potential

energy parameter set (35) with phi, psi cross term map correction (CMAP) (36) and modified

TIP3P waters (37). Electrostatic interactions were calculated via the Particle Mesh Ewald (PME)

method (38), using a sixth order spline interpolation for complementary function, with

κ = 0.34 Å-1

and a fast-Fourier grid density of ≈ 1 Å-1

. Cutoffs for the real space portion of the

PME calculation and the truncation of the Lennard-Jones interactions were 10 Å, with the latter

smoothed via a shifting function over the range of 8 Å to 10 Å. The SHAKE algorithm (39) was

used to constrain all covalent bonds involving hydrogen atoms. All simulations employed the

leapfrog algorithm and an integration step of 1 fs. Coordinates were saved every ps. Non-bond

and image lists were updated heuristically. All simulations were performed at constant pressure

and temperature (NPT ensemble) of 1 atm and 298 K, respectively. The mass of the thermal

piston was 20 000 kcal·mol-1

·ps2 and the mass of the pressure piston equaled 1000 amu. The net

translation and rotation of the systems were removed every 10 000 steps.

74

Systems setup – Coordinates for MD simulations were taken from the crystal structure of wild-

type oxy-TrHbN (PDB entry 1IDR). All ionizable residues were considered in their standard

protonation state at pH 7 and the histidines with the proton on ND1 position. Missing coordinates

from crystal structure were built using the internal coordinate definitions of CHARMM. For each

subunit, the carboxy terminal end was optimally positioned by performing 3 ns Langevin

dynamics with a 1 fs time step and a friction coefficient FBETA of 5 ps-1

while keeping

constrained all coordinates from the crystal structure. The converged structures were immersed

in a rhombic dodecahedron unit cell containing pre-equilibrated TIP3P water molecules (8330

molecules). Six sodium ions were added to neutralize the charge of the systems. Water molecules

within 2.8 Å of any protein atom were deleted yielding to about 24350 atoms for each system.

Prior to the initiation of MD simulations, the energy of the solvated systems was minimized with

two cycles of 500 steps of steepest descent followed by 500 steps of Adopted Basis Newton-

Raphson minimizations. During energy minimization, the protein coordinates were kept

constrained.

To increase sampling, two 20 ns trajectories were generated using the A and B crystal subunits in

absence of the coordinated dioxygen molecule. From these simulations, the last MD coordinates

were taken to produce two 20 ns trajectories of deoxy-TrHbN with and without a water molecule

in the DHP, for a total of four 20 ns trajectories. The water molecule was arbitrarily positioned in

the DHP in a cavity located between the heme iron and the B10 residue, and the initial position

was optimized with a short energy minimization, keeping all other coordinates constrained.

Because of their high structural similarities (31), mutant systems were built from equilibrated

wild-type coordinates. Three trajectories were produced for each deoxy form of both

Tyr(B10)Phe and Gln(E11)Val single mutants with a water molecule in the DHP, while five

trajectories were produced for the Tyr(B10)Phe/Gln(E11)Val double mutant. The number of

trajectories and their length were set in order to obtain a good sampling of the measured values.

For each mutant, one 10 ns trajectory without a water molecule in the DHP was performed.

Analysis of the DHP accessible volume – The accessible volumes located in the DHP were

studied using VOIDOO (40). The detected cavities were refined using a < 0.2 Å grid spacing and

a probe radius of 1.4 Å. One coordinate set every 10 ps was used for analysis.

75

5.5. Results and discussion

5.5.1. Kinetic data indicate that Tyr(B10) mainly contributes to the kinetic

barrier to ligand binding to TrHbN(Fe2+). – (i) O2 and CO binding to TrHbN

Combination of O2 and CO to most TrHbN mutants was too rapid to be measured by stopped-

flow spectrophotometry. As a consequence, these reactions were studied by laser flash-

photolysis. Since the previously published binding rate constants for O2 (kón) and CO (lón) of

TrHbN have been determined by stopped-flow spectrophotometry (12), we reexamined the

reactions using flash-photolysis. The measured reactions for O2 and CO corresponded to about

25 % and 63 %, respectively, of the expected changes in absorbance. The residuals of the fitted

curves generated by applying a single exponential mathematical term to the kinetic data were

nearly random, implying that under these conditions, O2 and CO combination processes are

monophasic (Fig. 5.3). As shown in Table 5.1, kón (55.8 x 106 M

-1·s

-1) value is higher than that

previously determined by stopped-flow spectrophotometry (kón = 25 x 106 M

-1·s

-1), indicating

that the starting deoxy forms are different.

5.5.2. O2 and CO binding to TrHbN mutants

With the exception of the double Tyr(B10)Phe/Gln(E11)Val mutant, all mutant proteins bound

O2 in a concentration-dependent manner requiring one exponential term to fit the pseudo-first-

order time courses. No reaction could be measured for the Tyr(B10)Phe/Gln(E11)Val mutant,

indicating either rapid geminate rebinding following O2 dissociation or failure to photodissociate

the bound O2. As shown in Table 5.1, kón values of Tyr(B10)Phe (540 x 106 M

-1·s

-1) and

Tyr(B10)Leu (621.2 x 106 M

-1·s

-1) mutants were ~ 10-fold higher than that of TrHbN indicating

that Tyr(B10) contributes significantly to the energy barrier to O2 binding. In contrast, kón values

for the oxygenation of Gln(E11)Ala (37.5 x 106 M

-1·s

-1) and Gln(E11)Val (32.6 x 10

6 M

-1·s

-1)

mutants were slightly lower than that of TrHbN (55.8 x 106 M

-1·s

-1). Substituting both Tyr(B10)

76

and Gln(E11) for Leu and Val respectively, creating the Tyr(B10)Leu/Gln(E11)Val mutant,

caused an additional increase of the kón rate to 1811.8 x 106 M

-1·s

-1.

All mutants bound CO in a concentration-dependent manner requiring a single exponential term

to fit the kinetic traces. As observed for O2, single substitutions at Tyr(B10) position resulted in a

~ 10-fold increase of the lón with values attaining 77.2 x 106 M

-1·s

-1 and 92.56 x 10

6 M

-1·s

-1 for

Tyr(B10)Phe and Tyr(B10)Leu, respectively (Table 5.1). In contrast, the Gln(E11)Ala and

Gln(E11)Val mutants showed only small changes in the lón values. The

Tyr(B10)Phe/Gln(E11)Val and Tyr(B10)Leu/Gln(E11)Val double mutants combined with CO

with similar rates (Table 5.1). These latter reactions are quite fast, approaching values for

diffusion-controlled reactions (41,42).

Table 5.1 shows that replacement of Tyr(B10) with either Phe or Leu results in a over an order of

magnitude increase of both kón and lón. That both ligands are similarly affected implies a direct

steric effect associated with the Tyr(B10). Replacement of Gln(E11) with either Ala or Val has

very little influence on the binding rates; however, for the double mutant combining the

Tyr(B10) and Gln(E11) replacements, there is a synergistic effect that substantially enhances

both kón and lón relative to the increase due to the Tyr(B10) substitutions alone. The question

remains as to what structural and/or dynamical processes are responsible for these side chain

specific effects on the combination rates.

As a first step, we examined the ns and slower recombination occurring subsequent to ligand

photodissociation using a 8 ns laser pulse. In many instances, geminate recombination which

occurs on the sub-microsecond time scale reflects the influence of the initial conformation prior

to substantial relaxation. By monitoring the geminate recombination on these faster time scales,

it is possible to establish whether the elements responsible for the very large differences in

combination rates are operational from the onset when the ligand is initially dissociated and

localized within the local environment near the heme binding site.

77

5.5.3. Geminate and solvent phase recombination

Figure 5.4 compares kinetic traces of the geminate and solvent phase recombination of CO to

TrHbN and several distal mutants displayed on a log-log plot. The rebinding to TrHbN (green

trace) consists of a single exponential phase. This phase, which slows with decreasing

concentrations of CO (not shown) is assigned to solvent phase recombination. The double mutant

Tyr(B10)Phe/Gln(E11)Val (red trace) shows two very fast phases. Only the second phase slows

in response to a decrease in CO concentration (data not shown) indicating that this kinetic phase

is a very fast solvent phase recombination. The even faster recombination phase is consistent

with the notion that it is a geminate recombination reaction based on both time scale and

insensitivity to the external CO concentration. The recombination trace for the oxy derivative of

this mutant showed two similar phases and geminate yields on the ns time scale (data not

shown). It was also concluded, based on the photolysis yield at 10 ns, that for the double

B10/E11 mutant as well as the wild-type protein, there is a faster subnanosecond geminate phase

for dioxygen that decreases the ns quantum yield relative to the CO derivatives. In contrast to the

CO derivative of the double B10/E11 mutant, which displays an exceptionally fast ns geminate

process with very large amplitude (> 0.8) that is among the largest for any CO derivative of an

Hb or Mb under ambient low viscosity conditions, there is almost no discernable geminate yield

under these conditions for TrHbN. The Tyr(B10)Leu/Gln(E11)Val double mutant exhibits very

similar enhanced kinetics to those from the Tyr(B10)Phe/Gln(E11)Val double mutant, both

under low and high viscosity conditions (data not shown) The two single mutants Tyr(B10)Phe

(black trace) and Gln(E11)Val (blue trace) manifest a measurable geminate process but with a

geminate yield in the range of 0.2.

The geminate recombination data show that the factors that are responsible for the large

differences in the combination rates and the solvent phase kinetics are operative at early times

subsequent to photodissociation. We now consider possible factors contributing to these

differences. The enhancement of the binding rates and the solvent phase recombination in going

from Tyr(B10) to Phe(B10) might be the result of a decrease in the effect of the tyrosine side

chain relaxing to a position that blocks access to the heme iron. If this relaxation was fast

enough, it could account for the very low geminate yield for TrHbN. The further enhancement in

78

combination rates seen for the double B10/E11 mutant could be attributable to a further

reduction in steric factors due a change in the positioning of the Phe(B10) side chain due to the

change in the E11 side chain.

There are several observations that raise questions about the validity of this side chain-based

steric explanation. The B10 side chain explanation can only account for the observed kinetics if

one also invokes modulation of the positioning of the side chain by the E11 side chain.

Arguments against that scenario come from the observation that the replacement of Gln(E11)

with valine or alanine has minimal effect on the on rates and on the geminate yield (data not

shown for Gln(E11)Ala, which is essentially identical to that of the Gln(E11)Val mutant).

Furthermore, the binding rates for Tyr(B10)Phe and Tyr(B10)Leu are very similar. Geminate

recombination studies comparing these two mutants both at low and high viscosity show very

little difference (data not shown) suggesting that if the B10 side chain was contributing through a

pure steric effect (due to the side chain alone), the leucine and phenylalanine side chains would

have behaved similarly with respect to this proposed steric interaction. A similar steric effect by

these two residues seems implausible given the difference in flexibility and volume of the two

side chains. Finally, both the Tyr(B10)Phe and Tyr(B10)Leu double mutants with Gln(E11)Val

show similar recombination kinetics. If the further substantial enhancement in combination rates,

solvent phase recombination rates, geminate yield and rates of geminate recombination were due

to the further reduction in a steric effect due to a change in the B10-E11 interaction, it would

seem very implausible that the two different B10 side chains would behave so similarly. While

these arguments are not definitive, they certainly weaken any explanation based solely on steric

effects arising solely from the side chains of the B10 and E11 residues. Given these points

together with the absence of large ligand-binding induced conformational changes involving the

proximal heme environment associated with TrHbN (11), we consider yet additional factors that

can contribute to the control of ligand binding kinetics.

Water in the DHP of Mb and Hb is known to significantly contribute to the kinetic barrier for

ligand binding (14-21). In these two cases, water occupies the DHP of the deoxy derivative and

only populates the DHP of the liganded species subsequent to ligand dissociation and the onset

of conformational fluctuations that open the so-called distal His(E7) gate. Thus, there is a delay

between the moment of the ligand dissociation and the reentry of water back into the DHP. It has

79

been claimed that this delayed water reentry process is essential for ensuring a high probability

of escape by the dissociated ligand once the ligand has accessed the Xe cavities of Mb and Hb.

Delayed occupancy of the DHP by water subsequent to ligand dissociation is also a significant

factor contributing to the geminate yield for the slower geminate phase arising from the

recombination after the ligand has access to the Xe cavities. If the dynamics of water controlling

occupancy of the DHP of TrHbN is a factor contributing to the observed kinetic patterns, it

would suggest that water participates very early in the recombination process and require that the

B10 and E11 side chains modulate its dynamics and/or occupancy factors. Thus, in contrast to

Mb, where water is observed to enter the DHP of Mb only after 50 to 100 ns, in the case of

TrHbN, the water would have to be present within a few ns if it is responsible for the low

geminate yield in wild-type TrHbN. The high yield and fast rates for the double B10/E11 mutant

would be attributable to the very low occupancy of water within the DHP. To test this

hypothesis, we have conducted both kinetic measurements on the photoproduct of the •NO

derivative of met-TrHbN where both the •NO and water can be directly followed and MD

simulations to establish the behavior of water within the DHP of TrHbN as a function of B10 and

E11 side chain substitutions.

Water controls ligand binding to ferric TrHbN – Combined mutagenesis and spectroscopic

studies indicated that Tyr(B10) and Gln(E11) residues stabilize the coordinated water molecule

in ferric TrHbN at 23°C and pH 7.5 (31). Accordingly, the optical spectra of ferric double

mutants bearing apolar residues at Tyr(B10) and Gln(E11) positions were found typical of ferric

heme proteins with no water coordinated to the iron atom (31).

We used laser-flash photolysis of the TrHbN(Fe3+

-NO) complex to study the kinetics of water

entry and binding to the heme iron at 23ºC and pH 7.5. Photodissociation of the horse heart

Mb(Fe3+

-NO) complex leaves the heme distal site in a ferric dehydrated state 5C Mb(Fe3+

) (21).

After •NO photolysis and escape a water molecule enters the DHP and binds to the heme iron

forming the aquomet Mb state [Mb(Fe3+

-H2O)]. At longer times, •NO displaces the bound water

molecule to reestablish the equilibrium Mb(Fe3+

-NO) complex.

80

For monitoring H2O kinetics in TrHbN, the experimental wavelength is the isobestic point

between TrHbN(Fe3+

-NO) and TrHbN Tyr(B10)Phe/Gln(E11)Val(Fe3+

) mutant, which is located

near the Soret band of native ferric TrHbN at 406 nm (Fig. 5.2). Figure 5.5A shows the changes

in absorbance at 408.5 nm. We interpret the experimental results in terms of three optical states:

TrHbN(Fe3+

-NO), TrHbN(Fe3+

-H2O) and TrHbN(Fe3+

). Accordingly, following •NO-photolysis

most of the •NO directly rebinds without leaving TrHbN. After the non-geminate fraction of •NO

escapes to solvent, forming a short-lived 5C ferric dehydrated state, a water molecule rebinds

very rapidly (1.49 x 108 s

-1), forming the TrHbN(Fe

3+-H2O) state. Finally, under •NO saturating

conditions (∼ 1.8 mM), the bound H2O is rapidly displaced (2.0 x 105 s

-1) leading to the decrease

in absorbance seen in Figure 5.5A. As expected, increasing the •NO concentration shortens the

duration of the TrHbN(Fe3+

-H2O) complex and has no effect on the rate of formation of

TrHbN(Fe3+

-H2O) (Fig. 5.5B). As shown in Fig. 5.5A and ref (21), water binding to 5C ferric

horse heart Mb following •NO-photolysis is significantly slower (5.7 x 106 s

-1), suggesting a

lower barrier for migration of water molecule in TrHbN.

Figure 5.5C shows the kinetic trace obtained when the reaction is monitored at 421 nm. This

wavelength corresponds to the maximum absorbance of the Soret band of the TrHbN(Fe3+

-NO)

species (Fig. 5.2). At this wavelength, we shall monitor the reaction: TrHbN(Fe3+

-NO) →

TrHbN(Fe3+

) → TrHbN(Fe3+

-H2O) → TrHbN(Fe3+

-NO). Initially a decrease in absorbance is

observed (2.1 x 108 s

-1), corresponding to the formation of 5C ferric dehydrated state from

TrHbN(Fe3+

-NO) followed by a small increase in absorbance (1.1 x 108 s

-1) associated to

[TrHbN(Fe3+

) → TrHbN(Fe3+

-H2O] and finally by a further increase in absorbance (2.3 x 105 s

-1)

corresponding to •NO replacing the bound water molecule.

Thus water appears to constitute the main barrier to ligand rebinding to TrHbN. Unlike Mb and

human HbA, where water does not appear to impact geminate recombination due to delayed

reentry of water into the DHP subsequent to ligand dissociation, in the case of TrHbN, the water

occupancy occurs on the time scale of the geminate recombination. This observation indicates

that water has access to the reactive site over the heme iron on a time scale that is much faster

than for Mb. This acceleration is consistent either with water being stabilized within the protein

at a site near the heme iron or with water being able to enter the DHP from the solvent on a ns

time scale. Although the state of hydration of the DHP of deoxy-TrHbN is not known, the

81

present flash-photolysis experiments with TrHbN(Fe3+

-NO) strongly suggest that a non-

coordinated water molecule, stabilized by Tyr(B10) and Gln(E11), may be close to the heme iron

in deoxy-TrHbN. To investigate the fate of a water molecule in the DHP of deoxy-TrHbN and its

distal mutants and to gain further insights into the role of Tyr(B10) and Gln(E11), MD

simulations were performed.

5.5.4. Molecular dynamics simulations suggest that water may constitute the

main kinetic barrier to ligand binding to TrHbN(Fe2+

)

All simulations showed stable trajectories with protein backbone r.m.s.d around 1 Å. The

positioning of the different DHP residues B10, E11, CD1 and the free DHP water molecule was

studied. Table 5.2 shows the average minimum interatomic distances between non- hydrogen

atoms and the heme-Fe atom extracted from the different trajectories produced. To measure

access to the heme-Fe atom (accessible volume), we used a probe of 1.4 Ǻ radius, which

approximates to the radius of a water molecule. The results are presented in Table 5.3 and are

expressed as the fraction of MD snapshots showing an accessible volume over the iron.

Wild-type TrHbN(Fe2+

) trajectories - Two 20 ns trajectories were produced for the wild-type

protein, and in both cases the water molecule was stabilized by strong hydrogen-bonds involving

both the Gln(E11) and Tyr(B10) residues. As a result, the water molecule occupied a main

position close to the iron atom, at a mean H2O-Fe distance of 3.5 Å. On some rare occasions

(0.7 % of MD frames) the water molecule left this main position to get closer to the Gln(E11)

side chain, creating an accessible volume over the iron atom (Table 5.3). In the absence of a

water molecule, the Tyr(B10) hydroxyl group was hydrogen-bonded to the OE1 atom of

Gln(E11). This H-bond pulled the Tyr(B10) side chain further away from the heme-Fe atom at a

mean minimum distance of 5.7 Å. This configuration prevailed in 89.3 % of the time. In this

configuration, the B10 residue does not hamper ligand coordination and the closest residue from

the iron atom is Phe(CD1) at 4.0 Å. As a consequence, an accessible volume over the heme iron

was found 10 times more often (7.1 % of the MD frames) than in TrHbN hydrated trajectories.

82

Tyr(B10)Phe(Fe2+

) mutant trajectories – In the Tyr(B10)Phe mutant, the Gln(E11) residue

pulled the water molecule away from the heme center, at a mean distance of 4.9 Å. The water

molecule no longer occupied a well defined position, being constantly in motion around the

Gln(E11) side chain. Also, the Phe(B10) residue showed increased flexibility and explored two

χ1 dihedral domains (χ1 in minus and trans), compared to only one for TrHbN. Consequently,

accessible volume over the heme-Fe atom was observed more frequently (~ 8-fold) than in

TrHbN (Table 5.3). The average volume was also larger than in TrHbN by ~ 25 Å3. Consistent

with MD simulations, kinetic data showed a similar increase in both O2 and CO combination

rates.

In the absence of a water molecule, the access to the heme-Fe atom was slightly reduced with

respect to TrHbN (Table 5.3), predicting lower or similar kón and lón for the Tyr(B10)Phe

mutant. This is due to the Gln(E11) side chain which moves closer to the heme iron ( by 0.9 Å),

as also observed in the crystal structure of the cyanomet derivative (31). MD data for the

Tyr(B10)Phe mutant are thus consistent with kinetic data if a water molecule is present in the

DHP.

Gln(E11)Val(Fe2+

) mutant trajectories – In the Gln(E11)Val mutant, the DHP water molecule

occupied a position similar to that seen in TrHbN (Table 5.2) and was stabilized by a strong

H-bond to the Tyr(B10) hydroxyl group. Access to the heme iron was slightly decreased with

respect to TrHbN, accounting for only 0.6 % of the MD frames analyzed. This is due to the

absence of electrostatic attraction by the E11 residue favoring water location near the Tyr(B10)

hydroxyl group. Consistent with these data, kón and lón for the Gln(E11)Val mutant were found

similar to those of TrHbN (Table 5.1).

In contrast, in the absence of a DHP water molecule, cavity formation over the heme-Fe atom in

the Gln(E11)Val mutant increased ∼ 3-fold compared to TrHbN. In this case the Val(E11) side

chain (5.8 Å) is unable to get as close to the heme-Fe atom as the Gln(E11) residue in the

Tyr(B10)Phe mutant (4.4 Å). In TrHbN, Tyr(B10) maintains Gln(E11) side chain at a greater

distance from the heme-Fe atom (4.9 Å) through H-bonding. As a consequence, the cavities

detected over the iron atom in the Gln(E11)Val mutant were larger by about 10 Å3 than those

83

formed in TrHbN (Table 5.3). Thus in contrast to kinetic data, MD simulations in absence of a

distal water molecule would predict an increase in O2 and CO combination rates.

Tyr(B10)Phe/Gln(E11)Val(Fe2+

) mutant trajectories – The B10/E11 double apolar

substitution had a dramatic effect on the DHP water molecule stabilization. In all simulations, the

water molecule rapidly escaped the protein matrix through the short tunnel. These results

indicate that a water molecule is unlikely to reside within the DHP of the B10/E11 apolar

mutants.

The analysis of the accessible volume revealed that apolar substitutions of the B10/E11 pair

increased the accessible volume, by more than 15 Å3, compared to TrHbN (Table 5.3).

Additionally, in the absence of a water molecule, over 54.7 % of the MD data showed an

accessible volume over the heme-Fe atom (Table 5.3). Overall, the MD results are in accord with

the measured k´on and l´on, which indicates that iron coordination of small gaseous substrates, in

this case should only be limited by their diffusion from the solvent to the tunnel and then to the

active site.

5.6. Conclusions

The present kinetic data and MD simulations indicate that the main barrier to ligand binding

from solvent and geminate phase to deoxy-TrHbN is the displacement of a non-coordinated

distal site water molecule, which is mainly stabilized by the Tyr(B10) residue. As observed for

TrHbN Tyr(B10)/Gln(E11) double mutants, once this kinetic barrier is eliminated, geminate

yield is dramatically increased and ligand binding is very rapid with rates approaching those

measured for diffusion-controlled reactions. Such proposal is further supported by the

observation that the rates measured for •NO binding to ferric heme-iron increases dramatically in

the Tyr(B10)Leu/Gln(E11)Val double mutant (1585 x 106 M

-1·s

-1) compare to wild-type TrHbN

(114.2 x 106 M

-1·s

-1), being as fast as CO and O2 binding to the deoxy-form of the double mutant

(not shown). These large combination rates almost certainly represent the upper limit for ligand

binding to a heme protein (44,45) and also indicate that the heme iron in TrHbN is highly

reactive. Such rapid access to the active site is attributed to the hydrophobic nature of the

tunnels, which may favor rapid docking and partitioning of the apolar gas into the polar distal

84

heme cavity. In turn, the rapid diffusion of apolar ligand to the active site may be responsible for

the efficient NOD reaction catalyzed by TrHbN (745 x 106 M

-1·s

-1). In addition to ligand

binding, water molecules in the DHP can participate actively in other important processes

including proton transfer reactions, catalysis, folding and redox processes. Recent quantum

mechanics/molecular mechanics and MD simulations with ferric TrHbN suggest that formation

of the Fe3+

-ONO2- complex triggers rapid hydration (few ns) of the distal heme cavity, which

causes weakening of the Fe3+

-O bond and rapid egress of the nitrate ion from the active site (43).

Thus water in the DHP facilitates the rapid release of NO3-, which is necessary to guarantee an

efficient NO detoxification and enhance survival of the microorganism under stress conditions.

Photolysis experiments with TrHbN(Fe3+

-NO) indicate that water rebinds to the distal heme site

at a rate of ∼ 1.49 x 108 s

-1. Similar experiments with Mb(Fe

3+-NO) and Mb(Fe

2+-CO) estimated

that water enters into the distal heme pocket at a rate of 5.7 x 106 s

-1 and 9 x 10

6 s

-1, respectively.

The large difference in the rates of water rebinding emphasizes a lower barrier for water in

TrHbN, which can be attributed in part to the electrostatic interactions of water with the distal

residues Tyr(B10) and Gln(E11). The difference in binding rates of water between Mb and

TrHbN may also be attributed to a much faster access to the DHP from the solvent due to the

absence of a distal gate in TrHbN (15,19,44,45). The results of the geminate recombination

studies are consistent with the water either being near the heme from the start or accessing the

DHP on an unprecedently fast time scale. Whatever the situation, these results point to an

important role for water in control of ligand reactivity in TrHbN.

5.7. Footnotes

We are grateful to Dr. Beatrice A. Wittenberg and Dr. Jonathan B. Wittenberg from the Albert

Einstein College of Medicine (NY, USA) for insightful discussions. This work was supported by

the National Sciences and Engineering Research Council (NSERC) grant 46306-01 (2005-2010),

the NIH grant 1-R01-AI052258 (2004-2007) (through Dr. Joel M. Friedman) and the Fonds

Québécois de la Recherche sur la Nature et les Technologies (FQRNT) grant 104897 to Dr.

Michel Guertin. Mario Milani is recipient of a post-doctoral fellowship supported through the

NIH grant 1-R01-AI052258 (2004-2007). Dr. Martino Bolognesi is grateful to CIMAINA

85

(Milano - Italy). Part of this study was supported by the Italian Ministry for University and

Scientific Research FIRB Project “Biologia Strutturale” to Dr. Martino Bolognesi, Contract

RBLA03B3KC. Patrick Lagüe is supported by the Canadian Foundation for Innovation (CFI)

grant 12428 and the Fonds Québécois de la Recherche sur la Nature et les Technologies

(FQRNT) grant 104897.

The abbreviations used are: BCG, bacillus Calmette-Guérin; Hb, hemoglobin; Mb, myoglobin;

TrHb, truncated hemoglobin; TrHbN, Mycobacterium tuberculosis truncated hemoglobin N;

MD, molecular dynamics; DHP, distal heme pocket; 5C, 5 coordinated; •NO, nitric oxide.

86

5.8. References

1. Brunori, M. (2001) Trends in Biochemical Sciences 26(1), 21-23

2. Brunori, M., Giuffre, A., Sarti, P., Stubauer, G., and Wilson, M. T. (1999) Cell Mol Life Sci

56(7-8), 549-557

3. Stevanin, T. M., Ioannidis, N., Mills, C. E., Kim, S. O., Hughes, M. N., and Poole, R. K.

(2000) J Biol Chem 275(46), 35868-35875

4. Cleeter, M. W., Cooper, J. M., Darley-Usmar, V. M., Moncada, S., and Schapira, A. H.

(1994) FEBS Lett 345(1), 50-54

5. Brown, G. C., and Cooper, C. E. (1994) FEBS Lett 356(2-3), 295-298

6. Gardner, P. R., Costantino, G., and Salzman, A. L. (1998) J. Biol. Chem. 273(41), 26528-

26533

7. Gardner, P. R., Costantino, G., Szabo, C., and Salzman, A. L. (1997) J. Biol. Chem. 272(40),

25071-25076

8. Kissner, R., Nauser, T., Bugnon, P., Lye, P. G., and Koppenol, W. H. (1997) Chem. Res.

Toxicol. 10(11), 1285-1292

9. Pfeiffer, S., Gorren, A. C. F., Schmidt, K., Werner, E., Hansert, B., Bohle, D. S., and Mayer,

B. (1997) J. Biol. Chem. 272(6), 3465-3470

10. Ouellet, H., Ouellet, Y., Richard, C., Labarre, M., Wittenberg, B., Wittenberg, J., and

Guertin, M. (2002) Proc Natl Acad Sci U S A 99(9), 5902-5907

11. Samuni, U., Ouellet, Y., Guertin, M., Friedman, J. M., and Yeh, S. R. (2004) J A Chem Soc

126(9), 2682-2683

12. Couture, M., Yeh, S. R., Wittenberg, B. A., Wittenberg, J. B., Ouellet, Y., Rousseau, D. L.,

and Guertin, M. (1999) Proc Natl Acad Sci U S A 96(20), 11223-11228

13. Dantsker, D., Samuni, U., Ouellet, Y., Wittenberg, B. A., Wittenberg, J. B., Milani, M.,

Bolognesi, M., Guertin, M., and Friedman, J. M. (2004) J Biol Chem 279(37), 38844-38853

14. Springer, B. A., Sligar, S. G., Olson, J. S., and Phillips, G. N., Jr. (1994) Chem Rev 94(3),

699-714

15. Vojtechovsky, J., Chu, K., Berendzen, J., Sweet, R. M., and Schlichting, I. (1999) Biophys J

77(4), 2153-2174

87

16. Rohlfs, R. J., Mathews, A. J., Carver, T. E., Olson, J. S., Springer, B. A., Egeberg, K. D., and

Sligar, S. G. (1990) J Biol Chem 265(6), 3168-3176

17. Olson, J. S., and Phillips, G. N., Jr. (1996) J Biol Chem 271(30), 17593-17596

18. Olson, J. S., and Phillips, G. N., Jr. (1997) J Biol Inorg Chem 2(4), 544-552

19. Scott, E. E., Gibson, Q. H., and Olson, J. S. (2001) J Biol Chem 276(7), 5177-5188

20. Goldbeck, R. A., Bhaskaran, S., Ortega, C., Mendoza, J. L., Olson, J. S., Soman, J., Kliger,

D. S., and Esquerra, R. M. (2006) Proc Natl Acad Sci U S A 103(5), 1254-1259

21. Cao, W., Christian, J. F., Champion, P. M., Rosca, F., and Sage, J. T. (2001) Biochemistry

40(19), 5728-5737

22. Christian, J. F., Unno, M., Sage, J. T., Champion, P. M., Chien, E., and Sligar, S. G. (1997)

Biochemistry 36(37), 11198-11204

23. Quillin, M. L., Li, T., Olson, J. S., Phillips, G. N., Jr., Dou, Y., Ikeda-Saito, M., Regan, R.,

Carlson, M., Gibson, Q. H., Li, H., and et al. (1995) J Mol Biol 245(4), 416-436

24. Quillin, M. L., Arduini, R. M., Olson, J. S., and Phillips, G. N., Jr. (1993) J Mol Biol 234(1),

140-155

25. Dantsker, D., Samuni, U., Friedman, J. M., and Agmon, N. (2005) Biochim Biophy Acta

1749(2), 234-251

26. Dantsker, D., Roche, C., Samuni, U., Blouin, G., Olson, J. S., and Friedman, J. M. (2005) J

Biol Chem 280(46), 38740-38755

27. Samuni, U., Dantsker, D., Roche, C. J., and Friedman, J. M. (2007) Gene 398(1-2), 234-248

28. Draghi, F., Miele, A. E., Travaglini-Allocatelli, C., Vallone, B., Brunori, M., Gibson, Q. H.,

and Olson, J. S. (2002) J Biol Chem 277(9), 7509-7519

29. Samuni, U., Roche, C. J., Dantsker, D., and Friedman, J. M. (2007) J Am Chem Soc 129(42),

12756-12764

30. Eich, R. F., Li, T., Lemon, D. D., Doherty, D. H., Curry, S. R., Aitken, J. F., Mathews, A. J.,

Johnson, K. A., Smith, R. D., Phillips, G. N., Jr., and Olson, J. S. (1996) Biochemistry

35(22), 6976-6983

31. Ouellet, Y., Milani, M., Couture, M., Bolognesi, M., and Guertin, M. (2006) Biochemistry

45(29), 8770-8781

88

32. Samuni, U., Roche, C. J., Dantsker, D., Juszczak, L. J., and Friedman, J. M. (2006)

Biochemistry 45(9), 2820-2835

33. Samuni, U., Dantsker, D., Khan, I., Friedman, A. J., Peterson, E., and Friedman, J. M. (2002)

J Biol Chem 277(28), 25783-25790

34. Brooks, B. R., Bruccoleri, R. E., Olafson, B. D., States, D. J., Swaminathan, S., and Karplus,

M. (1983) J Comput Chem 4(2), 187-217

35. MacKerell, A. D., Bashford, D., Bellott, M., Dunbrack, R. L., Evanseck, J. D., Field, M. J.,

Fischer, S., Gao, J., Guo, H., Ha, S., Joseph-McCarthy, D., Kuchnir, L., Kuczera, K., Lau, F.

T. K., Mattos, C., Michnick, S., Ngo, T., Nguyen, D. T., Prodhom, B., Reiher, W. E., Roux,

B., Schlenkrich, M., Smith, J. C., Stote, R., Straub, J., Watanabe, M., Wiorkiewicz-Kuczera,

J., Yin, D., and Karplus, M. (1998) J. Phys. Chem. B 102(18), 3586-3616

36. Mackerell, A. D., Jr., Feig, M., and Brooks, C. L., 3rd. (2004) J Comput Chem 25(11), 1400-

1415

37. Price, D. J., and Brooks III, C. L. (2004) J Chem Phys 121(20), 10096-10103

38. Feller, S. E., Pastor, R. W., Rojnuckarin, A., Bogusz, S., and Brooks, B. R. (1996) J. Phys.

Chem. 100(42), 17011-17020

39. Ryckaert, J.-P., Ciccotti, G., and Berendsen, H. J. C. (1977) Journal of Computational

Physics 23(3), 327-341

40. Kleywegt, G. J., and Jones, T. A. (1994) Acta Crystallographica Section D 50(2), 178- 185

41. Huie, R. E., and Padmaja, S. (1993) Free Radic Res Commun 18(4), 195-199

42. Blough, N. V., and Zafiriou, O. C. (1985) Inorg. Chem. 24(22), 3502-3504

43. Marti, M. A., Bidon-Chanal, A., Crespo, A., Yeh, S. R., Guallar, V., Luque, F. J., and Estrin,

D. A. (2008) J Am Chem Soc 130(5), 1688-1693

44. Milani, M., Pesce, A., Ouellet, Y., Ascenzi, P., Guertin, M., and Bolognesi, M. (2001) Embo

J 20(15), 3902-3909

45. Bolognesi, M., Cannillo, E., Ascenzi, P., Giacometti, G. M., Merli, A., and Brunori, M.

(1982) J Mol Biol 158(2), 305-315

89

Table 5.1 Kinetics constants for the reactions of TrHbN and its mutants with O2 and CO.

Protein k´on (O2)

x 106 M-1·s-1

l´on (CO)

x 106 M-1·s-1

TrHbN 25a

55.8 ± 4.1 (25 %)b

6.75a

7.80 ± 0.15 (63 %)

Tyr(B10)Phe 540.0 ± 14.3 (15 %) 77.2 ± 1.3 (74 %)

Tyr(B10)Leu 621.2 ± 7.0 (24 %) 92.56 ± 0.75 (83 %)

Gln(E11)Val 32.6 ± 1.0 (26 %) 6.81 ± 0.26 (56 %)

Gln(E11)Ala 37.5 ± 4.0 (29 %) 9.34 ± 0.27 (69 %)

Tyr(B10)Phe / Gln(E11)Val * 1119.2 ± 13.4 (33 %)

Tyr(B10)Leu / Gln(E11)Val 1811.8 ± 51.3 (8 %) 1148.4 ± 50.5 (15 %)

a Rate determined by stopped-flow experiment (12).

b Percentage of the expected amplitude measured for the reaction.

* Binding of O2 may occur in the dead time of the apparatus.

90

Table 5.2 Average minimum interatomic distances between non-hydrogen atoms and the heme irona.

Proteinb iron–residue distance (Å)

Water B10 E11 F(CD1)

TrHbN-d - 5.7 4.9 4.0

Tyr(B10)Phe-d - 6.7 4.4 4.4

Gln(E11)Val-d - 5.2 5.8 4.2

Tyr(B10)Phe/Gln(E11)Val-d - 6.6 5.5 4.8

TrHbN-w 3.5 5.4 5.3 4.9

Tyr(B10)Phe-w 4.9 7.2 4.4 4.7

Gln(E11)Val-w 3.8 5.2 5.5 4.9

Tyr(B10)Phe/Gln(E11)Val-w c - - - -

a Other distal heme pocket residues (Leu(E7), Phe(B9) and Val(G8)) showed longer distance from the

iron ranging from 5.5Å to 7.5Å.

b Protein with (w) and without (d) a water molecule in the distal heme pocket.

c The water molecule escaped too fast the distal heme pocket to get a proper measurement.

91

Table 5.3 . Cavity formation frequency and volume over the iron atom.

Proteina Frequency

(%)

Volume

(Å3)

TrHbN-d 7.1 46.8

Tyr(B10)Phe-d 5.0 62.2

Gln(E11)Val-d 19.5 58.3

Tyr(B10)Phe/Gln(E11)Val-d 90.7 81.8

TrHbN-w 0.7 50.9

Tyr(B10)Phe-w 5.4 77.5

Gln(E11)Val-w 0.6 74.0

Tyr(B10)Phe/Gln(E11)Val-w 43.5 69.6

a Protein with (w) and without (d) a water molecule in the distal heme pocket.

92

Figure 5.1 View of the distal heme pocket and the tunnels of cyanomet-TrHbN chain B under

xenon pressure (PDB 1S56). Besides the protein backbone (blue ribbon, with labelled α-helices),

the figure shows hydrogen-bonding (dashed red lines) between the distal residues Tyr(B10) and

Gln(E11) and the heme-bound cyanide. The path of the two tunnels is shown in orange. The

short tunnel (~ 8 Å) connects the heme distal site to the outer solvent space at a location

comprised between the central region of the G and H helices (left in the figure). The long tunnel

(~ 20 Å) extends from the heme distal cavity to a solvent access site located between the inter-

helical loops AB and GH (upper part of the figure); note the gating role of Phe(E15) on the long

tunnel. The arrows point to the tunnel entrance sites facing the solvent. The figure was produced

using the PyMOL software (Delano Scientific, USA)

93

Figure 5.2 Equilibrium absorption spectra of TrHbN(Fe3+

-H2O) (solid line), TrHbN(Fe3+

-NO)

(dashed line) and TrHbN Tyr(B10)Leu/Gln(E11)Val(Fe3+

) mutant (dashed and dotted line) at pH

7.5. The protein concentrations were 10 μM. The Tyr(B10)Leu/Gln(E11)Val double mutant (31)

is 5C in the ferric oxidation state and is analogous to the photoproduct generated when

TrHbN(Fe3+-NO) is dissociated. The isobestic point at 408.5 nm was used to follow the water

coordination subsequent to TrHbN(Fe3+

-NO) photolysis. At this wavelength we expect an

increase in absorbance when H2O binds to the heme iron and a decrease when •NO replaces

H2O. The experimental wavelength at 421 nm was employed to monitor the overall reaction:

TrHbN(Fe3+

-NO) → TrHbN(Fe3+

) → TrHbN(Fe3+

-H2O) → TrHbN(Fe3+

-NO).

94

Figure 5.3 The time courses of (panel A) O2 (68.7 μM) and (panel B) CO (50.7 μM)

recombination to TrHbN (5 μM in heme) following photolysis in 50 mM KPO4 pH 7.5 + 50 μM

EDTA at 23 °C. The figure shows the single exponential fits and residuals to the kinetic traces

measured at 411 and 420 nm for O2 and CO, respectively.

95

Figure 5.4 Kinetic traces showing the recombination of CO subsequent to nanosecond

photodissociation of the CO saturated derivatives of wild-type TrHbN and its distal mutants. The

recombination traces are displayed on a log-log plot with the Y axis corresponding to normalized

absorbance and the X axis to time subsequent to photodissociation. The traces are color coded as

follows: wild-type TrHbN in green, the Tyr(B10)Phe/Gln(E11)Val double mutant in red; the

Tyr(B10)Phe single mutant in black and the Gln(E11)Val single mutant in blue. All the samples

except the double mutant are solution phase at pH 7.5. The trace from the double B10/E11

mutant (red) is from a sample encapsulated in a thin porous sol-gel bathed in buffer. Essentially

identical kinetics were obtained for the solution phase sample of the double B10/E11 mutant but

due to the low concentration of the sample that trace was of poor quality.

96

Figure 5.5 Kinetic traces illustrating the absorbance changes following photodissociation of

TrHbN(Fe3+

-NO) and Mb(Fe3+

-NO) at 23 °C. Panel A shows the absorbance changes

corresponding to the water coordination processes followed at 408.5 nm (red) for TrHbN(Fe3+

-

NO) (8.31 μM in heme) and 410 nm (blue) for Mb(Fe3+

-NO) (7.37 μM in heme) preequilibrated

with 100% •NO. The solid lines are the results of the exponential fits (black). Panel B shows the

kinetic traces acquired at 410 nm for TrHbN(Fe3+

-NO) (9.6 μM in heme) preequilibrated with

25 % (black), 50 % (blue) and 100 % (red) •NO. Panel C shows kinetic traces obtained at

421 nm following photolysis of TrHbN(Fe3+

-NO) (8.31 μM in heme) preequilibrated with

100% •NO.

97

6.

Chapitre 6

Structural characterization of the tunnels of Mycobacterium

tuberculosis truncated hemoglobin N from molecular

dynamics simulations

6.1. Résumé

La structure de la forme oxygénée de TrHbN de Mycobacterium tuberculosis montre un réseau

de ponts hydrogène au site distal incluant la Tyr33(B10), la Gln58(E11) ainsi que l’O2 lié à

l’hème. De plus, la structure de TrHbN montre un réseau de cavités hydrophobes organisé dans

l’expace selon deux branches orthogonales. Dans les présents travaux, la structure et la

dynamique de la forme oxygénée et deoxygénée de TrHbN en présence d’un solvant explicite a

été étudiée à partir de 100 ns de simulation de dynamique moléculaire (DM). Les résultats

montrent que dépendamment de la présence ou absence d’une molécule d’O2 coordonnée au fer,

la chaîne latérale de Tyr33(B10) et celle de Gln58(E11) adoptent deux configurations distinctes

de concert avec la réorganisation du réseau de pont hydrogènes. En addition, nos données

indiquent que la Tyr33(B10) et la Gln58(E11) contrôlent la dynamique de la Phe62(E15). Chez

deoxy-TrHbN, la Phe62(E15) est restreinte à une configuration. Suivant la liaison à l’hème d’une

molécule d’O2, la conformation de Gln58(E11) change et la Phe62(E15) fluctue entre deux

configurations. Nous avons aussi réalisé une étude systématique des tunnels de TrHbN en

analysant des milliers d’instannés de trajectoire à l’aide de CAVER. Les résultats montrent que

la formation des tunnels résulte de la réorganisation dynamique des cavités hydrophobes. Les

analyses indiquent que la présence de ces cavités est liée à la structure rigide de TrHbN et ont

aussi mis en évidence deux autres tunnels non observés dans la structure cristalline, soient les

tunnels EH et BE, liant la surface au site actif de TrHbN. Les cavités ont un volume suffisant

pour accueillir et entreposer plusieurs molécules de ligand. La dynamique des tunnels est

contrôlée par la conformation de la chaîne latérale de Tyr33(B10), de Gln58(E11) et de

Phe62(E15). Aussi, en contraste avec de récents travaux récemment publiés, notre approche

98

systématique montre que la présence ou l’absence d’une molécule d’O2 ne contrôle pas

l’ouverture du tunnel long mais plutôt celle du tunnel EH. De plus, nos données ont mené à une

conclusion nouvelle et différente sur l’impact de la Phe62(E15) sur la configuration des tunnels.

Nous proposons que le tunnel EH et le tunnel long sont utilisés pour entreposer des ligands. Dans

l’ensemble, nos travaux poussent notre compréhension sur la fonction de TrHbN et sur la

diffusion des substrats à l’intérieur des protéines.

6.2. Abstract

The structure of oxygenated TrHbN from Mycobacterium tuberculosis shows an extended heme

distal hydrogen-bond network that includes Tyr33(B10), Gln58(E11), and the bound O2. In

addition, TrHbN structure shows a network of hydrophobic cavities organized in two orthogonal

branches. In the present work, the structure and the dynamics of oxygenated and deoxygenated

TrHbN in explicit water was investigated from 100 ns molecular dynamics (MD) simulations.

Results show that, depending on the presence or the absence of a coordinated O2, the Tyr33(B10)

and Gln58(E11) side chains adopt two different configurations in concert with hydrogen bond

network rearrangement. In addition, our data indicate that Tyr33(B10) and Gln58(E11) control

the dynamics of Phe62(E15). In deoxy TrHbN, Phe62(E15) is restricted to one conformation.

Upon O2 binding, the conformation of Gln58(E11) changes and residue Phe62(E15) fluctuates

between two conformations. We also conducted a systematic study of TrHbN tunnels by

analyzing thousands of MD snapshots with CAVER. The results show that tunnel formation is

the result of the dynamic reshaping of short-lived hydrophobic cavities. The analyses indicate

that the presence of these cavities is likely linked to the rigid structure of TrHbN and also reveal

two tunnels, EH and BE, that link the protein surface to the buried distal heme pocket and not

present in the crystallographic structure. The cavities are sufficiently large to accommodate and

store ligands. Tunnel dynamics in TrHbN was found to be controlled by the side-chain

conformation of the Tyr33(B10), Gln58(E11), and Phe62(E15) residues. Importantly, in contrast

to recently published works, our extensive systematic studies show that the presence or absence

of a coordinated dioxygen does not control the opening of the long tunnel but rather the opening

of the EH tunnel. In addition, the data lead to new and distinctly different conclusion on the

99

impact of the Phe62(E15) residue on TrHbN tunnels. We propose that the EH and the long

tunnels are used for apolar ligands storage. The trajectories bring important new structural

insights related to TrHbN function and to ligand diffusion in proteins.

6.3. Introduction

In Mycobacterium tuberculosis, the glbN gene encodes the truncated hemoglobin TrHbN.

Disruption of glbN in M. bovis BCG results in a dramatic reduction in the NO consuming

activity of stationary phase cells and impairs the ability of cells to protect aerobic respiration

from the inhibition by NO.1 This functional assessment is supported by the observation that

oxygenated TrHbN catalyses the very rapid oxidation of nitric oxide (NO) into nitrate with a

second-order rate constant k ≈ 745 μM-1

s-1

at 23°C.

The dioxygen ligand in TrHbN is fully buried within the distal site cavity. In both crystal

subunits (PDB accession code 1IDR2), the O2 is tilted by ≈ 110° relative to the Fe axial bond

pointing in the direction of residue Val94(G8). Thus, both oxygen atoms in the O2 are at

hydrogen bonding distance from the phenolic OH group of Tyr33(B10) (average 3.12 Å), which

also forms a hydrogen bond with the NE2 atom of Gln58(E11). Notably, resonance Raman

investigations on oxy-TrHbN have suggested that stabilization of the heme-bound O2 occurs

through a hydrogen bond between the Tyr33(B10) OH group and the proximal O atom of the

ligand.3 Accordingly, site-specific mutations of Tyr33(B10) to either Leu or Phe result in a shift

of the Fe–O bond stretching frequency from 560 to 570 cm-1

, that is, to a stretching frequency

identical to that of vertebrate and nonvertebrate oxygenated Hbs and Mbs.3,4

Kinetic analysis of

the Tyr33(B10)Phe mutant showed a 150-fold increase in the dissociation rate of O2 pointing to a

crucial role for this residue in O2 stabilization. The possible role of Gln58(E11) in O2

stabilization remains to be established.

In contrast to myoglogin and hemoglobin, ligand diffusion to the heme in TrHbN may occur via

an apolar cavity system connecting the heme distal cavity to two distinct protein surface sites.2 In

the crystal structure of TrHbN, the cavity system is organized in two roughly orthogonal

branches (hereafter referred as the long and short tunnels) linking the protein surface to the distal

100

heme pocket (Fig. 6.1). The access to the long tunnel is defined by surface residues Ile19(A15),

Ala24(B1), Val28(B5), Ala105(G19), and Val107(GH5) from the AB and GH loops (hinges).

Access to the short tunnel is defined by residues Phe91(G5), Ala95(G9), Leu116(H8), and

Ala119(H11) from the G and H helices. Treatment of TrHbN crystals under xenon pressure led

to binding of xenon atoms at five binding sites along the protein matrix tunnel (see Fig. 6.1)

supporting the potential role of the tunnels in diffusion and accumulation of low-polarity

molecules/ligands.5 The Xe1 and Xe5 binding sites are located along the long tunnel. The Xe2

binding site lies in the short tunnel while the Xe3 is located at the short tunnel entrance. Finally,

the Xe4 is located in a hydrophobic crevice leading to the short tunnel entrance.

Crystallographic studies on oxygenated2 and cyanomet

6 TrHbN derivatives revealed that in

TrHbN, the short and the long tunnels are separated by Phe62(E15), which can adopt two

conformations, referred as the open and closed conformations. In addition, molecular dynamics

simulations of TrHbN suggested that Phe62(E15) may act as a gating residue that would control

the diffusion of apolar substrate molecules along the long tunnel.5,7,8

It was also proposed that

dioxygen binding to the iron triggers structural fluctuations leading to the opening of the long

tunnel, a possible explanation to the higher bimolecular rate constant of the NO-dioxygenase

reaction (745 lM21s21) than that of the O2 binding (25 μM-1

s-1

).8

Although tunnels are observed in many other proteins,9–15

very little or nothing is known about

their formation and dynamics, nor the typical protein features necessary for their presence.

Relying simply on the crystal structure for the identification of the tunnels is not absolute as

protein dynamics is sometimes necessary for a tunnel to form.11,16

Similarly, the diffusion of

apolar ligands cannot always be predicted from the static crystal structure as they proceed

through packing defects arising at specific positions observable from protein dynamics.11

Recently, a study of the O2 migration pathways inside 12 monomeric globins13

led to the

conclusion that there is no direct relation between the conserved tertiary structure fold and the

shape and topology of O2 pathway networks. The only correlation between these pathways found

by the authors is the presence of hydrophobic residues. To our knowledge, there is not yet a

study describing the dynamical behavior of tunnels in proteins, and there is no structural

indication to what is necessary for tunnel formation. The systematic characterization of the

tunnel system in TrHbN is thus crucial to understand its function and is of fundamental interest

101

to other tunnel containing proteins. In the present work, a total of 100 nanoseconds of

simulations were generated on both oxy-TrHbN and deoxy-TrHbN. These simulations allowed

us to revisit the TrHbN tunnel system. First, we examined the impact of the presence of the O2

bound to the heme on the active site key residues Tyr33(B10) and Gln58(E11), and the

consequences for Phe62(E15) in the long tunnel. Next, we investigated tunnel formation from

the dynamics of cavities and the structural features of TrHbN leading to the formation of

cavities, that is, the rigidity of its structure and the mobility of residues lining the cavities.

Finally, we systematically characterized the different tunnels of TrHbN and their relationship

with the dynamics of the key residues. Trajectory analysis revealed that the tunnels are formed

from short-lived cavities of various shapes and volumes, and that TrHbN structure hosts a tunnel

system more complex than that which we first expected from the crystal structure. In addition to

the long and the short tunnels, two additional tunnels were identified. The Phe62(E15) residue

was found to control the dynamics of two of these tunnels, but in contrast to the conclusions of

previous studies,8 this is not triggered by the binding of the dioxygen to the iron.

6.4. Methods

Force field optimization of the oxygenated-heme atomic charges and Fe–O–O angle

parameter – CHARMM22 6-liganded heme force field parameters were developed by Kuczera

et al.17

to simulate CO-bound hemoprotein. To simulate O2-bound hemoprotein, the atomic

charges of heme prosthetic group as well as the coordinated oxygen and the Fe–O–O angle

parameter were optimized following the standard parametrization protocol for the CHARMM22

force field.18

Ab initio quantum mechanical (QM) calculations were performed using the

program Gaussian 03.19

The B3LYP/6-31G* level of theory was used for the initial geometry

optimization and subsequent single point calculations. This level of theory was applied

successfully to the parametrization of the CHARMM force field of iron–porphyrin systems.20

The atoms included for this procedure are those of the central iron–porphyrin ring and those of

the linking molecules O2 and imidazole. The heme side groups were omitted. Initial atom

coordinates were taken from the oxygenated TrHbN crystal structure (PDB accession code

102

1IDR 2). Missing hydrogen atoms were added using WebMO

21. The CHARMM atom types used

in this optimization are presented as supplemental material (Annexe 1, Fig. S1).

The optimized charges as well as the force constant for the Fe–O–O angle are presented in

Annexe 1, Table S1 and S2. The atomic charges were obtained from the optimized geometry

Mulliken charges, adjusted consistently with the electrostatic fitting procedure of the

parametrization protocol. The potential energy of interaction between the dioxygen ligand and a

water molecule was calculated from the difference in QM energies of a gas-phase system-water

complex and the isolated molecules. For this calculation, the single point calculations were

performed while keeping the intramolecular geometry of each molecule fixed. The force constant

for the Fe–O–O bond angle and the corresponding equilibrium angle were calculated from the

QM optimized geometries. The potential energy surface was obtained from increments of 10°

from 60° to 300°, with smaller increments around the minimum (122.22°) and the maximum

(180°).

Simulation details – Simulations were performed using CHARMM22 and the CHARMM22 all-

atom potential energy parameter set18

with phi, psi cross term map correction (CMAP)23

and

modified TIP3P waters.24

Electrostatic interactions were calculated via the Particle Mesh Ewald

method,25

using a sixth-order spline interpolation for complementary function, with κ = 0.34 Å-1

,

and a fast-Fourier grid density of ≈ 1 Å-1

. Cutoffs for the real space portion of the Particle Mesh

Ewald calculation and the truncation of the Lennard-Jones interactions were 10 Å, with the latter

smoothed via a shifting function over the range of 8 Å to 10 Å. The SHAKE algorithm26

was

used to constrain all covalent bonds involving hydrogen atoms. All simulations employed the

leapfrog algorithm and an integration step of 1 femtosecond (fs). Coordinates were saved every

picosecond (ps) for analysis. Nonbond and image lists were updated heuristically. All

simulations were performed at constant pressure and temperature (NPT ensemble) using Hoover

algorithm for temperature control.27

The mass of the thermal piston was 20,000 kcal•mol-1

•ps2

and the mass of the pressure piston equalled 1000 amu. All simulations were carried out at 298 K

and 1 atm. The net translation and rotation of the systems were removed every 10,000 steps.

103

Systems setup – Coordinates for MD simulations were taken from the crystal structure of wild-

type oxy-TrHbN (PDB entry 1IDR2). As there is no experimental structure of deoxy-TrHbN

available, the deoxy structure was obtained by deleting the oxygen molecule. This methodology

is supported by the low RMSD observed between the experimental structures of liganded and

unliganded states of hemoglobins available from the PDB database. As expected, no noticeable

reorganization of TrHbN in the absence of the heme-bound oxygen (except for residues

Tyr33(B10), Gln58(E11), and Phe62(E15)) was observed from the simulations (average RMSD

of 0.86 Å). It is worth mentioning that the same methodology was used by another group

studying the same protein.7,8

All ionizable residues were considered in their standard protonation state at pH 7 with neutral

histidines proton placed at the ND1 position. Missing coordinates from crystal structure were

built using the internal coordinates definition of CHARMM. For each subunit, the carboxy

terminal end was optimally positioned by performing 3 nanoseconds (ns) Langevin dynamics

with a 1 fs time step and a friction coefficient FBETA of 5 ps-1

while keeping constrained all

coordinates from crystal structure. The converged structures were immersed in a rhombic

dodecahedron unit cell containing pre-equilibrated TIP3P water molecules (8330 molecules). Six

sodium ions were added to neutralize the charge of the systems. Water molecules within 2.8 Å of

any protein atom were deleted yielding to about 24,350 atoms for each system. Before to the

initiation of MD simulations, the energy of the solvated systems was minimized with two cycles

of 500 steps of steepest descent followed by 500 steps of Adopted Basis Newton-Raphson.

During energy minimization, the protein coordinates were kept constrained.

To increase sampling, two trajectories were generated using A and B crystal subunits, either in

presence or absence of the coordinated oxygen, for a total of four trajectories of 25 ns. Data were

collected for the last 20 ns for further analysis. A 25 ns trajectory took about 1250 h to produce

on 8 AMD 2.2 GHz Opteron 248 processors interconnected with Infiniband network adapters.

Analysis of tunnels – CAVER28

was used to study TrHbN tunnels with a grid resolution of

0.5 Å. The dynamic character of each tunnel was determined using 2000 MD frames from each

simulation, covering the whole simulation over the last 20 ns (1 frame every 10 ps). Tunnel

104

profiles, that is, the average tunnel radius along its length, were calculated from the accessible

paths detected by CAVER.


6.5.1. Active site configurations

Resonance Raman spectrometry and X-ray crystallography revealed that in TrHbN both

Tyr33(B10) and Gln58(E11) can interact with the ligand.2,29

Also, previous molecular dynamics

simulations showed that the hydrogen bond network topology differs depending on the presence

of the iron bounded O2.8 Therefore, the conformations of Tyr33(B10) and Gln58(E11) were

studied depending on the presence or absence of a coordinated O2.

6.5.2. MD simulations of oxygenated TrHbN

Two trajectories were produced, one for each crystal subunit. The active site configuration

observed throughout these trajectories corresponded to that of the crystal structure, similarly to

recent MD simulations7,8

and is shown in Figure 6.2, top. The configuration of the active site can

be expressed by the distance separating the different chemical groups lying in the distal heme

pocket. A summary of the average distance separating the relevant atoms is shown in Table 6.1.

The average distances between Tyr33(B10) OH atom and the proximal and distal oxygens of the

bound O2 are 3.52 ± 0.01 Å and 2.83 ± 0.01 Å respectively. These distances are in agreement

with the strong interactions observed experimentally between Tyr33(B10) and the heme-bound

O2.29

It is noteworthy that these distances are slightly different than the distances observed in the

crystal structure where both oxgyens are practically equidistant to Tyr33(B10) OH atom (~3.1

Å).2 Partial charges on the proximal and distal oxygens of -0.18e and -0.32e, respectively, lead

the Tyr33(B10) OH atom to come closer to the distal oxygen. Similar distances (3.27 Å and 2.76

Å) were obtained with oxy-TrHbN simulations by Bidon-Chanal and coworkers.8 The very small

deviation of this residue from the PDB structure (RMSD of 0.52 A for backbone heavy atoms)

105

indicates that there is no reorganization of the distal heme pocket residues. Although the Fe–O–

O optimized angle is 122.22°, close to the angle values of the structures, the average value

observed from the simulations is 132.6 ± 0.01°.

6.5.3. MD simulations of the ferrous unliganded TrHbN

The absence of a coordinated dioxygen causes a reorganization of the active site configuration.

In this case, the side chain of Gln58(E11) experiences a lower steric hindrance and comes closer

to the iron atom (Fig. 6.2, bottom). Following rotation of the Gln58(E11) side chain, the

Tyr33(B10) phenolic OH group becomes hydrogen bonded to the OE1 atom of Gln58(E11). This

major configuration predominates and occurs 89.3 ± 4.9% of the time. A minor configuration

corresponding to that found in oxy-TrHbN, is observed 8.7 ± 4.9% of the time. As for oxy-

TrHbN trajectories, the H-bond involving the HE21 or HE22 atom of the Gln58(E11) residue

with O atom of Tyr33(B10) is observed but is weak because of the 3.76 Å distance (Table 6.1)

and the more obtuse angle of 48°.

6.5.4. Gln58(E11) and Phe62(E15) dynamics are linked

The two conformations of the Phe62(E15) side chain observed in crystal structures suggest

mobility of the Phe62(E15) residue. Gln58(E11) is within contact distance of the Phe62(E15)

residue.2,6

Phe62(E15) was proposed to act as a gate controlling ligand diffusion in the long

tunnel.2,7,8

Furthermore, extended molecular dynamics simulations show that position of the

Gln58(E11) residue influences Phe62(E15) dynamics.8 Therefore, the side-chain dihedral angles

of these two residues were calculated from MD simulations. The results, expressed according to

the rotameric species nomenclature of Lovell et al,30

are given in Table 6.2 and represented in

Figures 6.3 and 6.4.

106

In oxy-TrHbN the Gln58(E11) side chain is confined to two rotamers, tp-100° and mm100°.

Both show the same hydrogen bonding with the Tyr33(B10) residue (Annexe 1, Fig. S2). The

side chain of Phe62(E15) moves within the long tunnel by visiting two χ1 domains, [-180°: -

140°] and [-110°: -60°] (Table 6.2, Fig. 6.3, top and Fig. 6.4, bottom left), corresponding to the

t80° (χ1 ‘‘trans’’) and m-30°/m-85° (χ1 ‘‘minus’’) rotamers, respectively. For the remainder of

the text, the Phe62(E15) t80° rotamer will be referred to as the T (‘‘trans’’) state and the m-

30°/m-85° rotamers as the M (‘‘minus’’) state. When Phe62(E15) is in the T state, the phenyl

ring is located farther from the heme and χ2 is mainly found in the domain [0°: +90°]. In

contrast, when Phe62(E15) is in the M state, the phenyl ring is closer to the heme and χ2 is

unrestricted. The M state prevails in oxy-TrHbN accounting for 84.8 ± 2.5% of the

conformations, with an average χ1 of -84.2 ± 1.2°. The difference in potential energy between M

and T states is about 0.6 kcal/mol in favour of the M state.

Deoxy-TrHbN behaves differently. Here, Gln58(E11) adopts two new rotamers (Table 6.2): tt0°

(χ1 ‘‘trans’’ and χ2 ‘‘trans’’) and mt-30° (χ1 ‘‘minus’’, χ2 ‘‘trans’’). It is noteworthy that the tp-

100° and mm100° rotamers adopted by Gln58(E11) in the oxy complex are rare, and that the tt0°

and mt-30° conformations found in the deoxy state are more typical.30

The residue Phe62(E15) is

almost restricted to the M state and the χ2 angle mainly allowed within the domain [-75°: +15°]

(Fig. 6.3, bottom , and Fig. 6.4, bottom right). The restriction of Phe62(E15) to the M state is a

direct consequence of the displacement of the Gln58(E11) OE1 atom (Fig. 6.2), which

experiences a lower steric hindrance than in oxy-TrHbN. As a consequence, the χ1 average angle

of the Phe62(E15) residue in the M state is -74.6 ± 1.2°. The relation between Gln58(E11) and

Phe62(E15) was also observed experimentally from mutants. The crystal structure of single

mutants Tyr33(B10)Phe (PDB ID: 2GKM) and Gln58(E11)(Ala/Val) (PDB ID: 2GKN and

2GLN), and the double mutant Tyr33(B10)Phe/Gln58(E11)Val (PDB ID: 2GL3) shows only one

Phe62(E15) conformation (M state).29

These experimental observations clearly support that the

dynamics of Gln58(E11) and Phe62(E15) are linked in wild type TrHbN and that Tyr33(B10) is

involved.

In conclusion, our MD simulations clearly demonstrated that Phe62(E15) and Gln58(E11)

dynamics are linked (see Fig. 6.4), and that Tyr33(B10) conformation is involved. In addition,

the ligand state of the heme also influences the dynamics of Phe62(E15) and Gln58(E11). Upon

107

O2 binding, the conformation of Gln58(E11) and Tyr33(B10) change, and consequently residue

Phe62(E15) fluctuates between the M and T states. Some of these observations are in accord

with earlier reports8,31

; however, our data lead to very different conclusions on the tunnel

dynamics (vide infra).

6.5.5. Characterization of cavities and tunnels

Dynamic modelling of tunnels from cavities – The crystal structure of TrHbN treated at high

xenon pressure shows that the protein harbors several xenon atoms at distinct sites while

showing low root-mean-square deviation (RMSD) of the backbone chain position.5 This

indicates that the TrHbN two-on-two helical fold is rigid. Figure 6.5 presents snapshots of

cavities in the protein taken from the four MD trajectories produced in this work (a movie is

available as supplemental material and at CHARMM-GUI web site32,33

). These cavities show

various shapes and volumes, and are short-lived appearing and disappearing repeatedly. On very

rare occasions, a cavity was found wide open extending from the protein surface to the distal

heme pocket. These cavities form along particular paths, hereafter referred to as tunnels (vide

infra).

To determine whether TrHbN backbone is involved in the formation of cavities, the backbone

1H–

15N order parameters (S

2) were calculated. Here, S

2 measures the degree of spatial restriction

of the 1H–

15N bond vector; its value varies from 0 to 1, where lower values indicate larger

amplitudes of internal motion.34

Typically, structured regions (α-helices and β-strands) show an

average S2 near 0.85 while exposed loops and terminal regions exhibit lower values.

35 Values of

S2 were calculated from an ensemble generated with the last 20 ns of each of the four trajectories

according to the ‘‘model-free’’ formalism introduced by Lipari and Szabo36

and the procedure

described in 37

. Briefly, S2 were calculated as the plateau value of the autocorrelation function:

108

)(lim 2

2 tCSt

Where

)()(

)]()([)(

33

22

trr

tPAtC

Here, A is a constant such that C2(0) = 1; P2 is the second order Legendre polynomial,

P2[x] = 1/2(3x2 – 1); and r(τ) is the N–H bond length at time τ. The angle brackets represent

the average over the ensemble. The unit vectors μ(τ) and μ(τ + t) describe the orientation of the

N–H vector at time τ and (τ + t) in relation to a fixed reference frame. To construct this frame,

global translational and rotational motions were removed from the trajectory by a RMSD

optimized superposition of all the backbone heavy atoms (N, C, Ca, and O) on the first

coordinate frame. Thus, S2 reflects the internal motion of the peptide plan. The recently

developed CMAP correction to the CHARMM22 force field23

greatly improved the agreement

between the MD-derived and NMR-derived dynamical parameters.38

Although a similar

agreement is expected for the S2 values obtained in the present study, validation with

experimental data is under way.

As shown in Figure 6.6, the pre-A-A loop, the E-F loop, and the loops contiguous to the C-helix

exhibit low values of S2. Not surprisingly, helical regions show the highest values of S

2. In

particular, helices B, E, G, and H that surround the cavities show an average S2 of 0.90 ± 0.03,

which is higher than the typical value of 0.85 found for structured regions. Both TrHbN termini

show maximum flexibility, notably the C-terminus, which does not adopt any secondary

structure pattern. These results emphasize that the TrHbN two-on-two a-helical fold is rigid. This

rigidity is likely necessary to generate empty volumes in the protein matrix. The inspection of the

trajectories revealed that the side-chain motions of the residues lying along the tunnels are

responsible for the dynamic reshaping of cavities. It is noteworthy that for some residues

[Ile19(A15), Val28(B5), and Ile119(H11)], the rotamers adopted by the side chains are not

typical.

109

Tunnel characterization – This section presents a systematic study of TrHbN tunnels. As

previously reported, protein dynamics is sometimes necessary to allow the formation of

tunnels.11,16

In the same concept, TrHbN cavities form along particular discrete paths in the

protein matrix leading to the formation of tunnels. The different tunnels observed are presented

in Figure 6.7. In addition to the long and the short tunnels,2 we report two tunnels, named EH

and BE, that join the protein surface to the buried distal heme pocket. These tunnels are not

present in the crystallographic structure, and are analogous to the E helix/H helix and B helix/E

helix tunnels reported by Golden and Olsen39

using Locally Enhanced Sampling MD. In contrast

to our results, these tunnels were observed only when both A and B subunits of the

crystallographic structure were present in the system.

Tunnels linking the protein surface to the distal heme pocket were rarely observed using the

usual 1.4 Å probe radius. Because substrate molecules may diffuse transiently from different

cavities, the tunnels do not need to be continuously open at the same time. Also, the bottlenecks

and broadenings located along the different tunnels may govern the substrate diffusion process.

Thus, it is important to analyze both the narrower and the broader regions of each tunnels.

CAVER used here28

is specifically designed to find and analyze paths leading from buried

protein cavities to the outside solvent. Because CAVER does not depend on a given probe

radius, it is able to pass through bottlenecks to find a path.

The profile of each tunnel, that is the average tunnel radius along the tunnels, was produced from

the analysis of thousand of MD snapshots (see Method) and are given in Figure 6.8. The length

of each tunnel as well as the average volume of the different profiles in oxy-TrHbN and deoxy-

TrHbN are summarized in Table 6.3. All distances were calculated from an arbitrarily chosen

position located at 3.8 Å from the iron centre above the pyrrole NB atom. This position was used

as the starting point for each profile as it is common to all tunnels and located outside the volume

occupied by the ligand.

Long tunnel – The long tunnel (Fig. 6.7, A) has a length of 20 Å, and as a consequence has the

biggest average volume. Two profiles were observed for the long tunnel (Fig. 6.8, A), according

to the Phe62(E15) state (M or T). When Phe62(E15) is in the M state, the profile shows a

bottleneck of 1.2 Å radius located at 6.3 Å from the starting point. This is followed by a

110

broadening of 1.6 Å radius at 11.0 Å from the origin [Fig. 6.8, A, profile P1 (circles)]. The

broader part of this profile corresponds to the Xe1 binding site reported by Milani et al.5 (see also

Fig. 6.1). Recall that Phe62(E15) is restricted to the M state in deoxy-TrHbN and fluctuates

between the M and T states in oxy-TrHbN. As a consequence, the P1 profile was observed for

99.6% of the time in deoxy-TrHbN, while it accounts for 84.8 ± 2.5% of the time in oxy-TrHbN.

The P1 profile is inverted to generate the P2 profile when Phe62(E15) is in the T state. The

broader region now is closer to the distal pocket (radius of 1.6 Å, centered at 6.0 Å) and is

followed by a bottleneck (radius of 1.2 Å, centered at 11.6 Å), [Fig. 6.8, A (squares)]. The broad

region of the P2 profile comprises the Xe5 binding site5 (Fig. 6.1). Interestingly, the Xe5 binding

site, not seen in the B subunit and observed in the A subunit of the crystal structure, has a low

occupancy factor of 0.30 (PDB ID. 1S56), which is in accord with the low frequency of

occurrence of profile P2 in oxy-TrHbN simulations. The averaged tunnel volume is similar for

both profiles, with values of 169 ± 4 Å3 for profile P1, and 173 ± 2 Å

3 for profile P2.

As mentioned earlier and as shown in Figure 6.3, the Phe62(E15) χ2 adopts a wide range of

values. To determine the impact of this dihedral angle on the profiles of the long tunnel, the

bottleneck radii were extracted with respect to χ2. As can be seen in Figure 6.9, the bottleneck

radius of the long tunnel varies according to χ2, both for the M or T Phe62(E15) states.

Physically, when χ2 varies between 50° and +90°, and between -90° and -40°, the phenyl ring of

Phe62(E15) tends to be parallel to the long tunnel axis, and therefore the average radius of the

bottleneck radius increases up to ~1.3 Å. On the other hand, when the phenyl ring is orthogonal

to the long tunnel axis (χ2 range of values [-40°: +40°]) the bottleneck radius decreases below 1.1

Å and therefore the tunnel becomes too narrow to permit the diffusion of a substrate. These

observations are in agreement with the open/closed concept from the crystallographic structure

(PDB ID: 1IDR).5 In this concept the Phe62(E15) closed conformation has dihedral angle values

of χ1 = -155.0° and χ2 = 36.5°, and the open conformation, with the phenyl ring closer to the

heme, has dihedral values of χ1 = -90.8° and χ2 = -68.4°.

Our results are in disagreement with recent MD simulations from Estrin’s group7,8

for the

closed/open concept and for the long tunnel width. From their simulations, the authors identify

the closed state with the Phe62(E15) phenyl ring closer to the heme, whereas the open state has

the Phe62(E15) phenyl ring lying roughly parallel to the axis of the tunnel. Further, Crespo et al.7

111

report that the opening of the long tunnel increases its narrowest width to values of ≈ 3.4 Å of

diameter, which is much larger than the values reported in Figure 6.9. These disagreements

might be explained by a reorganization of the pre-A and A helices reported by Bidon-Chanal et

al.8 In the crystal structure, the pre-A and A helices stabilize the region near the long tunnel

entrance with multiple salt bridges and hydrogen bonds involving residues Arg6, Ser14, Asp17,

Lys18, Gly68, and Glu114. Bidon-Chanal et al.8 report an important reorientation of the pre-A

helix to form a salt bridge between residues Arg10 and Glu70. The atoms involved in this salt

bridge are separated by ≈ 25 Å in the crystal structure; thus, this reorganization involves the

rupture of all the interactions stabilizing the region near the long tunnel entrance present in the

crystal structure. This salt bridge is not always present, and according to their observations, in its

absence the dynamical fluctuations of the backbone facilitates the relative displacement helices B

and E, which has an impact on the dynamics of Phe62(E15) residue and the long tunnel. In our

simulations, the preA helix was found flexible (Fig. 6.6), but no such reorganization was

observed.

Finally, Figure 6.9 also clearly indicates that the orientation of the Phe62(E15) χ2 dihedral angle

is the principal determinant controlling the opening of the long tunnel. This is consistent with the

observations from the trajectories of the long tunnel fully open without discontinuity in either M

or T state for both oxy and deoxy-TrHbN (recall that the T state is mostly populated by oxy-

TrHbN). Again, our results are in disagreement with the mechanism proposed by Bidon-Chanal

et al.8 where upon dioxygen binding Phe62(E15) side chain changes conformation to open the

long tunnel. In Ref. 8 there is no systematic study of the Phe62(E15) χ2 dihedral angle on the

impact of the opening of the long tunnel.

EH tunnel – The surface entrance of the EH tunnel (Fig. 6.7, C) is defined by residues

Phe61(E14), Ala65(E18), Val118(H10), and Leu122(H14), situated between the E and H helices.

As observed for the long tunnel, the Phe62(E15) side chain controls substrate access from the EH

tunnel to the distal heme pocket. The EH tunnel is approximately 15 Å long, and its volume

varies according to the conformation of the Phe62(E15) side chain (Table 6.3). As shown in

Figure 6.8, the Phe62(E15) T state favors the opening of the EH tunnel whereas the M state

drastically impairs communication with the distal heme pocket. The profiles, extracted as

function of the Phe62(E15) M and T states, are shown in Figure 6.8, B. The first 4 Å from the

112

starting point are common to both the EH and the long tunnels. In the M state, the average

volume of the EH tunnel is 118 ± 1 Å3, and its profile [Profile P1, Fig. 6.8, B (circles)] exhibits

one bottleneck located at 7.3 Å from the starting point, close to the Phe62(E15) residue.

When Phe62(E15) is in the T state, the average tunnel volume increases to 147 ± 2 Å3, and the

profile [profile P2, Fig. 6.8, B (squares)] shows now an enlargement comprising the Xe5 and the

Xe2 binding sites. As shown from the plot of the average bottleneck radius as function of the

Phe62(E15) χ2 for both the M and T states (Fig. 6.9), the opening of the EH tunnel is controlled

by χ1 but not by χ2. Figure 6.9 also shows that the bottleneck is always broader when the

Phe62(E15) is in the T state than in the M state.

When the Phe62(E15) is in the M state, the EH tunnel often merges with the long tunnel, giving

the LEH tunnel (Fig. 6.7, D). Because the M state predominates in the trajectories, the diffusion

of ligands through the EH tunnel is likely to occur most of the time in a two-step mechanism:

during the first, the substrate penetrates the protein matrix and then, crosses the Phe62(E15)

barrier. This is the first report of the EH tunnel. In the A-chain of the crystal structure of the

oxygenated form (PDB ID 1IDR), the Phe62(E15) residue shows two conformations that

correspond to the M and the T states. In agreement, in the crystal structure, the EH tunnel is

observed only when the Phe62(E15) position corresponds to the T state. Interestingly, the crystal

structure of Chlamydomonas eugametos shows the long and the EH tunnel (PDB entry 1DLY).40

Short tunnel - The surface entrance of the short tunnel (Fig. 6.7, B) is defined by residues

Phe91(G5), Ala95(G9), Leu116(H8), and Ala120(H12) and is situated between the G and H

helices. The short tunnel (Fig. 6.7, B) has a length of about 13 Å and has an average volume of

_116 ± 2 Å3 (Table 6.3). The radius of the short tunnel does not vary significantly (Fig. 6.8, B),

with only one bottleneck of 1.10Å located near to the protein surface and defined by the residues

Val94(G8), Leu98(G12), Ile119(H11), and the heme. The Xe2 and Xe3 binding sites5 are

enclosed in the short tunnel and located on either sides of this bottleneck. Trajectory analysis

revealed that the Ile119(H11) have a highest mobility than either Val94(G8) and Leu98(G12).

For the majority of the MD frames analyzed, Ile119(H11) partly hindered the short tunnel space.

The residue Ile119(H11) was found displaced for the MD frame having a continuously open

113

short tunnel. In some occasions, the short tunnel was found shifted in the direction of the

Phe91(G5) residue.

BE tunnel - The BE tunnel (Fig. 6.7, E) extends from between Tyr33(B10) and Phe46(CD1)

side chains and reaches the solvent between the B and E helices. The entrance of the BE tunnel is

defined by Tyr33(B10) backbone and by the side chains of Leu37(B14) and Met51(E4). The BE

tunnel has a length of 10.5 Å and shows two profiles according to Tyr33(B10) conformation

(Table 6.3). The first profile [profile P1, Fig. 6.8, D (circles)] occurs when the χ2 dihedral of

Tyr33(B10) lies within the domain [+75°: +135°]. In this conformation the tunnel has an average

volume of 69 ± 2 Å3 and the profile shows a bottleneck of 0.97 Å of radius at 7.1 Å from the

reference point in the distal site, making it the narrowest tunnel found in TrHbN. This bottleneck

is close to the protein surface and is located near the side chain of Leu37(B14) and Met51(E4)

and the backbone of Tyr33(B10). The BE tunnel is considerably enlarged when the aromatic ring

of the Tyr33(B10) becomes parallel to the ring of Phe46(CD1). This particular movement creates

a large opening of the BE path over the first 6 Å [profile P2, Fig. 6.8, D (squares)] increasing the

volume of the cavity by 13 Å3. Nevertheless, the narrow bottleneck near the protein surface is

kept. This movement is rare and occurs only in the simulations of deoxy-TrHbN.

6.6. Concluding remarks

Our results clearly show that (1) TrHbN hosts a complex tunnel system; (2) tunnels in TrHbN are

not static, but are rather a result of a dynamic reshaping of the cavities along a given path and (3)

the rigid structure of TrHbN backbone is required to generate and maintain the cavities. MD

simulations strongly suggest that gaseous substrates would access the active site through definite

pathways that are rarely open from the protein surface to the distal heme pocket as suggested

from the X-ray crystal structures. Rather, substrates would migrate through short-lived cavities

of various shapes and volumes.

Contrary to Estrin’s group,7,8

our data lead to different conclusions regarding the impacts of the

iron bounded dioxygen and the Phe62(E15) residue on the opening/closing of TrHbN tunnels.

First, we found that opening or closing of the long tunnel cannot be connected to the fluctuations

114

of Phe62(E15) residue around χ1 dihedral angle. Rather these fluctuations cause only a

displacement of the bottleneck along the long tunnel. Second, our data show that fluctuations

around the χ2 dihedral angle, which are associated with the phenyl ring rotation, control the

opening of the long tunnel. Third, the opening and closing of the newly discovered EH tunnel

was found to depend on the χ1 dihedral angle while χ2 dihedral angle has no effect.

Finally, the tunnels average volumes are large enough to accomodate few small apolar ligands.

As the long and the EH tunnels are the largest, and their opening is controlled via the residue

Phe62(E15) and the presence of an iron bounded dioxygen, we propose that these tunnels can be

used as storage to sustain the active site with ligands.

6.7. Acknowledgments

We are grateful to Dr. Joel M. Friedman, Dr. Beatrice A. Wittenberg, and Dr. Jonathan B.

Wittenberg for insightful discussions.

6.8. References

1. Ouellet H, Ouellet Y, Richard C, Labarre M, Wittenberg B, Wittenberg J, Guertin M.

Truncated hemoglobin HbN protects Mycobacterium bovis from nitric oxide. Proc Natl Acad

Sci USA 2002; 99:5902–5907

2. Milani M, Pesce A, Ouellet Y, Ascenzi P, Guertin M, Bolognesi M. Mycobacterium

tuberculosis hemoglobin N displays a protein tunnel suited for O2 diffusion to the heme.

EMBO J 2001;20:3902–3909.

3. Yeh SR, Couture M, Ouellet Y, Guertin M, Rousseau DL. A cooperative oxygen binding

hemoglobin from Mycobacterium tuberculosis. Stabilization of heme ligands by a distal

tyrosine residue. J Biol Chem 2000;275:1679–1684.

4. Couture M, Das TK, Lee HC, Peisach J, Rousseau DL, Wittenberg BA, Wittenberg JB,

Guertin M. Chlamydomonas chloroplast ferrous hemoglobin. Heme pocket structure and

reactions with ligands. J Biol Chem 1999;274:6898–6910.

115

5. Milani M, Pesce A, Ouellet Y, Dewilde S, Friedman J, Ascenzi P, Guertin M, Bolognesi M.

Heme-ligand tunneling in group I truncated hemoglobins. J Biol Chem 2004;279:21520–

21525.

6. Milani M, Ouellet Y, Ouellet H, Boffi MGA, Antonini G, Bocedi A, Mattu M, Bolognesi M,

Ascenzi P. Cyanide binding to truncated hemoglobins: a crystallographic and kinetic study.

Biochemistry 2004;43:5213–5221.

7. Crespo A, Marti MA, Kalko SG, Morreale A, Orozco M, Gelpi JL, Luque FJ, Estrin DA.

Theoretical study of the truncated hemoglobin HbN: exploring the molecular basis of the NO

detoxification mechanism. J Am Chem Soc 2005;127:4433–4444.

8. Bidon-Chanal A, Marti MA, Crespo A, Milani M, Orozco M, Bolognesi M, Luque FJ, Estrin

DA. Ligand-induced dynamical regulation of NO conversion in Mycobacterium tuberculosis

truncated hemoglobin-N. Proteins 2006;64:457–464.

9. Huang X, Holden HM, Raushel FM. Channeling of substrates and intermediates in enzyme-

catalyzed reactions. Ann Rev Biochem 2001;70:149–150.

10. Ciacchi L, Payne M. The entry pathway of O2 into human ferritin. Chem Phys Lett

2004;390:491–495.

11. Cohen J, Kim K, King P, Seibert M, Schulten K. Finding gas diffusion pathways in proteins:

application to O2 and H2 transport in CpI [FeFe]-hydrogenase and the role of packing defects.

Structure 2005;13:1321–1329.

12. Furse KE, Pratt DA, Schneider C, Brash AR, Porter NA, Lybrand TP. Molecular dynamics

simulations of arachidonic acid-derived pentadienyl radical intermediate complexes with

COX-1 and COX-2: insights into oxygenation regio- and stereoselectivity. Biochemistry

2006;45:3206–3208.

13. Cohen J, Schulten K. O2 migration pathways are not conserved across proteins of a similar

fold. Biophys J 2007;93:3591–3600.

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

Exploring molecular oxygen pathways in Hansenula polymorpha copper-containing amine

oxidase. J Biol Chem 2007;282:17767–17776.

15. Saam J, Ivanov I, Walther M, Holzhutter H-G, Kuhn H. Molecular dioxygen enters the active

site of 12/15-lipoxygenase via dynamic oxygen access channels. Proc Natl Acad Sci USA

2007;104:13319–13324.

16. Amara P, Andreoletti P, Jouve HM, Field MJ. Ligand diffusion in the catalase from Proteus

mirabilis: a molecular dynamics study. Prot Sci 2001;10:1927–1935.

116

17. Kuczera K, Kuriyan J, Karplus M. Temperature dependence of the structure and dynamics of

myoglobin. A simulation approach. J Mol Biol 1990;213:351–373.

18. Mackerell AD, Bashford D, Bellott M, Dunbrack RL, Evanseck JD, Field MJ, Fischer S, Gao

J, Guo H, Ha S, Joseph-Mccarthy D, Kuchnir L, Kuczera K, Lau FTK, Mattos C, Michnick

S, Ngo T, Nguyen DT, Prodhom B, Reiher WE, Roux B, Schlenkrich M, Smith JC, Stote R,

Straub J, Watanabe M, Wiorkiewicz-Kuczera J, Yin D, Karplus M. All-atom empirical

potential for molecular modeling and dynamics studies of proteins. J Phys Chem B

1998;102:3586–3616.

19. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, Montgomery

Jr.JA, Vreven T, Kudin KN, Burant JC, Millam JM, Iyengar SS, Tomasi J, Barone V,

Mennucci B, Cossi M, Scalmani G, Rega N, Petersson GA, Nakatsuji H, Hada M, Ehara M,

Toyota K, Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda Y, Kitao O, Nakai H, Klene

M, Li X, Knox JE, Hratchian HP, Cross JB, Bakken V, Adamo C, Jaramillo J, Gomperts R,

Stratmann RE, Yazyev O, Austin AJ, Cammi R, Pomelli C, Ochterski JW, Ayala PY,

Morokuma K, Voth GA, Salvador P, Dannenberg JJ, Zakrzewski VG, Dapprich S, Daniels

AD, Strain MC, Farkas O, Malick DK, Rabuck AD, Raghavachari K, Foresman JB, Ortiz JV,

Cui Q, Baboul AG, Clifford S, Cioslowski J, Stefanov BB, Liu G, Liashenko A, Piskorz P,

Komaromi I, Martin RL, Fox DJ, Keith T, Al-Laham MA, Peng CY, Nanayakkara A,

Challacombe M, Gill PMW, Johnson B, Chen W, Wong MW, Gonzalez C, Pople JA.

Gaussian 03, Revision C.02. Wallingford, CT: Gaussian, Inc.; 2004.

20. Autenrieth F, Tajkhorshid E, Baudry J, Luthey-Schulten Z. Classical force field parameters

for the heme prosthetic group of cytochrome c. J Comput Chem 2004;25:1613–1622.

21. WebMO 7.0. Available at: www.webmo.net.

22. Brooks BR, Bruccoleri RE, Olafson BD, Karplus DJ. Charmm: a program for

macromolecular energy, minimization, and dynamics calculations. J Comput Chem

1983;4:187.

23. Mackerell AD, Feig M, Brooks CL. Extending the treatment of backbone energetics in

protein force fields: Limitations of gas-phase quantum mechanics in reproducing protein

conformational distributions in molecular dynamics simulations. J Comput Chem 2004;

25:1400–1415.

24. Price DJ, CLB, III. A modified TIP3P water potential for simulation with Ewald summation.

J Chem Phys 2004;121:10096–10103.

25. Darden T, York D, Pedersen L. Particle mesh Ewald: an Nlog(N) method for Ewald sums in

large systems. J Chem Phys 1993;98: 10089–10092.

26. Ryckaert J-P, Ciccotti G, Berendsen HJC. Numerical integration of the cartesian equations of

motion of a system with constraints: molecular dynamics of n-alkanes. J Comput Phys

1977;23:327–341.

http://www.webmo.net/

117

27. Feller SE, Zhang Y, Pastor RW, Brooks BR. Constant pressure molecular dynamics

simulation: the Langevin piston method. J Chem Phys 1995;103:4613–4621.

28. Petrek M, Otyepka M, Banas P, Kosinova P, Koca J, Damborsky J. CAVER: a new tool to

explore routes from protein clefts, pockets and cavities. BMC Bioinform 2006;7:3161.

29. Ouellet Y, Milani M, Couture M, Bolognesi M, Guertin M. Ligand interactions in the distal

heme pocket of Mycobacterium tuberculosis truncated hemoglobin N: roles of TyrB10 and

GlnE11 residues. Biochemistry 2006;45:8770–8781.

30. Lovell SC, Word JM, Richardson JS, Richardson DC. The penultimate rotamer library.

Proteins 2000;40:389–408.

31. Bidon-Chanal A, Martı´ MA, Estrin DA, Luque FJ. Dynamical regulation of ligand migration

by a gate-opening molecular switch in truncated hemoglobin-N from Mycobacterium

tuberculosis. J Am Chem Soc 2007;129:6782–6788.

32. Jo S, Kim T, Iyer V, Im W. 2007 CHARMM-GUI: a Web-based Graphical User Interface for

CHARMM. Available at: http://www.charmm-gui.org

33. Daigle R. 2008. Mycobacterium tuberculosis hemoglobin N cavities. Available at:

http://www.charmm-gui.org/?doc5gallery&id533.

34. Cavanagh J, Fairbrother WJ, Palmer IAG, Skelton NJ. Protein NMR spectroscopy: principles

and practice. San Diego, CA: Academic Press, Inc.; 1996.

35. Palmer AG. NMR probes of molecular dynamics: overview and comparison with other

techniques. Annu Rev Biophys Biomol Struct 2001;30:129–155.

36. Lipari G, Szabo A. Model-free approach to the interpretation of nuclear magnetic resonance

relaxation in macromolecules. 1. Theory and range of validity. J Am Chem Soc

1982;104:4546–4559.

37. Buck M, Karplus M. Internal and overall peptide group motion in proteins: molecular

dynamics simulations for lysozyme compared with results from X-ray and NMR

spectroscopy. J Am Chem Soc 1999;121:9645–9658.

38. Buck M, Bouguet-Bonnet S, Pastor RW, MacKerell AD. Importance of the CMAP

correction to the CHARMM22 protein force field: dynamics of hen lysozyme. Biophys J

2006;90:L36–L38. Available at: http://dx.doi.org/10.1529/biophysj.105.078154.

39. Golden SD, Olsen KW. Identification of ligand-binding pathways in truncated hemoglobins

using locally enhanced sampling molecular dynamics. GLOBINS AND OTHER NITRIC

OXIDE-REACTIVE PROTEINS, PART B 2008;437:459–475.

http://www.charmm-gui.org/

http://www.charmm-gui.org/?doc5gallery&id533

http://dx.doi.org/10.1529/biophysj.105.078154

118

40. Pesce A, Couture M, Dewilde S, Guertin M, Yamauchi K, Ascenzi P, Moens L, Bolognesi

M. A novel two-over-two alpha-helical sandwich fold is characteristic of the truncated

hemoglobin family. EMBO J 2000;19:2424–2434.

41. DeLano WL. 2002. The PyMOL Molecular Graphics System. Available at:

http//www.pymol.org.

119

Table 6.1 Interatomic Distances Between Relevant Atoms From Trajectories

Atom pair* Oxy-TrHbN Deoxy-TrHbN major Deoxy-TrHbN minor

B10-E11-N 3.30 ± 0.03 2.83 ± 0.04 3.76 ± 0.04

B10-E11-O 5.37 ± 0.02 1.87 ± 0.04 5.83 ± 0.04

B10-O1 3.52 ± 0.02 – –

B10-O2 2.83 ± 0.01 – –

B10-Iron 5.20 ± 0.02 5.73 ± 0.10 5.16 ± 0.07

E11-N-Iron 4.78 ± 0.05 3.79 ± 0.08 4.70 ± 0.04

E11-N-Pyrrol-A 4.46 ± 0.04 3.87 ± 0.08 4.26 ± 0.03

E11-N-Pyrrol-B 3.99 ± 0.04 3.54 ± 0.03 3.95 ± 0.06

* E11–N, side chain amide N atom; E11–O, side chain amide O atom; B10, phenolic O atom; O1, proximal oxygen;

O2, distal oxygen; pyrrole, heme pyrrolic N

atom.

120

Table 6.2 Distribution of the Different Rotameric Species Encountered During Simulations for

Q(E11) and F(E15)

Amino acid Rotamera Oxy-TrHbN Deoxy-

TrHbN

Deoxy-

TrHbN-minor

Q(E11) tp-100° 75.7 – 80.5

mm100° 21.8 – 13.5

tt0° – 13.0 –

mt-30° – 80.7 –

outliers/othersb 2.7 6.3 6.0

F(E15) t80° 11.6 0.3 9.8

m-30° 32.4 53.3 33.7

m-85° 52.4 46.3 54.2

outliers 3.6 0.1 2.3

a Nomenclature according to Lovell et al.

30

b Results include 1.6% and 2.7% of rotamer mm-40° in oxy-TrHbN and deoxy-TrHbN

respectively.

121

Table 6.3 Tunnels Physical Properties

Tunnels Length (Å) Profile Volume (Å3)

Long 20.0 P1

P2

170 ± 4

173 ± 2

Short 13.0 – 116 ± 2

EH 15.0 P1

P2

118 ± 1

147 ± 2

BE 10.5 P1

P2

69 ± 2

83 ± 4

122

Figure 6.1 TrHbN structure (PDB entry 1IDR, subunit A). Helical regions are identified. The

B,E,G and H helices are colored in blue, green, yellow, and purple, respectively. The five Xenon

binding sites5 from crystal structure PDB ID 1S56 are represented by the blue hard spheres. The

long and the short tunnels are indicated by the shaded zones. The picture was generated using

PyMOL.41

123

Figure 6.2 Active site configurations from typical MD frames for oxy-TrHbN (top) and deoxy-

TrHbN (bottom). The heme, the proximal histidine, Phe62(E15), Gln58(E11), and Tyr33(B10)

are represented with balls and sticks. Hydrogen bonds are represented by dashed lines with their

corresponding length. Additional pictures from different rotamers are available as supplemental

material (Annexe 1, Fig. S3). Pictures were generated using PyMOL.41

The stereo view

representation of this figure is also available as supplemental material (Annexe 1, Fig. S3).

124

Figure 6.3 Phe62(E15) χ1 as function of χ2 for oxy-TrHbN (top) and deoxy-TrHbN (bottom).

The T state corresponds to the v1 dihedral domain [-180°:-140°] while it is [-110°: -60°] for the

M state. By convention, χ2 range is [-90°: +90°] for phenylalanine and tyrosine. For each

configuration, 2000 MD frames were randomly selected.

125

Figure 6.4 Gln58(E11) (top) and Phe62(E15) (bottom) χ1 dihedral angle as function of simulation

time for oxy-TrHbN (left) and deoxy-TrHbN (right). Only the results from subunit A are shown.

126

Figure 6.5 Different snapshots of cavities in TrHbN. The BEG, and H helices are colored in blue,

green, yellow, and purple, respectively. The cavities are represented in grey. The pictures were

generated using PyMOL.41

127

Figure 6.6 Backbone 1H-

15N order parameters as function of residue sequence number calculated

from trajectories of A-TrHbN (continuous) and B-TrHbN (dotted) in oxy-TrHbN (top) and

deoxy-TrHbN (bottom). TrHbN helices are indicated by the boxes and their corresponding code

letter. Order parameters of residues 2 and 130 to 136 did not converge.

128

Figure 6.7 Representation of the long (A), short (B), EH (C), LEH (D), and BE (E) tunnels.

Helices B, E, G, and H are colored in blue, green, yellow, and purple, respectively. The pictures

were generated using PyMOL41

and PyMOL plugin CAVER28

.

129

Figure 6.8 Profiles generated for each tunnel leading from distal heme pocket to the bulk solvent.

Profiles of the (A) long and (B) EH tunnels were extracted regarding to the Phe62(E15) M

(circles) and T (squares) states. (C) Averaged profile of the short tunnel. (D) BE tunnel profiles

extracted from MD frames having the typical Tyr33(B10) χ2 dihedral domain (within [70°:

135°]) (circles) and from flipping events of the aromatic ring (squares). The relative position of

some Xenon binding sites are shown.5 Standard errors typically fall below 0.01 Å. Profiles were

generated using CAVER.28

130

Figure 6.9 Averaged minimum radius (bottleneck) of the (A) long and (B) EH tunnels according

to the Phe62(E15) χ2 dihedral. MD frames having the Phe62(E15) in the T (grey) and the M

(black) states were used. Data were collected using CAVER.28

A

B

131

7.

Chapitre 7

Theoretical Investigations of Nitric Oxide Channeling in

Mycobacterium tuberculosis Truncated Hemoglobin N

7.1. Résumé

L’hémoglobine tronquée du groupe I TrHbN de Mycobacterium tuberculolsis catalyse

l’oxydation de l’oxyde nitrique en nitrate selon la constante de vitesse d’ordre 2

k’NOD ≈ 745µM-1

•s-1

à 23°C (communément appelée NOD pour « nitric oxide dioxygenase

reaction »). Il a été proposé que cette haute efficacité soit associée à la présence de tunnels

hydrophobes prenant place à l’intérieur de la structure de TrHbN en permettant la diffusion

rapide des substrats vers le site actif. Dans ce travail, nous avons étudié les mécanismes de

diffusion du •NO à l’intérieur de la structure de TrHbN dans le contexte de la réaction NOD en

utilisant deux approches différentes. Des simulations de dynamiques moléculaires de TrHbN ont

été réalisées en présence de molécules de •NO explicites. La diffusion de •NO du solvant

jusqu’au site actif a été observée pour chacune des simulations réalisées. Ces simulations ont

révélé que les •NO interagissent avec des régions spécifiques de la surface de TrHbN constituées

de chaînes hydrophobes et situées aux entrées des tunnels. L’entrée de •NO et la diffusion

interne se sont produites par les tunnels Long, Court et EH identifiés précédemment. Le tunnel

Court a été préférentiellement utilisé pour atteindre le site actif. Cette préférence est attribuée à la

topologie en entonnoir ainsi qu’au caractère hautement hydrophobe couvrant une large zone

autour de l’entrée de ce tunnel. Ces propriétés favorisent la formation fréquente de cavités à

l’interface entre le solvant et la protéine suffisamment grandes pour accueillir trois molécules de

•NO. Ceci accélère la capture du •NO et son entrée subséquente. L’importance du caractère

hydrophobe est soulignée davantage par la comparaison avec un mutant ayant les entrées mutées

par des résidus polaires. Une carte complète des sentiers de diffusion du •NO à l’intérieur de

TrHbN a été calculée et il a été démontré que les •NO diffusent d’une cavité xénon vers une

autre. Ce schéma est en parfait accord avec la carte tridimensionnelle de l’énergie libre calculée

132

par échantillonnage implicite de ligands (ligand implicit sampling). Les trajectoires ont démontré

que le •NO modifie significativement la dynamique de résidus clefs : la Phe62(E15) proposé

comme une barrière contrôlant le trafic à l’intérieur du tunnel Long ainsi que la Ile119(H11)

localisée à l’entrée du tunnel Court. Il est important de noter que la diffusion du •NO est

beaucoup plus rapide que celle reportée précédemment chez la myoglobine. Les résultats

présentés dans ce travail contribuent à l’avancement des connaissances sur les mécanismes de

diffusions des substrats gazeux apolaires à l’intérieur des protéines.

7.2. Abstract

Mycobacterium tuberculosis group I truncated hemoglobin TrHbN catalyzes the oxidation of

nitric oxide (•NO) to nitrate with a second-order rate constant k ≈745µM-1

•s-1

at 23°C (nitric

oxide dioxygenase reaction). It was proposed that this high efficiency is associated with the

presence of hydrophobic tunnels inside TrHbN structure that allow substrate diffusion to the

distal heme pocket. In this work, we investigated the mechanisms of •NO diffusion within

TrHbN tunnels in the context of the nitric oxide dioxygenase reaction using two independent

approaches. Molecular dynamics simulations of TrHbN were performed in the presence of

explicit •NO molecules. Successful •NO diffusion from the bulk solvent to the distal heme

pocket was observed in all simulations performed. The simulations revealed that •NO interacts

with TrHbN at specific surface sites, composed of hydrophobic residues located at tunnel

entrances. The entry and the internal diffusion of •NO inside TrHbN were performed using the

Long, Short, and EH tunnels reported earlier. The Short tunnel was preferentially used by •NO to

reach the distal heme pocket. This preference is ascribed to its hydrophobic funnel-shape

entrance, covering a large area extending far from the tunnel entrance. This funnel-shape

entrance triggers the frequent formation of solvent-excluded cavities capable of hosting up to

three •NO molecules, thereby accelerating •NO capture and entry. The importance of

hydrophobicity of entrances for •NO capture is highlighted by a comparison with a polar mutant

for which residues at entrances were mutated with polar residues. A complete map of •NO

diffusion pathways inside TrHbN matrix was calculated, and •NO molecules were found to

diffuse from Xe cavity to Xe cavity. This scheme was in perfect agreement with the three-

133

dimensional free-energy distribution calculated using implicit ligand sampling. The trajectories

showed that •NO significantly alters the dynamics of the key amino acids of Phe62(E15), a

residue proposed to act as a gate controlling ligand traffic inside the Long tunnel, and also of

Ile119(H11), at the entrance of the Short tunnel. It is noteworthy that •NO diffusion inside

TrHbN tunnels is much faster than that reported previously for myoglobin. The results presented

in this work shed light on the diffusion mechanism of apolar gaseous substrates inside protein

matrix.

7.3. Introduction


inactivating key enzymes such as the terminal respiratory oxidases (1–5) and the iron/sulfur

protein aconitase (6,7). •NO also combines at near-diffusion-limited rate, with superoxide

produced by respiring cells to form the highly oxidizing agent peroxynitrite (8,9). •NO-


The truncated hemoglobin N (TrHbN) from Mycobacterium tuberculosis (Mtb) is thought to play

pivotal roles in the cellular metabolism of this organism during stress and hypoxia. TrHbN is

expressed during the stationary phase of Mycobacterium bovis BCG (10) and Mtb H37Ra (11).

In Mtb H37Ra, the activity of the glbN gene encoding TrHbN is upregulated by the general

nitrosative stress inducer, nitrite, by the •NO releaser sodium nitroprusside and by hypoxia. The

activity of the glbN gene is also enhanced during Mtb H37Ra invasion of THP-1 activated

macrophages (producing •NO) (11). Recent studies by our laboratory indicated that TrHbN has a

potent ability to detoxify •NO to nitrate (nitric oxide dioxygenase (NOD) reaction) and to protect

aerobic respiration from the inhibition by •NO in stationary phase cells of M. bovis BCG (10).

The high rate of •NO oxidation (k’NOD ≈ 745 µM-1

•s-1

at 23°C) catalyzed by oxygenated TrHbN

suggests that dioxygenation of •NO may be one of the vital defense systems in Mtb for coping

with the toxic effects of •NO, and may be important for allowing the intracellular survival of the

bacterium in macrophages. This hypothesis is also supported by the observation that expression

of TrHbN in a •NO-sensitive mutant of Salmonella enterica enhances the survival of the mutant

under nitrosative stress conditions and during growth within macrophages (12).

134

The oxygenated TrHbN crystal structure (13) presents two tunnels connecting the distal heme

pocket to the bulk solvent: the Short tunnel (ST) and the Long tunnel (LT). Molecular dynamics

(MD) simulations (14) revealed two additional tunnels: EH (EHT) and BE (BET) (Fig. 7.1, top).

Previous MD simulations suggested that gaseous substrates such as O2, CO, and •NO would

access the active site through these apolar tunnels, which are by nature composed of short-lived

cavities of various shapes and volumes (14). We also proposed that the character of these tunnels

makes them suitable to store ligands accessible to the active site. A similar proposition was

invoked to account for the bimodal solvent phase recombination of CO with the truncated Hb

from Paramecium caudatum (15). Finally, the configuration of the active site allows the bound

O2 to remain optimally oriented and stabilized by Tyr33(B10) and Gln58(E11) for reaction with

•NO (10,13,14,16,17). These singular TrHbN characteristics are believed to generate the

unequalled NOD rate constant, which is 15-fold and 34-fold faster than that of horse-heart Mb

and sperm-whale Mb, respectively (10,18).

Previous work using steered MD simulations aimed at understanding O2 and •NO diffusion in

TrHbN (19). The results suggested that O2 reaches the active site of deoxy-TrHbN through the

ST while •NO accesses bound O2 through the LT (referred as the dual-path mechanism).

However, due to the limitations of the method (a biasing force drives ligand diffusion along

predefined coordinates), important aspects of ligand diffusion before the actual NOD reaction

were not addressed in this study. How ligands interact with TrHbN surface, how they enter the

protein matrix, and how they diffuse to the active site are events that may influence the NOD

reaction. In this work, we address these specific issues using unbiased explicit MD simulations

where •NO molecules were initially placed in the bulk solvent. In addition, we present a detailed

description of •NO diffusion processes along the different pathways, as well as the free-energy

cost associated with •NO diffusion inside the different TrHbN channels. Two major findings

arise from this study: the hydrophobic nature of entrances is responsible for •NO capture before

diffusion to DHP; and the dual-path mechanism is refuted.

135

7.4. Methods

Molecular dynamics (MD) simulations MD simulations were performed using CHARMM

software (20) with the CHARMM22 all-atom potential energy parameter set (21) and O2-bound

heme force-field parameters (14). Simulations were performed as described previously (14). A

complete description of the simulation protocol is given as Supporting Material enclosed in

Annexe 2. A total of 260 ns of trajectories were produced: two trajectories of 30 ns each without

free •NO molecule and 10 trajectories of 20 ns each including 10 free •NO molecules. An

additional 20-ns MD trajectory of a TrHbN multiple polar mutant was carried in absence of •NO.

Mutations were designed to increase polarity of tunnel entrances without filling cavities (this

design is discussed in the Supporting Material enclose in Annexe 2). The mutations consisted in

the replacement of hydrophobic amino acids located at tunnel entrances by polar residues:

Ala95(G9)Ser and Ala120(H12)Asp for ST, and Ala24(B1)Ser for LT, and Val118(H10)Asp for

EHT. The trajectory was calculated using NAMD (22). Only the last 15 ns of the trajectory were

used for analysis.

7.4.1. Analysis

To study •NO interactions with TrHbN surface, the time of contact between •NO molecules and

all TrHbN atoms was calculated using whole trajectories and a distance cutoff of 3.5 Å. The

solvent-excluded accessible volumes at the tunnel entrances were calculated using the VOIDOO

software with a probe radius of 1.4 Å (23). The probability density (or probability distribution)

of •NO at the surface and inside TrHbN was determined using all •NO coordinates along the last

15 ns of the 10 simulations performed, giving a total of 150,000 sets of coordinates. A distance

cutoff of 3 Å from the •NO center-of-mass was taken for the calculation of •NO probability

density.

136

7.4.2. Implicit ligand sampling

The ligand potential of mean force (PMF) inside TrHbN was determined using the implicit

ligand sampling (ILS) method (24) included in the open source VMD 1.8.6 software package

(25). The errors on PMF values were estimated as described in Cohen et al. (24). A drawback of

the method concerns the overestimation of energy barriers due to the absence of the ligand (24).

For each ILS calculation, five sets of 5000 MD frames covering entire equilibrium trajectories of

TrHbN in absence of free ligand were used. Details about ligand parameters used with this

method are given as Supporting Material (Annexe 2).

The impacts of free •NO molecule on the residues lining the tunnels were studied. The rotameric

distribution of each residue was collected as function of the presence or absence of •NO inside

respective internal cavities (Xe1, Xe2, Xe3, Xe5, and EHc). For the residues that displayed a

different behavior in response to the presence of a •NO molecule, individual ILS computations

were performed using a subset of 5000 representative MD snapshots (e.g., according to different

rotamers). To increase sampling, MD snapshots were selected from simulations of oxy-TrHbN

with and without free NO molecules. For the snapshots taken from simulations with explicit

•NO, only the part of the trajectory before the first •NO entry inside the protein core was

considered. If >5000 MD representative snapshots were available, a random selection was

performed. The binding affinities (Kb) of the ligands were estimated from the PMF maps

obtained from the ILS calculations using the relation (26,27)

β ( ) ( ')w r w r

bK dre

(1)

where r represents the positions of the grid maps, β = 1/kBT, w the PMF between the ligand and

the protein, and r’ is a reference position far away in the bulk. In our calculations, w(r’) was

calculated as the average PMF of the ligand in a water box (more details in the Supporting

Material enclose d in Annexe 2). Unless otherwise noted, all affinity numbers given in this study

refer to Kd = 1/Kb.

137


•NO interacts with specific regions of the TrHbN surface. The times of contact between •NO

molecules and TrHbN surface are represented in Fig. 7.2. This figure shows that tunnel

entryways constitute discrete domains with enhanced affinity for •NO due to the hydrophobic

nature of the residues at these positions. •NO capture is promoted by formation of short-lived

hydrophobic cavities at the surface of TrHbN, creating free solvent volumes, as presented in Fig.

7.3 for ST entrance. The solvent-excluded volumes arise from the energy cost related to the

hydrophobic hydration (28), the excluded volume size being in relation with the enriched content

of hydrophobic residues at the tunnel entrances. The favorable interactions between •NO and the

hydrophobic cavities are also related to the hydrophobic hydration, leading to the association of

apolar entities (28). The enhanced affinity observed at tunnel entrances is thought to be important

for the capture of apolar ligands. Consequently, it is expected that entrances delimited by polar

residues may result in lower affinities for apolar ligands. To verify this, a mutant was designed

where the polarity near the ST, LT, and EHT entryways was increased with polar residues (see

Methods).

The formation frequency and average solvent-excluded volumes are given in Table 7.1 for each

of the entrances, for both the WT and the polar mutant. The biggest solvent-excluded volumes

(average volume of 89 Å3) were observed at the ST entrance of WT in 43% of the MD snapshots

(Table 7.1), providing enough room to host up to three •NO molecules. Smaller solvent-excluded

volumes were also observed at the entrance of LT, EHT, and BET, although with a lower

incidence. Because of the presence of polar residues located at the entrances, solvent-excluded

volumes are much lower and less frequent for the ST entrance of the polar mutant. The effect is

less significant for LT and EHT entrances due to an increased side-chain dynamics experienced

by the mutated residues in the solvent, similar to what was reported in Leroux et al. (29).

The particular funnel shape of the ST entrance (Fig. 7.3 c), combined with its hydrophobic

nature, is responsible for higher free volumes than for other tunnel entrances. To our knowledge,

this is the first time that such physical characteristics at the surface of a protein are reported. We

hypothesize that this physical property contributes to the high efficiency of the NOD reaction

catalyzed by TrHbN.

138

7.5.1. •NO enters the protein matrix using the ST, LT, and EHT

•NO enters the protein matrix using the ST, LT, and EHT At least one •NO molecule reached the

distal heme pocket (DHP) from the bulk solvent in each of the 10 simulations performed. The

first •NO entry typically occurred within the first 3 ns. In agreement with the particular character

of the ST entrance, the first •NO to reach the DHP came from the ST in six of ten occurrences;

the LT and the EHT were used twice each. The time required to diffuse from the protein surface

to the DHP ranged from hundreds of picoseconds to several nanoseconds, the shorter times being

associated with the ST. The BET was not used although several short-lived contacts between

•NO molecules and the BET entryway were observed, likely due to the narrow bottleneck (<1 Å)

of this tunnel (14).

The greater propensity of •NO to use ST to reach DHP observed in this work is clearly in

disagreement with previous modeling studies, where the authors reached the conclusion that the

diffusion of •NO occurs preferentially via LT in oxy-TrHbN (19). This disagreement will be

addressed in the next section.

Ruscio et al. used a very similar approach to study CO entry and diffusion within Mb (30). In

contrast to what was observed in this article, a CO molecule reached the active site in only one-

third of the 48 90-ns trajectories. In the other cases, CO entered Mb matrix but could not reach

the active site. This comparison highlights the higher efficiency of substrate diffusion from the

solvent to the DHP in the TrHbN matrix than in Mb. This higher efficiency is in agreement with

the 15-fold (horse-heart Mb) and 34-fold (sperm-whale Mb) higher bimolecular rate constant for

the NOD reaction (10,18).

7.5.2. Diffusion through tunnels

MD simulations showed that tunnels in TrHbN are not open channels but are formed of

neighboring hydrophobic cavities that are temporally interconnected due to side-chain flexibility

139

(thermal fluctuations) (14,31). Therefore, it is expected that ligand diffusion through the different

tunnels takes place by hopping from one cavity to another.

Occupancy of •NO in the protein matrix was calculated using two approaches: ILS (24), using

trajectories without •NO, and the probability distributions from the trajectories with •NO (see

Methods). The results are presented in Fig. 7.4, a and b, for the probability distributions and ILS,

respectively. Interestingly, both methods show that the TrHbN matrix contains cavities with high

occupancy for •NO, and that most of these cavities correspond to the Xe binding sites identified

by x-ray crystallography (32). In addition to Xe binding sites, the distal heme pocket and a cavity

lying inside the EHT were found favorable to •NO (Fig. 7.4). The latter cavity is located between

the EHT entrance and the side chain of the Phe62(E15) residue. The ILS isosurface maps (Fig.

7.4 b) confirmed that ST, LT, and EHT are the most favorable routes. Fig. 7.1 (bottom) shows all

diffusion pathways observed from the trajectories calculated using explicit •NO. Every •NO that

reached the DHP was found to transit through the Xe2 site, which is located at the intersection of

the ST, LT, and EHT.

The possibility that •NO molecules have local impacts on TrHbN, and therefore influence its

diffusion through tunnels, was investigated. Because TrHbN matrix can hold more than one •NO

molecule at the same time, and as it is expected that substrate concentrations are relatively low in

vivo, this specific study focused on the simpler cases, i.e., when only one •NO was inside

TrHbN. To determine the effect of local impacts, the side-chain dihedral angles of all the

residues lining the tunnels were monitored in presence or absence of •NO near that residue.

These residues, as well as the Xe binding sites, are sketched in Fig. 7.1 (bottom). Only two

residues were influenced by the presence of •NO: Ile119(H11), located along the ST, between

Xe2 and Xe3 sites and shaping part of the EHT; and Phe62(E15), located along the LT. The

rotamers adopted in the different cases are reported in Table 7.2.

In the next subsections, the diffusion process through the different tunnels is analyzed separately.

The PMF profiles for each tunnel are presented in Fig. 7.5. Movie S1, Movie S2, and Movie S3,

illustrating •NO diffusion in each of the three tunnels, extracted from the trajectories, are

available at http://www.cell.com/biophysj/supplemental/S0006-3495(09)01450-7.

140

Short tunnel - Diffusion through the ST begins by docking to the surface (Fig. 7.5 a, at ~14 Å

with a value of 3.5±0.3 kcal/mol), in the funnel-shape entryway, close to Xe3. Entry is

performed by transit from Xe3 to Xe2. For this process, a •NO molecule must cross a bottleneck

region (radius of 1.2 Å) defined by the side chain of Ile119(H11), Leu98(G12), and the heme

pyrrole B (vinyl group) (14). MD simulations in absence of free •NO revealed that the

Ile119(H11) side chain adopts four rotamers (mm, mt, pp, and pt) with various distributions

(Table 7.2). The pp and pt rotamers are rarely found inside a-helices, whereas the mm and mt

rotamers are more typical (33). In the absence of •NO, mm and pt rotamers dominate with 56%

and 37% of the conformations, respectively. When a •NO is docked inside Xe2 or Xe3 (ST

entrance), the mm and mt rotamers increase significantly (Table 7.2). Fig. 7.5 b shows the PMF

profiles according to the Ile119(H11) rotamers. The mt and mm rotamers are found to be the

most favorable for •NO diffusion inside ST, from Xe3 to Xe2, with energy barriers of

2.7±0.6 kcal/mol and 3.5±0.6 kcal/mol, respectively, centered at 9 Å. This result is in agreement

with the TrHbN crystallographic structure under high xenon pressure, where a xenon atom

occupies Xe3 while Ile119(H11) displays the mt rotamer (PDB ID No. 1S56, B chain) (32). On

the other hand, pt and pp rotamers are unfavorable, causing a high free energy barrier of >5 -

0.8/+0.6 kcal/mol. These unfavorable free energy profiles agree with the observation of explicit

•NO molecules docked for several hundreds of picoseconds while Ile119(H11) adopted pt and pp

rotamers. Interestingly, when a •NO molecule docks inside Xe1 (in the LT, Fig. 7.1, top), the

occurrence of mm rotamers decreases, accounting for 18% of the conformations, whereas it

increases to 72% for pt rotamer (Table 7.2). This has the effect of significantly reducing

diffusion through ST.

Long tunnel - Entry into the LT is performed by transition from the solvent to the Xe1 binding

site. This process is unhampered and fast because of the wide LT entrance. This is in agreement

with our earlier report on TrHbN tunnels dynamics, where the LT was found to display the

widest entrance with an average aperture of ~1.8 Å radius (14).

To reach Xe2 from Xe1, a •NO molecule must transit by the Xe5 binding site (Fig. 7.1, bottom).

However, the tunnel radius at the Xe5 binding site is quite narrow due to the side chains of

Val29(B6), Phe62(E15), and Leu98(G12) residues so that rotation of the Phe62(E15) phenyl ring

(fluctuation of the χ2 dihedral angle) is required. Before the arrival of •NO, as observed in a

141

previous work (14), the Phe62(E15) side chain explores two χ1 domains (sweeping motion)

while a wide range of values is allowed for χ2. The two χ1 domains were referred as the M and T

states. The M state regroups two rotameric species, m-85 and m-30, whereas the T state is mainly

characterized by the t80 rotamer. In this rotameric nomenclature, m stands for gauche negative

and t for trans (33). The phenyl ring fills Xe1 when in T state, while for the M state, it is located

between Xe5, Xe2, and EHc cavities. The radius of the LT (and EHT), and consequently their

opening, is determined by χ2 (see Fig. 7.9 of (14).).

The impact of •NO on the side-chain dynamics of Phe62(E15) is represented in Fig. 7.6 and the

rotamers observed for the different cases are reported in Table 7.2. When a •NO molecule

occupies Xe1, the Phe62(E15) side chain is almost limited to the M state (98% of occupancy)

and χ2 is restricted to a narrower range: [-90°:-60°] and [80°:90°] (χ2 ranges from -90° to 90° by

convention) (Fig. 7.6). This range corresponds to the maximum radii for LT (~1.3 Å), while EHT

radius is at a minimum (~1.0 Å) (14). Therefore, LT would be in the open state with this

configuration, in agreement with the observation of the passage of one •NO to Xe5 (more on this

in the next paragraph). Similarly, when a •NO occupies Xe5, the Phe62(E15) side chain is almost

limited to the T state (86.3% of occupancy, the T state has a broader definition than the t80

rotamer, which has a limited χ2 range), and χ2 is restricted to the range [+20°:+75°]. This range

of conformations corresponds to a radius between 0.9 Å and 1.1Å, similar to the BET average

radius where no diffusion was observed. In fact, no diffusion of •NO from Xe5 to Xe1 was

observed via the Phe62(E15) barrier. We conclude that LT was closed in this conformation.

Instead, •NO can escape via EHT to Xe1 (more on this in the next subsection).

Because •NO has significant effects on Phe62(E15) dynamics, PMF profiles were calculated

when Phe62(E15) is either in the M or the T state. The PMF profiles are given in Fig. 7.5 c. For

both M and T states, the entry is found favorable for docking of •NO, as reported in a previous

section, with a docking free energy of -2.7±0.3 kcal/mol. For the M state (Fig. 7.5 c, solid line),

the PMF profile revealed that the Xe1 and Xe2 binding sites are favorable to •NO, with docking

free energies of -3.7±0.3 kcal/mol and -3.6±0.3 kcal/mol, respectively. This is reflected in the

high probability of finding •NO at these positions in explicit simulations (Fig. 7.4 a). To diffuse

from Xe1 to Xe2 in the LT, a •NO molecule must overcome a free energy barrier of 3.2±0.6

kcal/mol, indicating that •NO molecules seldom cross the Phe62(E15) side chain. Only one event

142

was observed over the 200 ns of trajectories. This energy barrier comes from the steric

encumbrance of Phe62(E15), which has to change its conformation at the time of the transit

between Xe1 and Xe5 to accommodate a •NO at Xe5. Many events were observed where a •NO

tried to transit from Xe1 to Xe5 but came back to Xe1. For the T state, the PMF profile is

inverted, showing a maximum free energy level in the tunnel of -1.7±0.3 kcal/mol near Xe1

while the most favorable region is found between Xe5 and Xe2, with a potential well of -4.0±0.3

kcal/mol. Even if the free energy barrier is lower by 2.1±0.6 kcal/mol, this path was not used by

•NO because Phe62(E15) quickly returns to M state as a •NO enters the LT. Due to the high

energy barriers involved for the path Xe1-Xe5-Xe2, the path via EHT (Xe1-EHC-Xe2) was used

more frequently. This path is presented in the next subsection.

EH tunnel - As for the LT, two PMF profiles were calculated according to the Phe62(E15)

conformation. The PMF profiles are given in Fig. 7.5 d. For this tunnel, •NO diffuses from the

bulk solvent to an internal cavity (EHc) (Fig. 7.1, bottom), which does not correspond to any

identified Xe binding site. This cavity is located between the EHT entrance and the side chain of

the Phe62(E15) residue. To reach EHc from the bulk, a •NO must cross a bottleneck region of

1.3 Å radius at the protein surface.

As for the LT, EHT shows two different PMF profiles according to the Phe62(E15)

conformation. For both PMF profiles, the first few Å between the protein surface and the EHc do

not differ (Fig. 7.5 d). The EHc was found favorable to •NO, with a free energy of -3.1±0.3

kcal/mol (≈-4.5±0.3 kcal/mol from the solvent), and no escape of •NO to the solvent was

observed. When EHc is filled with a •NO, the Phe62(E15) side chain fluctuates between M and T

states, with a prevalence for the M state. When Phe62(E15) is in T state, there is no barrier

between EHc and Xe2, and frequent exchanges between these sites were observed from the

trajectories (Fig. 7.1, bottom, blue arrow). In contrast, when Phe62(E15) is in the M state, the

PMF profile shows an energy barrier of 4.0±0.6 kcal/mol. In this case, frequent exchanges

between Xe1 (located in LT) and EHc are observed (Fig. 1, bottom, yellow arrow), although a

small free energy barrier of 1.5±0.6 kcal/mol (relative to EHc) exists between the two sites (data

not shown).

143

DHP - Once a •NO reaches the DHP, it remains at contact distance with the bound O2 and

Val94(G8) side chain for up to hundreds of picoseconds before returning to Xe2. Since the

energy barrier between these two sites is low (0.8±0.6 kcal/mol), frequent Xe2 ↔ DHP

transitions were observed. The presence of •NO in the DHP does not trigger motions or

reorganization of nearby side chains.

7.5.3. Ligand binding affinities

The binding affinities for •NO and O2 ligands were estimated from the PMF maps obtained from

the ILS calculations (see Methods). The affinities were calculated for the whole protein (global

affinities) and for the entrances at the surface of the protein, for both WT and the polar mutant.

The results are presented in Table 7.1. WT global affinities for •NO and O2 were evaluated as

6.20±0.04 mM and 8.72±0.04 mM, respectively. It is noteworthy that the high precision

estimated for the ligand affinities arises from the methodology used and does not account for the

error due to the imprecision of the force field and the tendency of MD simulations to

undersample. Under identical experimental conditions, •NO affinity may always be 25–75%

greater than for O2 for the same conditions due to the chemical nature of the ligands, the Van der

Waals interactions being more significant for •NO. It is expected that the affinities of free

ligands should be similar for deoxy-TrHbN. The affinity of the heme-iron for the ligands was not

calculated in this work. However, a much larger affinity for O2 arises upon binding to the heme-

iron. A change in its chemical character upon binding leads to a stabilization of the Fe-O2

complex by Tyr33(B10) (10). In vivo, ligand concentration in activated macrophages are

believed to be low (~1 mM) (34), suggesting that tunnels may be free of ligand at equilibrium.

However, such affinity increases the chances by >73 times to find a •NO molecule bound to

TrHbN compared to the solvent.

As expected from times of contact (Fig. 2), •NO occupancies and ILS results (Fig. 4), the

affinities presented in Table 7.1 also reveal that the entrances are favorable to •NO (see also

Annexe 2, Fig. S1). Quantitatively, ST entrance has the highest affinity (Kb), accounting for as

much as the affinities of LT and EHT taken together. The gradient of affinity of ST entrance

144

extends far from the tunnel entrance and covers a large area (Annexe 2, Fig. S1). This large area

may play an important role by trapping •NO molecule and thus increasing the use of ST.

Binding affinities for the WT and the polar mutant are compared in Table 7.1. The global affinity

of the mutant for •NO is only slightly lower than for the WT because the tunnels, unchanged in

the mutant, account for ~50% of the global affinity. The biggest differences are observed for the

ST entrance, where the affinity of the mutant is >13 times lower than for the WT (67.6±1.1 mM

for the WT, and 926±33 mM for the mutant). Since the affinity of the ST entrance of the mutant

is greatly reduced we predict that ST should be almost unused by apolar ligands to reach the

DHP.

On the other side, the affinity of LT entrance was modestly lowered while EHT entrance affinity

was mostly unchanged. The limited decrease of affinity observed for the LT entrance is because

the entrance is defined by backbone residues from AB and GH hinges. As a consequence,

substitution of a side chain can only partially affect the character of the entrance.

Similarly, the unchanged affinity of the mutant EHT entrance is due to the Asp118(H10) side

chain that moved toward the solvent, triggering enlargement of the tunnel entrance. The move

toward the solvent of a mutated residue at the entrance of a tunnel, leading to a limited change in

protein kinetics, was also observed for a NiFe hydrogenase (29).

Despite a dramatic ST entrance affinity decrease observed for the mutant, we predict only a

twofold reduced NOD activity. Effectively, the almost unchanged affinities of the LT and EHT

entrances, accounting for as much as the ST entrance of the WT, would continue to supply the

active site with apolar ligands. This example illustrates the complexity to plan efficient

mutagenesis of a protein containing multiple tunnels, as in Johnson et al. (35).

7.5.4. Comparison of the different paths

The free energy difference between the solvent and the DHP is ~4.5±0.3 kcal/mol in favor of

DHP. When the tunnels are compared between each other, the ST has the biggest

145

solvent-excluded volume at its entrance, the lowest energy barrier, and also the shortest distance

to reach the DHP. Therefore, it is natural that this tunnel is the most used by •NO to reach DHP,

as observed from explicit simulations. However, one should not neglect the LT/EHT in their

capacity to supply ligands to the active site. Effectively, their numerous binding sites can

accommodate up to three ligands, which can reach DHP at the appropriate time.

As in our previous study of TrHbN tunnels (14), the results presented here disagree with the

results reported by Bidon-Chanal et al. (19). Effectively, using steered MD , they came to the

conclusion that LT is the most favorable diffusion pathway for •NO to reach DHP. In this study,

the high energy barrier near the Phe62(E15) side chain explains why LT is rarely used by •NO to

reach DHP. Bidon-Chanal et al. did not take into account the more favorable EHT pathway for

•NO diffusion to the DHP. In addition, Bidon-Chanal et al. observed a progressive rise in free

energy for the migration of •NO through ST, making it less suitable than LT (surprisingly, their

PMF values for ST near the DHP indicate that this position is less favorable to •NO than in the

solvent). This result is clearly in disagreement with the results obtained using two different

approaches (explicit MD simulations and ILS) in this work. Moreover, our MD simulations with

explicit NO molecules clearly showed that LT and ST merge at the Xe2 binding site with the

same binding energy, which is not the case in the work of Bidon-Chanal et al. Both MD and ILS

approaches identified the ST as the most favorable route for •NO to reach oxy-TrHbN DHP. This

is explained by the presence of the funnel-shape hydrophobic entrance of the ST, the shorter

diffusion length, and the lowest energy barrier experienced by •NO.

It is interesting to note that the diffusion of •NO from the solvent to EHc, Xe1, or Xe2 involves a

decrease in free energy of ~4.5 ± 0.3 kcal/mol. Such an energy value ensures efficient capture of

•NO in TrHbN matrix, preventing the release to the solvent. Over the 200 ns of trajectories, only

two such events involving a single •NO was observed and proceeded through the ST (events

involving multiple •NO, energetically more complex, were observed but will not be reported in

this study). We propose that this efficient capture of •NO by TrHbN is related to its high NOD

catalytic.

There are few other resolved group I TrHb structures from the protozoan Paramecium caudatum

(pc-TrHbN–PDB 1DLW), the unicellular alga Chlamydomonas eugametos (ce-TrHbN–PDB

146

1DLY), and the cyanobacterium Synechocystis sp. (ss-TrHbN–PDB 1S69). All these structures

display significant internal empty volumes (pc-TrHbN: 180 Å3, ce-TrHbN: 400 Å

3, and mt-

TrHbN: 265 Å3) defined by hydrophobic residues (32). The hydrophobic character of these

residues is conserved (36). On the other hand, the hydrophobic nature of the residues lining the

tunnel entrances is not as well conserved between the different TrHbNs. The free volume present

in ce-TrHbNs structure is organized into two tunnels corresponding to the LT and EHT. In

addition, ce-TrHbN accommodates a large hydrophobic funnel-shaped surface at the ST

entrance, suggesting that this tunnel may be functional. Three additional diffusion routes were

found from modeling studies of pc-TrHbN: between helices B and G; between helices B and E;

and between the heme and helix C (37). TrHbs from groups II and III do not display tunnels.

These observations suggest that the redundancy in access ways in TrHbN is an evolved feature

of the group I TrHb.

Finally, during the review process, an article bearing on •NO diffusion in TrHbN was published

by Mishra and Meuwly (38). This study is based on multiple 2-ns MD simulations of TrHbN

with one •NO starting in different cavities (total of 24 2-ns trajectories). Some important

differences arise from both works. First, unlike our work, they observed many escapes of •NO to

the solvent and rare •NO captures from TrHbN, which is surprising given the energy difference

of ~4.5 kcal/mol in favor of the protein matrix observed in this work. Second, they observed •NO

diffusion from Xe2 to the solvent, through helices C and H, and did not observe diffusion

through the EHT. Finally, the heme force field parameters are not convenient for simulations of

oxygenated hemoprotein, favoring unlikely conformation of DHP residues as well as

Phe62(E15) (14). Substantial methodological differences between the works could explain these

discrepancies.

7.6. Conclusion

The simulations presented here show that substrate diffusion in TrHbN follows discrete paths

showing enhanced affinity for apolar gaseous ligands. Diffusion toward the TrHbN active site

(DHP) begins with favorable hydrophobic interactions at the protein surface, corresponding to

147

tunnel entryways. The strongly hydrophobic funnel-shape entrance of ST makes it the most

favorable tunnel in TrHbN. This funnel constitutes a discrete surface domain with enhanced

affinity for •NO. •NO capture is promoted by formation of short-lived hydrophobic cavities at the

protein surface, creating free solvent volumes sufficient to host up to three •NO molecules

simultaneously. These molecules, once captured, do not escape and enter the ST. Once inside the

protein, •NO diffuses from one cavity to another. Most of these cavities correspond to

experimental xenon binding pockets (32).

When docked inside specific cavities (Xe1, Xe2, Xe3, and Xe5), •NO was found to affect the

dynamics of Phe62(E15) and Ile119(H11) residues in LT and ST, respectively. In contrast to

what was observed in Bidon-Chanal et al. (19), •NO diffusion in the LT was found to be

hindered by the Phe62(E15) side-chain obstruction, refuting the dual-path mechanism proposed

by Bidon-Chanal et al. (19). Moreover, •NO entering the LT preferentially bypasses the

Phe62(E15) barrier by passing through EHT. Our results suggest that these paths would be better

suited to provide substrates to the active site once the reaction cycles have begun.

Our MD simulations suggest that •NO diffusion between the TrHbN surface and the active site is

fast. In all 10 simulations performed, a •NO molecule reached the distal heme pocket within 20

ns. This contrasts with a similar earlier work using Mb that required more simulations and longer

trajectories to study a similar process (30). This is in accord with the fact that the bimolecular

rate constant for the NOD reaction catalyzed by TrHbN is ~15-fold higher than that of horse

myoglobin (10).

148

7.7. Supporting material

Three movies, one figure, and two tables are available at

http://www.biophysj.org/biophysj/supplemental/S0006-3495(09)01450-7.

The authors thank Beatrice A. Wittenberg and Jonathan B. Wittenberg for helpful discussion.

This work was supported by the Natural Sciences and Engineering Research Council of Canada

(grant No. 46306-01 (2005–2010)), the Fonds Québécois de la Recherche sur la Nature et les

Technologies (grant No. 104897), and the Canada Foundation for Innovation (grant No. 12428).

R.D. is supported by a postgraduate scholarship from the Fonds Québécois de la Recherche sur

la Nature et les Technologies (scholarship No. 106627).

7.8. References

1. Brunori, M. 2001. Nitric oxide, cytochrome-c oxidase and myoglobin. Trends Biochem. Sci.

26:21–23.

2. Brunori, M., A. Giuffre, P. Sarti, G. Stubauer, and M. T. Wilson. 1999. Nitric oxide and

cellular respiration. Cell. Mol. Life Sci. 56:549–557.

3. Stevanin, T. M., N. Ioannidis, C. E. Mills, S. O. Kim, M. N. Hughes, et al. 2000.

Flavohemoglobin hmp affords inducible protection for Escherichia coli respiration, catalyzed

by cytochromes bo’ or bd, from nitric oxide. J. Biol. Chem. 275:35868–35875.

4. Cleeter, M. W. J., J. M. Cooper, V. M. Darleyusmar, S. Moncada, and A. H. V. Schapira.

1994. Reversible inhibition of cytochrome-c oxidase, the terminal enzyme of the

mitochondrial respiratory-chain, by nitric-oxide - implications for the neurodegenerative

diseases. FEBS Lett. 345:50–54.

5. Brown, G. C., and C. E. Cooper. 1994. Nanomolar concentrations of nitric-oxide reversibly

inhibit synaptosomal respiration by competing with oxygen at cytochrome-oxidase. FEBS

Lett. 356:295–298.

6. Gardner, P. R., G. Costantino, and A. L. Salzman. 1998. Constitutive and adaptive

detoxification of nitric oxide in Escherichia coli - role of nitric-oxide dioxygenase in the

protection of aconitase. J. Biol. Chem. 273:26528–26533.

7. Gardner, P. R., G. Costantino, C. Szabo, and A. L. Salzman. 1997. Nitric oxide sensitivity of

the aconitases. J. Biol. Chem. 272:25071–25076.

http://www.biophysj.org/biophysj/supplemental/S0006-3495(09)01450-7

149

8. Kissner, R., T. Nauser, P. Bugnon, P. G. Lye, and W. H. Koppenol. 1997. Formation and

properties of peroxynitrite as studied by laser flash photolysis, high-pressure stopped-flow

technique, and pulse radiolysis. Chem. Res. Toxicol. 10:1285–1292.

9. Pfeiffer, S., A. C. F. Gorren, K. Schmidt, E. R. Werner, B. Hansert, et al. 1997. Metabolic

fate of peroxynitrite in aqueous solution - reaction with nitric oxide and pH-dependent

decomposition to nitrite and oxygen in a 2:1 stoichiometry. J. Biol. Chem. 272:3465–3470.

10. Ouellet, H., Y. Ouellet, C. Richard, M. Labarre, B. Wittenberg, et al. 2002. Truncated

hemoglobin HbN protects Mycobacterium bovis from nitric oxide. Proc. Natl. Acad. Sci.

USA. 99:5902–5907.

11. Pawaria, S., A. Lama, M. Raje, and K. L. Dikshit. 2008. Responses of Mycobacterium

tuberculosis hemoglobin promoters to in vitro and in vivo growth conditions. Appl. Environ.

Microbiol. 74: 3512–3522.

12. Pawaria, S., G. Rajamohan, V. Gambhir, A. Lama, G. C. Varshney, et al. 2007. Intracellular

growth and survival of Salmonella enterica serovar Typhimurium carrying truncated

hemoglobins of Mycobacterium tuberculosis. Microb. Pathog. 42:119–128.

13. Milani, M., A. Pesce, Y. Ouellet, P. Ascenzi, M. Guertin, et al. 2001. Mycobacterium

tuberculosis hemoglobin N displays a protein tunnel suited for O2 diffusion to the heme.

EMBO J. 20:3902–3909.

14. Daigle, R., M. Guertin, and P. Lague. 2009. Structural characterization of the tunnels of

Mycobacterium tuberculosis truncated hemoglobin N from molecular dynamics simulations.

Proteins: Struct. Funct. Bioinf. 75:735–747.

15. Samuni, U., D. Dantsker, A. Ray, J. B. Wittenberg, B. A. Wittenberg, et al. 2003. Kinetic

modulation in carbonmonoxy derivatives of truncated hemoglobins: the role of distal heme

pocket residues and extended apolar tunnel. J. Biol. Chem. 278:27241–27250.

16. Crespo, A., M. A. Marti, S. G. Kalko, A. Morreale, M. Orozco, et al. 2005. Theoretical study

of the truncated hemoglobin HbN: exploring the molecular basis of the NO detoxification

mechanism. J. Am. Chem. Soc. 127:4433–4444.

17. Ouellet, Y., M. Milani, M. Couture, M. Bolognesi, and M. Guertin. 2006. Ligand interactions

in the distal heme pocket of Mycobacterium tuberculosis truncated hemoglobin N: roles of

TyrB10 and GlnE11 residues. Biochemistry. 45:8770–8781.

18. Scott, E. E., Q. H. Gibson, and J. S. Olson. 2001. Mapping the pathways for O2 entry into

and exit from myoglobin. J. Biol. Chem. 276:5177–5188.

150

19. Bidon-Chanal, A., M. A. Marti, A. Crespo, M. Milani, M. Orozco, et al. 2006. Ligand-

induced dynamical regulation of NO conversion in Mycobacterium tuberculosis truncated

hemoglobin-N. Proteins. 64:457–464.

20. Brooks, B. R., R. E. Bruccoleri, B. D. Olafson, D. J. States, S. Swaminathan, et al. 1983.

CHARMM: a program for macromolecular energy, minimization, and dynamics calculations.

J. Comput. Chem. 4:187–217.

21. MacKerell, A. D., D. Bashford, M. Bellott, R. L. Dunbrack, J. D. Evanseck, et al. 1998. All-

atom empirical potential for molecular modeling and dynamics studies of proteins. J. Phys.

Chem. B. 102:3586–3616.

22. Phillips, J. C., R. Braun, W. Wang, J. Gumbart, E. Tajkhorshid, et al. 2005. Scalable

molecular dynamics with NAMD. J. Comput. Chem.26:1781–1802.

23. Kleywegt, G. J., and T. A. Jones. 1994. Detection, delineation, measurement and display of

cavities in macromolecular structures. Acta Crystallogr. D Biol. Crystallogr. 50:178–185.

24. Cohen, J., A. Arkhipov, R. Braun, and K. Schulten. 2006. Imaging the migration pathways

for O2, CO, NO, and Xe inside myoglobin. Biophys. J. 91:1844–1857.25.

25. Humphrey, W., A. Dalke, and K. Schulten. 1996. VMD—visual molecular dynamics. J. Mol.

Graph. 14:33–38.

26. Chandler, D., and L. R. Pratt. 1976. Statistical mechanics of chemical equilibria and

intramolecular structures of nonrigid molecules in condensed phases. J. Chem. Phys.

65:2925–2940.

27. Roux, B., and T. Simonson. 1999. Implicit solvent models. Biophys.Chem. 78:1–20.

28. Hummer, G., S. Garde, A. E. Garcıa, A. Pohorille, and L. R. Pratt. 1996. An information

theory model of hydrophobic interactions. Proc. Natl. Acad. Sci. USA. 93:8951–8955.

29. Leroux, F., S. Dementin, B. Burlat, L. Cournac, A. Volbeda, et al. 2008. Experimental

approaches to kinetics of gas diffusion in hydrogenase. Proc. Natl. Acad. Sci. USA.

105:11188–11193.

30. Ruscio, J. Z., D. Kumar, M. Shukla, M. G. Prisant, T. M. Murali, et al. 2008. Atomic level

computational identification of ligand migration pathways between solvent and binding site

in myoglobin. Proc. Natl. Acad. Sci. USA. 105:9204–9209.

31. Saam, J., I. Ivanov, M. Walther, H.-G. Holzhutter, and H. Kuhn. 2007. Molecular dioxygen

enters the active site of 12/15-lipoxygenase via dynamic oxygen access channels. Proc. Natl.

Acad. Sci. USA. 104:13319–13324.

151

32. Milani, M., A. Pesce, Y. Ouellet, S. Dewilde, J. Friedman, et al. 2004. Heme-ligand

tunneling in group I truncated hemoglobins. J. Biol. Chem. 279:21520–21525.

33. Lovell, S. C., J. M. Word, J. S. Richardson, and D. C. Richardson. 2000. The penultimate

rotamer library. Proteins. 40:389–408.

34. Lewis, R. S., S. Tamir, S. R. Tannenbaum, and W. M. Deen. 1995. Kinetic analysis of the

fate of nitric oxide synthesized by macrophages in vitro. J. Biol. Chem. 270:29350–29355.

35. Johnson, B. J., J. Cohen, R. W. Welford, A. R. Pearson, K. Schulten, et al. 2007. Exploring

molecular oxygen pathways in Hansenula polymorpha copper-containing amine oxidase. J.

Biol. Chem. 282:17767–17776.

36. Vuletich, D. A., and J. T. J. Lecomte. 2006. A phylogenetic and structural analysis of

truncated hemoglobins. J. Mol. Evol. 62:196–210.

37. Cohen, J., K. W. Olsen, and K. Schulten. 2008. Finding gas migration pathways in proteins

using implicit ligand sampling. Methods Enzymol. 487:439–457.

38. Mishra, S., and M. Meuwly. 2009. Nitric oxide dynamics in truncated hemoglobin: docking

sites, migration pathways, and vibrational spectroscopy from molecular dynamics

simulations. Biophys. J. 96:2105–2108.

39. DeLano, W.L. 2002. The PyMOL Molecular Graphics System. http://www.pymol.org.

http://www.pymol.org/

152

Table 7.1 Calculated affinities for NO and solvent-excluded volume at tunnel entrances detected

for TrHbN and multiple polar mutant.

TrHbN Mutant

Tunnel Kd solvent-excluded

volume‡ Kd

solvent-excluded

volume‡

(mM) (%) (Å3) (mM) (%) (Å

3)

Global*

6.20 ± 0.04

(8.72 ± 0.04)†

7.92 ± 0.24

(10.5 ± 0.5)†

Surface 12.3 ± 0.03 18.5 ± 0.3

ST entrance 67.6 ± 1.1 43 89 ± 60 926 ± 33 7.8 45 ± 26

LT entrance 103 ± 1 32 66 ± 40 159 ± 10 25 61 ± 35

EHT entrance 199 ± 2 24 56 ± 38 196 ± 18 20 65 ± 41

* Affinities are calculated for the tunnels and the surface.

† Numbers in parentheses are for the calculated affinity for O2 (Surface and Tunnels).

‡The formation frequency and the average solvent-excluded volume detected.

153

Table 7.2 Rotamers observed for two residues upon the absence or presence of •NO molecule in

specific cavities.

Residue Rotamer empty Xe1 Xe2 Xe3 Xe5 EHc

Ile119(H11) mm 56 18 75 34 60 40

mt 2 3 9 37 3 17

pt 37 72 14 23 31 37

pp 3 4 1 0 2 1

others/outliers 2 3 1 6 4 5

Phe62(E15) m-30 30 19 22 29 1 40

m-85 50 79 44 45 8 35

t80 15 0 25 13 75 13

outliers 5 2 9 13 16 12

154

Figure 7.1 TrHbN structure (PDB entry 1IDR, subunit A). (Full legend on next page)

155

Figure 7.1 (legend) (top) TrHbN structure (PDB entry 1IDR, subunit A). The different tunnels

observed in TrHbN from MD simulations are represented by the orange surface (14). The B, E,

G and H helices are represented by blue, green, yellow and purple, respectively. The picture was

generated using PyMOL (39). (bottom) Schema of cavity-to-cavity diffusion routes found in the

present work. The amino acids separating the different cavities are identified. Diffusion in LT,

EHT and ST are represented by the green, blue and red arrows, respectively. The yellow arrow

indicates diffusion between EHc and the Xe1. BET was not used by •NO over 200 ns total MD

simulation time. The letter ’S’ indicates the solvent. Briefly, a •NO molecule can reach the distal

heme pocket using the ST, LT or EHT. Diffusion occurs by hopping from a cavity to another.

Using the ST, •NO leaves the bulk and enters the funnel-shape entrance nearby Xe3. Then, •NO

hops to Xe2 from where it can hop to the DHP (gray arrow). In the DHP, •NO is within contact

distance of the bound O2. Diffusion in the LT is quite different. First, •NO molecule enters the

protein matrix and docks inside Xe1, and constrains the Phe62(E15) residue to stay in the M

state. Diffusion toward Xe2 via Xe5 is sterically compromised but not impossible. This is

performed by hoping from Xe1 to Xe5 and then to Xe2 (Xe1→Xe5→Xe2). This scenario

involves a free energy barrier of 3.2±0.6 kcal/mol. An alternative route with a lower barrier

exists: Xe1→EHc→Xe2. The diffusion Xe1→EHc involves a free energy barrier of ~1.5±0.6

kcal/mol. When •NO occupies EHc, Phe62(E15) is unconstrained and therefore can adopt either

the M or the T state. In the T state, there is no free energy barrier between EHc and Xe2 and

diffusion between these two sites is allowed. For the EHT, diffusion begins with the •NO

passage from the bulk to the EHc. The next diffusion step toward the DHP occur as described for

LT. •NO escape to the bulk is unlikely because of the free energy cost (≈4.5±0.3 kcal/mol). Only

two such events were observed through the ST. A •NO molecule can also leaves DHP to dock

inside farther cavities (Xe2→EHc→Xe1). Diffusion through Xe5 (Xe2→Xe5→Xe1) is sterically

compromised by Phe62(E15) side chain which is restricted to the T state when •NO reaches Xe5

(Xe1 cavity is filled by the Phe62(E15) aromatic ring).

156

Figure 7.2 Time of contact between •NO molecules and TrHbN relative to the MD simulation

time. Atom coordinates were taken from the oxygenated TrHbN crystal structure (PDB entry

1IDR, subunit A). The entryway to the LT, EHT, ST and BET are indicated. The picture was

generated using PyMOL (39).

157

Figure 7.3 Representative solvent-excluded volume (105 Å3) formed over the ST entrance. a)

Overall view, b) zoomed view and c) side view of the solvent-excluded over the funnel-shaped

ST entrance. The solvent-excluded volume is represented by the gray mesh. The B, E, G and H

helices are colored in blue, green, yellow and purple, respectively. The picture was generated

using PyMOL (39).

158

Figure 7.4 (a) Density probability of •NO derived from explicit MD simulations and (b) implicit

ligand PMF for •NO inside TrHbN calculated from MD frames having the Phe62(E15) in the

(top) M state and (bottom) T state. In (a), higher •NO probability densities are represented by

colored isosurfaces. In (b), the free energy isosurfaces correspond to regions of measured PMF

of -1.5±0.3 kcal/mol (blue), -2.5±0.3 kcal/mol (yellow) and -3.5±0.3 kcal/mol (orange),

0 kcal/mol corresponding to ligand in vacuum. Tunnel entrances and cavities are indicated in a

and b, respectively. Pictures were generated using (a) PyMOL (39) and (b) VMD (25).

159

Figure 7.5 PMF profiles for •NO diffusion in ST (a) regardless and (b) as function of different

Ile119(H11) rotamers. For these ILS profiles, only 2988 and 3263 MD frames were available for

the mt and pp rotamers, respectively. Smaller samplings lead to overestimated PMF (24). PMF

profiles for the (c) LT and (d) EHT were calculated as function of Phe62(E15) M state (solid

line) and T state (dashed). The letter ’S’ indicates the TrHbN surface location.

160

Figure 7.6 Phe62(E15) χ1 and χ2 dihedral angles as function of the simulation time. The arrow,

at t ≈ 5.3 ns, indicates the entry of one •NO molecule in LT. This •NO diffused quickly from the

protein surface to the Xe1 binding site where it stayed for the remaining of the trajectory.

161

8.

Chapitre 8

Structure and Dynamics of Mycobacterium tuberculosis

Truncated Hemoglobin N: Insights from NMR Spectroscopy

and Molecular Dynamics Simulations

8.1. Résumé

L’activité oxyde nitrique dioxygénase (NOD) de l’enzyme TrHbN de Mycobacterium

tuberculosis (TrHbN-Fe2+


–OH2 + NO3-) protège la respiration aérobie

contre l’inhibition par le •NO. L’activité élévée de TrHbN a été attribuée en partie à la présence

de plusieurs cavités hydrophobes dynamiques permettant la partition et la diffusion des substrats

gazeux •NO et O2 au site actif. Nous avons étudié la relation existant entre ces cavités et la

dynamique de la protéine en utilisant la spectroscopie de résonance RMN en solution et par des

simulations de dynamique moléculaire (DM). Les résultats provenant des deux approches

indiquent que la protéine est principalement rigide avec des mouvements très limités des liens

amides (N-H) sur l’échelle de temps ps-ns. Ceci indique que la diffusion et la partition des

substrats à l’intérieur de TrHbN doit être controllées par les mouvements des chaînes latérales.

Des analyses de type « Model-free » ont aussi révélé la présence de mouvement lents (µs-ms),

non-observés en DM, pour plusieurs résidus positionnés le long des hélices B et G incluant le

résidu distal Tyr33(B10). Toutes les structures et les données de dynamique moléculaire

d’hémoglobines tronquées ayant une exension en N-terminal appelée région « pre-A » suggère

que celle-ci possède une structure secondaire stable en hélice alpha. De plus, une étude récente a

attribué un rôle crucial pour cette hélice pre-A pour l’activité NOD. Or, nos données RMN en

solution ont montré clairement que dans des conditions proches de celles physiologiques, ces

résidus n’adoptent pas une conformation en hélice alpha et sont plutôt significativement

désordonés. La conformation observée dans les cristaux serait attribuable à des contacts

cristallins. Aussi, le manque d’ordre dans la pre-A ne signifie pas pour autant que cette région ne

joue pas un rôle fonctionnel important mais s’il existe, que celui-ci ne peut être expliqué par la

162

conformation en hélice de ces résidus. De plus, les simulations de DM futures ne devraient pas

être démarrées avec la pre-A avec une conformation en hélice afin d’éviter des biais basés sur

des structures initiales erronnées. Ce travail constitue la première étude dynamique de la

structure d’une hémoglobine tronquée par spectroscopie RMN en solution.

8.2. Abstract

The potent nitric oxide dioxygenase (NOD) activity (TrHbN-Fe2+


–

OH2 + NO3-) of Mycobacterium tuberculosis truncated hemoglobin N (TrHbN) protects aerobic

respiration from inhibition by •NO. The high activity of TrHbN has been attributed in part to the

presence of numerous short-lived hydrophobic cavities that allow partition and diffusion of the

gaseous substrates •NO and O2 to the active site. We investigated the relation between these

cavities and the dynamics of the protein using solution NMR spectroscopy and molecular

dynamics (MD). Results from both approaches indicate that the protein is mainly rigid with very

limited motions of the backbone N−H bond vectors on the picoseconds-nanoseconds timescale

indicating that substrate diffusion and partition within TrHbN may be controlled by side-chains

movements. Model-free analysis also revealed the presence of slow motions (μs-ms), not

observed in MD simulations, for many residues located in helices B and G including the distal

heme pocket Tyr33(B10). All currently known crystal structures and molecular dynamics data of

truncated hemoglobins with the so-called pre-A N-terminal extension suggest a stable alpha-

helical conformation that extends in solution. Moreover, a recent study attributed a crucial role to

the pre-A helix for NOD activity. However, solution NMR data clearly show that in near-

physiological conditions, these residues do not adopt an alpha-helical conformation and are

significantly disordered, and that the helical conformation seen in crystal structures is likely

induced by crystal contacts. Although, this lack of order for the pre-A does not disagree with an

important functional role for these residues, our data show that one should not assume an helical

conformation for these residues in any functional interpretation. Moreover, future molecular

dynamics simulations should not use an initial alpha-helical conformation for these residues in

order to avoid a bias based on an erroneous initial structure for the N-termini residues. This work

163

constitutes the first study of a truncated hemoglobin structure and dynamics performed by

solution NMR spectroscopy.

8.3. Introduction

Tuberculosis infects one-third of the world’s population. Most infected individuals fail to

progress to complete disease because the TB bacilli are maintained in a latency state by the

immune system. During latent infection mycobacteria are exposed to low O2 concentrations and

to nitric oxide (•NO) produced by the immune system of the host. Since •NO can inhibit or

inactivate key enzymes such as the terminal respiratory oxidases and the iron/sulfur protein

aconitase and can generate secondary reactive nitrogen species displaying varied reactivity and

toxicity (1-7), •NO-metabolizing reactions are thus required for Mycobacterium tuberculosis

(Mtb) to fight •NO poisoning. The truncated hemoglobin N (TrHbN) from the pathogenic

bacterium Mtb has a potent ability to detoxify •NO to nitrate (nitric oxide dioxygenase reaction

(NOD)) and to protect aerobic respiration from the inhibition by •NO in stationary phase cells of

M. bovis BCG (8). The high rate of •NO oxidation (kNOD ≈ 745 µM-1

s-1

at 23°C) catalyzed by

oxygenated TrHbN and the large affinity (Kd = 8 nM) for O2 suggests that the NOD reaction

may be one of the vital defense systems in Mtb for coping with the toxic effects of •NO under the

low O2 concentration (1-4 µM) prevailing in infected lesions (9). An understanding of the

structure and dynamics characteristics leading to this high reactivity is essential, hence the need

to study the molecular mechanisms controlling ligand/substrate access to the distal heme pocket

(DHP).

A striking feature of TrHbN is the presence of multiple narrow hydrophobic tunnels connecting

the active site to distinct protein surface sites. Fig. 8.1 shows the tunnels that have been

identified using MD simulations (10) and X-ray crystallography (11): Short Tunnel (ST), Long

Tunnel (LT), EH Tunnel (EHT) and BE Tunnel (BET). These tunnels are not open channels but

are formed of short-lived hydrophobic neighboring cavities of various shapes and volumes (10)

that partition substrates from solvent. These cavities are temporally interconnected due to side-

chain flexibility. The high rigidity of TrHbN backbone as observed in MD simulations (10)

164

enables these numerous cavities. Interestingly, except for one, all these cavities correspond to the

Xe binding pockets identified by X-ray crystallography (11). Dynamic hydrophobic tunnels of

narrow diameter, as observed in TrHbN, may also prevent the formation of hydrogen bonded

water clusters that can hinder passage of other molecules through the tunnel. Water molecules

and other polar substrates in hydrophobic tunnels typically have very rapid sub-nanosecond

transit times that insure that water will not hinder passage of less polar substrates. In contrast to

the Xe cavities in myoglobin, the hydrophobic tunnels in TrHbN constitute the trajectory linking

the DHP to the solvent (10, 12-15)[170-174].

As we reported earlier, the local entrance of the tunnels also contributes to the overall reactivity

both through substrate selectivity and through the dynamics controlling the transit between the

external solvent and the internal tunnel (12). The hydrophobicity and polarity of the cluster of

residue side chains localized at the tunnel entrance can create domains that favor local

enhancements in the concentration of potential substrates (12). Reactants such as •NO are

soluble in aqueous environments but have higher solubility in hydrophobic environments. Thus

hydrophobic patches in the region of the tunnel entrance on an otherwise hydrophilic/polar

surface may serve to create an enhanced build up of •NO at the proposed site of entry. Our recent

results (12) show •NO molecules interacting with oxygenated TrHbN preferentially at the

entrance of the tunnels. In TrHbN the hydrophobic nature of the entrances forces water

molecules partitioning away from the protein surface and thus favors •NO capture.

A functional consequence of the tunnels is that •NO reaches the bound O2 without the need for

important structural changes allowing TrHbN to catalyze NOD reaction at a rate approaching

that of diffusion-controlled reactions. Interestingly, myoglobin requires the distal His64(E7) to

swing out of the DHP, resulting in a much slower NOD activity (kNOD ≈ 45 µM-1

s-1

at 23°C).

Another unusual feature of TrHbN is the presence of the so-called pre-A helix, located at the N-

terminus. This “floating” tail was recently assigned a functional role by influencing the diffusion

of ligands to the active site (16). The authors proposed that the deletion of the pre-A helix alters

protein dynamics, especially the conformational state of the Phe62(E15) residue restricting the

passage of NO to the DHP, thus affecting the ability of TrHbN for •NO detoxification (16).

However, since the protruding and rigid conformation of the pre-A helix observed in the crystal

165

lattice could be questionable, it may be ill-advised to interpret the functional role of the pre-A

helix based on the unusual conformation observed in the crystal structure.

To this day, no experimental data are available on the structure and dynamics of TrHbN in

solution. The present work describes the first study of a truncated hemoglobin by solution NMR

spectroscopy. The structure and dynamics of TrHbN in the cyanomet form were studied using

NMR chemical shifts, relaxation data acquired at three magnetic fields, and NMR amide

exchange experiments. The cyanomet form was studied since the majority of currently available

structures for TrHbN, including the structure with Xe atoms highlighting the tunnel network,

were obtained in complexes with cyanide in the ferric state, and that cyanide represents a

valuable diatomic ligand model system. In order to evaluate the structural integrity of the pre-A

helix in solution, we have studied both the wild type protein and the ∆pre-A mutant. To

complement the NMR data, the characterization was carried out in combination with MD

simulations.

8.4. Material and methods

8.4.1. NMR

Protein expression, labeling, and purification - The genes coding for the mature wild-type

TrHbN protein (136 residues) as well as for the Δpre-A mutant (lacking residues 1 to 12) were

cloned in E. coli BL21(DE3) cells. Proteins were prepared by reconstitution of the apoprotein

with heme in a manner similar to the one described by Scott and Lecomte (17). Details of protein

preparation methods can be found in Supporting Material (Annexe 3).

Samples preparation and NMR spectra recording - 15

N- or 15

N/13

C-uniformly labeled protein

samples were prepared at a concentration of 0.8 mM in a 20 mM KPO4 pH 7.5 buffer containing

50 μM EDTA, 3 mM KCN, 10% D2O, 1X Complete protease inhibitors (Roche) and 0.1 mM

DSS. All experiments were performed at 26.4°C (calibrated using MeOH) on Varian INOVA

600 (Université Laval, Québec, Canada), 500, and 800 (Québec / Eastern Canada High Field

166

NMR Facility, McGill University, Montréal, Canada), all equipped with z-axis, pulsed-field

gradient, triple-resonance cold probe.

NMR resonance assignment - For backbone and side-chains resonances assignment, the

following spectra were collected using pulse sequences from Biopack (Varian Inc, Palo Alto,

CA): 2D 15

N-HSQC, 2D 13

C-HSQC (both aliphatic and aromatic), 3D HNCO, 3D

CBCA(CO)NH, 3D HNCACB, 3D C(CO)NH, 3D HC(CO)NH, 3D HCCH-TOCSY, 3D 13

C-

NOESY-HSQC, and 3D HNHB. All these experiments were performed on a Varian INOVA 600

spectrometer (Université Laval, Québec, Canada). All spectra were processed using

NMRPipe/NMRDraw (18) and analyzed within NMRView (19).

15N spin relaxation experiments -

15N-R1,

15N-R2, and {

1H}-

15N NOE

NMR relaxation

experiments were recorded at proton frequencies of 500, 600, and 800 MHz. Pulse sequences

from the Kay group were used (20, 21). More details are available in the Supporting

Material (Annexe 3).

Model-free analysis - The model-free analysis (22, 23) was performed using an axially

symmetric diffusion tensor within the program MODELFREE 4.20 (A.G. Palmer III, Columbia

University, New York, NY). Details are available in the Supporting Material (Annexe 3).

Amide exchange experiments - Amide exchange experiments were performed as described in

Morin and Gagné (24). Data were recorded at 600 MHz at both pH 7.5 and 8.5. At pH 7.5, a total

of 45 15

N-TROSY-HSQC spectra (BIOPACK, Varian, Palo Alto, CA) were recorded. Details

can be found in Supporting Material (Annexe 3).

8.4.2. Molecular dynamics simulations

Force field optimization of the cyanide-bound heme atomic charges and Fe-C-N angle

parameter - CHARMM22 (25) force field lacks parameters for the cyanide-bound heme. To

simulate CN-bound TrHbN in this study, the atomic charges of heme prosthetic group as well as

the cyanide and the Fe-C-N angle parameter were optimized following the standard

167

parameterization protocol for the CHARMM22 force field (25). The same procedure was

previously applied for O2-bound heme force-field parameters (10) and details are given in

Supporting Material (Annexe 3).

Systems and simulation setup - Initial coordinates were taken from wild type cyanomet TrHbN

crystallographic structure (PDB 1RTE) (14). Crystallographic water and sulfate ions were

ignored. Hydrogen atoms were added using CHARMM’s HBUILD facility (26). All ionisable

residues were considered in their standard protonation state at pH 7.0 with neutral histidine

protons placed at the ND1 position. The crystal unit cell contained two TrHbN molecules (A and

B chains). Both chains were used individually to start two independent 85 ns MD simulations in

order to increase sampling (trajectories identified hereafter as A-TrHbN and B-TrHbN). The first

5 ns were considered as equilibration time. Complete protocol is available in Supporting


Trajectory analysis - The protein N-H bond S2 parameters were calculated using the M2 method

described in Fisette et al. (27). Briefly, method M1 was used to verify convergence of the

backbone amide autocorrelation function (C1(t) (approximating the ensemble using the last 80 ns

of simulations for t < 40 ns). When a function did not converge, S2 was estimated as the average

of the last 500 ps. In M2 method, all structure snapshots from the two trajectories were combined

and randomized. The correlation function decays to its plateau value immediately after C1(0).

The S2 estimate therefore takes into account 160 ns of simulation time. Backbone amide group

accessible surface area (ASA) was calculated using a probe radius of 1.4 Å at every 10 ps.

Backbone amide hydrogen bond occupancy was analyzed with various N-H acceptor distance

cutoffs (2.0, 2.2 and 2.4 Å) and a minimum N-H acceptor angle of 120°.

168

8.5. Results

8.5.1. NMR

Protein expression, labeling, and purification - Using the protocol described in Supplementary

Material (Annexe 3), ~20 mg of highly pure reconstituted TrHbN (r-TrHbN) per liter of M9

medium were obtained for the wild type protein and ~80 mg per liter for the reconstituted Δpre-

A mutant (Δpre-A r-TrHbN). According to LC-MS analysis, incorporation of 13

C and 15

N in our

samples was ≥ 96%. The samples were stable for several weeks at room temperature. Kinetics

parameters for O2 and CO binding and dissociation of r-TrHbN were determined (data not

shown) and found identical to those already published. Resonance Raman spectra for the O2 and

CO r-TrHbN complexes also indicate that the DHP is quite identical to that of TrHbN. Dynamic

light scattering measurements were made at several concentrations to ensure that there was no

aggregation/oligomerization in the NMR sample. The protein was found as a monomer in

solution at the concentration used for NMR. We also noticed that Met1 was cleaved for the wild

type protein but was still present for the Δpre-A mutant, as confirmed by mass spectrometry.

1H,

15N, and

13C resonance assignments for TrHbN cyanomet - As shown in Fig. 8.2, the 2D

15N-HSQC spectrum of r-TrHbN cyanomet is of high quality, with a good dispersion of the N-H

resonances. Backbone resonances assignment was completed for cyanomet r-TrHbN at 97%,

99%, 96%, and 96% respectively, for backbone amides, Cα, Cβ, and carbonyls. Missing amide

assignments are for Leu3, Arg6, Lys9, Gly74 as well as for the 7 proline residues (Pro12, Pro69,

Pro71, Pro76, Pro108, Pro121(H13), and Pro135). Most of the missing amides are located in the

pre-A region of the protein (residues 2 to 9) as they were too weak to be visible in the spectra.

Assignment for the amide group of Gly74 was not possible because of overlapping with other

glycine residues. Missing assignments for Cα are for Gly2 and Ser5. For Cβ, missing assignments

are Ser52(E5), Val107(GH5), and Ser109(H1). Finally, missing assignments for carbonyls are

Gly2, Leu4, Ser5, Arg8, Arg53(E6), and Thr73. For side chains, 75% of the resonances

assignment was completed. Chemical shifts have been deposited in the Biological Magnetic

Resonance Data Bank (BMRB) under accession number 17226.

169

15N spin relaxation data -

15N spin relaxation data were recorded at three magnetic fields (500,

600, and 800 MHz proton Larmor frequencies) in order to better characterize the dynamics of r-

TrHbN in solution. Mean values for 15

N-R1, 15

N-R2 and {1H}-

15N NOE are presented in

Table 8.1 and all experimental values are plotted in Fig. 8.3, A-D. Values obtained are

homogenous and show the same pattern for all the magnetic fields, with smaller R2 for residues

in N- and C- termini and in loops, as expected for these more mobile regions. Raw relaxation

data for each residue at the three fields are available in Table S2 of the Supporting

Material (Annexe 3). 15

N spin relaxation data for r-TrHbN have been deposited in the BMRB

under accession number 17226.

Model-free analysis - It was possible to get high-quality data for 101 out of the 127 non-proline

observable residues at the lowest magnetic field of 500 MHz, overlapping or too broad

resonances being responsible for non-analyzed amides. Only these 101 residues which have data

at the three magnetic fields were used to perform the model-free analysis.

Examination of the inertia tensor of r-TrHbN indicates an asymmetric protein, with relative

moments of 1.00: 0.96: 0.45. Many diffusion tensors were tested, using either the program

relax (28, 29) (local τm, sphere, prolate spheroid, oblate spheroid, and ellipsoid) or Quadric (A.

G. Palmer III, Columbia University) (isotropic, axial, and anisotropic). The best model appeared

to be prolate spheroid with relax and axial with Quadric, which are mostly the same.

Consequently, an axially symmetric (spheroid) tensor was used for the model-free analysis.

Global optimization of the tensor showed that TrHbN tumbles anisotropically in solution with a

D||/D value of 1.45 and a global tumbling time of 10.0 ns, similar to the results obtained for

other globins of the same size and shape (30, 31).

For local motions, the data were fitted to the five following model-free models: (m1) [S2], (m2)

[S2, τe], (m3) [S

2, Rex], (m4) [S

2, τe, Rex], (m5) [S

2f, S

2s, τe], where S

2 (=Sf

2Ss

2) is the square of the

generalized order parameter characterizing the amplitude of internal motions, Sf2 and Ss

2 are the

squares of the order parameters for the internal motions on the fast (ps) and slow (ns) timescales,

respectively, τe is the effective correlation time for internal motions, and Rex is an additional

parameter added to contributions to observed R2 from conformational exchange and pseudo-first-

order processes occurring on the microsecond-to-millisecond timescale (32). Following the final

170

model-free model selection step, 46 residues were fitted with model m1, 23 with model m2, 9

with model m3, 6 with model m4, and 17 residues with model m5. Models m1 and m2 are the

simplest models and were used to fit most of the residues. These models are representative of

residues exhibiting fast pico- to nanoseconds motions for their N-H vectors. Results from the

model-free analysis are presented in Table S3 (Annexe 3).

NMR order parameters, S2 -The S

2 generalized order parameter characterizes the amplitude of

internal picoseconds-nanoseconds timescale motions of N-H bond vectors. S2 values vary from 0

for a completely disordered vector, to 1 for totally restricted vector (33). S2 values obtained for r-

TrHbN are plotted in Fig. 8.3, E, and mapped onto the protein structure in Fig. 8.4, A.

Considering all residues, an average order parameter (S2) of 0.84 was obtained for r-TrHbN. In

particular, we see that helices B, and E are the most rigid of the protein with an average S2 value

of 0.91, which is higher than the average value of 0.88 reported for α-helices by Goodman et al.

(34). Helices F and G appeared to be the most flexible, with average S2 of 0.84 and 0.86,

respectively. Order parameters for other helices are typical.

As expected, a lower degree of motion restriction was observed for loops as well as for both N-

and C-termini, for which a very high degree of mobility is observed with average S2

values of

0.42 and 0.21, respectively. S2 order parameters for r-TrHbN have been deposited in the BMRB

under accession number 17226.

Slow motions - The Rex term in model-free models m3 and m4 accounts for slow μs to ms

motions. Fifteen residues of r-TrHbN were fitted with one of these models, indicating the

presence of chemical exchange occurring in the slow μs to ms timescale in the vicinity of these

residues (Glu30(B7), Tyr33(B10), Gly83, Thr87(G1), His90(G4), Phe91(G5), Val94(G8),

Ala95(G9), Leu98(G12), Ala99(G13), Ala101(G15), Leu102(G16), Ala105(G19), Ile112(H4))

(see Fig. 8.3 F and Fig. 8.4 B). Most of these slow motions are observed for N-H bond vectors of

residues located in helices B and G, and pointing toward the active site of the protein, some of

them located directly in or at the entrance of a tunnel. Active site residues requiring an Rex term

include Tyr33(B10) and Val94(G8) while residues defining tunnel entrances include Phe91(G5)

and Ala95(G9) for the ST tunnel and, Glu30(B7) and Ty33(B10) for the BET tunnel.

171

Two-timescale motions - Seventeen residues were better characterized by using model m5

(Leu7, Glu11, Ile13, Ala24(B1), Ala44, Glu70, Ala75, Gln79, Val80(F7), Arg84, Gly106,

Asp125, Gly129, Glu130, Thr133, Ala134, Val136), indicating the presence of motions

occurring at two different timescales (Sf2 = ps, Ss

2 = ns) for these N-H vectors. All these residues

are located in flexible regions of the protein such as N- and C-termini, loops, or at the beginning

of helices.

NMR amide exchange experiments - Amide protons can exchange using one of two regimes:

EX1 or EX2 (41). In the EX1 regime, exchange rates are pH independent and give kinetics

information on the exchange process, i.e. the exchange rate (kex) is equal to the opening rate for

the local structure protecting the N-H group from solvent access (kop). In the EX2 regime,

thermodynamics information can be extracted from pH dependent exchange rates. The extracted

information provides the following parameters, which correspond to different ways to present the

same information: ΔGHX (the free energy for the opening of the local structure protecting the N-

H group), Kop (the equilibrium constant) and PF (the Protection Factor for the exchange,

Kop = 1/PF). Generally, under non denaturing conditions (i.e. high stability), amide protons

exchange from the EX2 regime. A simple approach for determination of the exchange regime

consists in measuring exchange rates at two pH. If rates from the highest pH are faster, then the

exchange proceeds using the EX2 regime at the lowest pH. As can be seen from Fig. 8.3, J,

amide exchange at pH 7.5 proceeds from the EX2 regime. Indeed, the rates at pH 8.5 are in

average 30±10 times faster than rates at pH 7.5 (while theory would suggest a difference of 10X

between rates from both pH). From these data, ΔGHX, Kop and PF can be extracted to

characterize the exchange process at pH 7.5. Fig. 8.3, K, shows a plot of ΔGHX values according

to the residue number and these values are mapped on the protein structure on Fig. 8.4, C.

Quantitative data is available for only 52 residues, while semi-quantitative data is available for

60 residues (12 too slow, 48 too fast for being quantified). The pre-A region appears as the less

stable portion in TrHbN. Indeed, from the 5 residues that could be characterized in the pre-A

region, neither had a measurable exchange rate, i.e. all exchanged too fast to be characterized.

This means that N-H groups in this region are poorly protected from exchange, with ΔGHX lower

than 5 kcal/mol. This is in agreement with missing resonances and very low S2 values observed

for this part of the protein. Other short helices such as the A, C, F and H’ helices have low ΔGHX.

On the contrary, helices B, E, G and H have ΔGHX of higher value, even including residues for

172

which exchange rates could not be measured accurately due to the limited time allowed for

exchange (~42 days). Exchanging on an intermediate timescale, the H helix has ΔGHX between 7

and 10 kcal/mol. Amide exchange data have been deposited in the BMRB under accession

number 17226.

NMR investigation of the Δpre-A mutant - A recent study concluded that the pre-A region

may play an important function by altering the dynamics of the protein core and thus ligand

diffusion (16). In the latter work, the excision of pre-A region triggered changes in LT dynamics,

especially for Phe62(E15) side-chain, leading to the blockade of the LT and NO access to the

heme-bound oxygen. To validate these conclusions we generated a mutant protein corresponding

to the pre-A mutant described in (16) (lacking residues 1-12) and investigated its structure and

dynamics by NMR spectroscopy. pre-A r-TrHbN was as stable as r-TrHbN and showed

identical kinetic binding properties. Comparison of 15

N-HSQCs from the wild type and the Δpre-

A mutant indicated that changes in the electronic environment are limited to residues

neighboring the mutation site. Analysis of the C sidechain chemical shifts revealed that only

two residues had significant chemical shift change upon pre-A deletion : Asp17 (0.23 ppm) and

His22 (0.34 ppm). Asp17 C is near the deletion and 6.6Å from the terminal NH3+ in the

deletion mutant. On the other hand, His22 is far from pre-A residues in the crystal structure

(nearest group is the charged extremity of Arg6 at 9.2Å). We therefore found little evidence that

the pre-A region was interacting strongly with other parts of the protein. Fig. 8.5 illustrates the

chemical shifts differences (Δδ) between the 1H

15N resonances from r-TrHbN and Δpre-A r-

TrHbN.

Solution structure of the pre-A “helix” - As shown in Fig. 8.3, E, all N-termini residues up to

residue 13 have low order parameters, with steadily decreasing order parameters going toward

the N-termini as typically seen in unstructured extremities of several proteins. The region

includes the so-called pre-A “helix” that was observed in all TrHbN crystal structures. In order to

evaluate the helical character of the pre-A region in solution, we have used the program SSP,

which uses NMR chemical shifts to calculate a single residue-specific secondary structure

propensity score (36). As shown in Fig. 8.3, N, the score obtained for pre-A residues in TrHbN is

around 0, implying an absence of α-helical secondary structure predisposition in solution,

consistent with our NMR spin relaxation and amide exchange data.

173

8.5.2. Molecular dynamics simulation and comparison with NMR results

Stable trajectories were obtained in both molecular dynamics simulations performed. Average

backbone root mean square deviation (RMSD) for residues in α-helices stabilized at ~0.7 Å.

MD order parameters of N-H bond vectors - Both MD trajectories were used to calculate the

generalized order parameter S2 for backbone N-H bonds (sampling from a total of 160 ns).

Comparisons between MD and NMR data are shown in Fig. 8.3, H-I. S2 values, N-H bond

internal autocorrelation function and local motions are available in Annexe 3, Table S4 and

Fig. S1. Individual MD simulations produced very similar S2 datasets. Significant S

2 differences

between these two simulations are observed for the pre-A region and for some residues located

in C-E and E-F loops. For some residues located in loops, the autocorrelation function C1(t) did

not converged during simulations (residues Asp39, Glu70, Gly74). S2 estimates for these

residues would have required more and/or longer MD simulations to converge.

S2 obtained from MD simulations are in very good agreement with those obtained from NMR, as

shown in Fig. 8.3, H-I. The average MD S2 obtained for the four α-helices forming the 2-on-2

fold and enclosing the tunnels (B, E, G and H helices – S2 of 0.89) is the same as from NMR and

reflects very restricted motions in the ps-ns timescale. RMSD between MD and NMR is 0.10

using all residues, 0.16 for loop residues, and 0.06 for residues in α-helices. Similar agreements

were obtained in a recent study on the of E. coli β-lactamase TEM-1(27). The main discrepancy

concerns the pre-A region where MD-derived S2 are much higher than those obtained from

NMR. Other discrepancies concern some residues located within C-E and E-F loops. These latter

discrepancies are explained by the non-convergence of the autocorrelation function for several

residues.

Further agreement of results from NMR and MD concerns residues exhibiting two timescale

motions. Of the seventeen residues fitted with model m5 in the model-free analysis (two-

timescale motions on the ps-ns timescale) nine were observed to have similar motions in MD

simulations. This is shown in MD simulations by N-H vectors exploring more than a single

conformation (see residues Glu11, Ile13, Glu70, Arg84, Gly106, Gly129, Glu130, Thr133, and

Ala134 in Fig. S1(Annexe 3) ).

174

Hydrogen bonds occupancies compared to amide exchange rates - Backbone N-H hydrogen

bond occupancy was calculated for the different simulations and is shown in Fig. 8.3, M. As

expected, occupancy is higher in regions with well formed secondary structure. Particularly,

higher occupancies are found in helices E and G correlated with slower amide exchange rates. As

for S2 parameters, however, discrepancies are found for the pre-A region. Indeed, the pre-A

region remained structured as an α-helix along MD simulations and thus high H-bond

occupancies (Fig. 8.3, M) and low solvent exposure (Fig. 8.3, L) were calculated. This is in

striking contrast with the very fast experimental amide exchange rates.

8.6. Discussion

To this day the structure and dynamics of TrHbN have been studied by several techniques such

as X-ray crystallography, resonance Raman spectroscopy, or MD simulations. The present work

presents the first structure and dynamics study of TrHbN using solution NMR spectroscopy.

NMR data indicate that the 2-on-2 fold may provide sufficient rigidity to host several tunnels.

High S2 parameters derived from model-free analysis and MD simulations show that motions of

backbone amide vectors in TrHbN are very limited on the ps-ns timescale. As we already

proposed high rigidity of the backbone may enable hosting the numerous cavities that form the

gas pathways connecting the DHP to the solvent (10). The high rigidity of the B, E, G and H

helices, which form the characteristic 2-on-2 helical core of the truncated hemoglobin fold, may

prevent the optimization of internal side-chains Van der Waals interactions that would otherwise

result in tighter side-chain packing and the loss of tunnels. At the same time, the free internal

volume of tunnels coupled with thermal fluctuation may allow flexibility of the side-chains

making the tunnels, which is mandatory for gas migration (10, 12, 37-41). This structural feature

could also be a prerequisite in other tunnel-containing proteins. Since most of TrHbN tunnel

residues contain methyl groups (alanine, valine, leucine and isoleucine), NMR characterization

of these side-chain motions using 2H spin relaxation within these methyl groups may represent a

great biophysical interest. For example, if NMR reveals side-chain motions for residues that

appear tightly packed in MD simulations, other substrate/product diffusion routes may be

175

formed. Afterward, new models could be tested by performing biased MD simulations

techniques or by performing extended MD simulations to capture non-frequent motions.

Dynamics in tunnels and in the active site - Rex terms were observed for 13 residues located in

helices B, G, and H (Fig. 8.4, B). Among these we find the distal residue Tyr33(B10), shaping

the BE tunnel, and involved in ligand binding and stabilization (15, 42-43). Since many catalytic

processes occur on this timescale, the presence of such slow motions for this tyrosine is most

likely relevant for the function of the protein.

Fig. 8.4, B, also illustrates that most of the residues having a Rex term are found in helix G which

together with helix H define the ST tunnel, and that all these residues point toward the protein

core and not toward the solvent. Such co-localized slow motions could reflect the presence of a

molecule slowly diffusing in the tunnels of the protein. Another potential explanation for the

presence of these slow motions would be a translational movement of this helix or of the

neighboring H helix in solution. Such a motion, would not contribute to the NMR-derived S2, but

it would modulate the chemical shifts of amides at the interface and, depending on the timescale,

would lead to the observation of Rex terms in the model-free analysis. This kind of motion may

increase significantly internal free volume allowing larger substrates to penetrate the protein

core. Consequently, this motion may also be necessary for bulkier NO3- release. As a

comparison, important structural changes were observed upon binding of cyanide to the ferric

form of the truncated hemoglobin Synechocystis PCC 6833 (Syn-TrHb) (44, 45). In its ferric

form, Syn-TrHb has His46(E10) side chain bound to the heme iron (46), with important residues

implicated in ligand binding pointing toward the solvent (Tyr22(B10) and Gln43(E7)) or being

far above the heme iron (Gln47(E11)) (47). Following displacement of His46(E10) by cyanide,

significant reorganization of the B and E helices occurs, causing Tyr22(B10) and Gln43(E7) to

enter the DHP and formation of a H-bonding network between Tyr22(B10), Gln43(E7),

Gln47(E11) and the bond cyanide (44, 45). Interestingly, no tunnels were found in the 6-

coordinate crystal structure of SynHb, while the LT and EHT are evident in the cyanomet form

(44, 47). These two structural conformations imply important changes in internal side chains

packing. In this view, μs-ms motions in r-TrHbN could have repercussions on the dynamics and

organization of the tunnel, which govern substrate/product diffusion to and out of the DHP.

176

We have attempted to quantify these μs-ms motions observed from spin relaxation using CPMG

relaxation dispersion experiments, without success (data not shown). This absence of dispersion

in the CPMG relaxation dispersion profiles is still consistent with the Rex parameters extracted

from model-free analysis. Indeed, in the model-free analysis scheme we used, the contribution

from μs-ms motions on R2 is assumed to be quadratically dependent on the magnetic field

strength, i.e. in the fast-exchange limit. Of course, the potential contribution from motions in the

slow-exchange limit, which is independent of the magnetic field, is not totally excluded from

these data. However, the flat profiles from CPMG relaxation dispersion experiments confirm that

the μs-ms motions detected from model-free analysis are most probably outside the slow-

exchange limit, i.e. in the fast-exchange limit, probably with kex > 10000 s-1

since CPMG (with

νcpmg up to 1000 Hz) could not quench the exchange.

Pre-A “helix” is not an helix - The main structural divergence between results from NMR and

previous results from X-ray and MD concerns both the rigidity and the presence of a secondary

structure for the pre-A region (residues 2-9). This region forms a somewhat unusual α-helix that

extends away from the globular structure in all X-ray structures of TrHbN and is quite stable in

MD simulations that use X-ray coordinates for an initial structure. NMR data showed no

evidence for such a stable helix, or even helical propensity, in solution. Experimental data

supporting disorder and the lack of secondary structure for the so-called pre-A “helix” are: 1)

Missing or very weak resonances in the NMR spectra, attributed to broadened peaks and/or fast

exchange with the solvent; 2) Very low S2 parameters, reflecting highly disordered N−H vectors;

3) N-H exchanging too fast to be characterized by amide exchange experiments, meaning a very

low protection factor for this region (thus no or low residency H bonds); 4) The propensity for

secondary structure from chemical shifts indicates no helical conformation for this region.

The presence of such an α-helical structure for the pre-A region in the TrHbN X-ray structures

could be favored by the high ionic strength conditions used for crystallization as well as by the

crystal contacts. Indeed, careful observation of the packing of the pre-A region in the crystal

lattice shows contacts with other chains which may favor an helical conformation. Although an

helical conformation could potentially be transiently populated in solution, this would have to be

extremely low since we saw no signs of helical propensity under the physiological conditions

used for the NMR study.

177

Only small localized changes were observed following the removal of the pre-A region; this is in

agreement with CD studies proposing that this region does not contribute to the structural

integrity of the protein (16). The CD data from Lama et al. (16) actually support a disordered

pre-A model. They noted that “the content of random coil decreased when pre-A was deleted

from M. tuberculosis HbN and increased when pre-A region was added to M. smegmatis HbN”.

We believe that this random coil changes in their CD can be attributed to the removal/addition of

a disordered pre-A region. Our pre-A results partially agree with the presence of an interaction

between Arg10 and Glu70 previously detected (35). Our data show that the C chemical shift of

Glu70 does not change upon deletion of the pre-A, hence suggesting that the environment of the

Glu70 sidechain is not affected by the presence or absence of the pre-A region. As mentioned

above, only two C chemical shifts changes were noted upon deletion of the pre-A region, one of

them being His22 (0.34ppm shift). A potential cause for this shift may be transient polar

interactions in the native protein between Glu70 (which is hydrogen bonded to His22 in the

crystal structure) and one or more of the positively charged groups in the disordered pre-A

(Arg6, Arg8 or Arg10), similar to the transient interaction observed by MD between Arg10 and

Glu70 (35).

However, our data disagree with one result from the molecular dynamics simulations done by

Lama et al. (16). In this study, it was shown that the excision of the pre-A region results in

distinct changes in the protein dynamics. They observed that the Phe62(E15) gate is trapped in

the closed conformation in the mutant. From our NMR data, there is no significant change in the

amide and C chemical shifts for Phe62(E15) and nearby residues in the pre-A mutant.

Moreover, we have recorded R2 parameters at 600 MHz for this pre-A mutant and the R2 value

observed for Phe62(E15) is the same as the mean R2 in both proteins. This implies a similar

degree of slow timescale (μs-ms) dynamics in both situations.

178

8.7. Conclusion

This study reports the first experimental characterization of the dynamics of a truncated

hemoglobin. The characterization of the structure and dynamics of ferric Mycobacterium

tuberculosis TrHbN bound to cyanide were done by two complementary techniques, NMR and

MD simulations. This stable cyanomet form provides a good diatomic ligand system model, and

corresponds to the most studied form by X-ray crystallography to date. Although the nature of

the hydrogen bond network with the ligand on the distal side differs slightly with that observed

in the ferrous oxygen-bound structure (14), X-ray crystallography and MD simulations suggest

that the behavior of tunnels and pre-A conformation does not vary significantly between the

oxygen-Fe(II) and cyanide-Fe(III) forms of TrHbN. Consequently, the results presented here can

be extended to the more physiological oxygen-Fe(II) form.

Results obtained from both NMR and MD are in a very good agreement, validating previous

results as well as our protocols. The results emphasize that TrHbN is mainly rigid, especially for

residues forming the 2-on-2 fold showing very limited motions of the N-H vectors in the ps-ns

timescale probed (average S2 for secondary structures of 0.89). Model-free analysis revealed the

presence of unexpected and localized motions in the µs-ms timescale taking place at the DHP

residue Tyr33(B10) and also all along the G-helix. These motions could have important

repercussions on the dynamics of the tunnels and the active site and thus, on ligand diffusion and

kinetics properties. Amide exchange experiments, performed at pH 7.5 and 8.5, also evidenced

the very high stability of TrHbN 2-on-2 fold. While X-ray structures show pre-A as an

α−helix (14-15), our NMR data revealed this N-terminal region to be very disordered and

showed no evidence for the presence of any secondary structure. The novel pre-A conformation

presented here is not in contradiction with a vital role for the pre-A region, as was recently

proposed (16). However, we believe it would be ill-advised to assume an helical conformation

and structural order for residues 2-9 in attempting to interpret the functional role of the pre-A

region. Moreover, we recommend that future MD simulations of TrHbN should not be carried

with an initial helical conformation for the pre-A region; an initial random conformation would

likely be more accurate. Taken together, these findings are of great interest toward our

179

understanding of TrHbN function. This work could also be of interest for other gas-interaction

and tunnel-containing enzymes.

Since most of TrHbN tunnel residues contain methyl groups (alanine, valine, leucine and

isoleucine) it may now be possible to follow 2H spin relaxation within these methyl groups in

order to probe internal side chain flexibility.

This study is thus the first step of a wide analysis of TrHbN dynamics in various forms by NMR

techniques.

8.8. Acknowledgments

The authors thank Prof. Martino Bolognesi and Olivier Fisette for stimulating discussions. This

work was supported by the Natural Sciences and Engineering Research Council of Canada

(NSERC), the Fonds québécois de recherche sur la nature et les technologies (FQRNT), and

PROTEO, the Quebec Network for Research on Protein Function, Structure, and Engineering.

8.9. Supporting information

Additional details, optimized CN-bound heme atomic charges, 15

N spin relaxation data, model-

free analysis results, MD-derived dynamics parameters, amide N-H bound local motions in MD

simulations, ∆pre-A R2 data, ∆pre-A 1H-

15N HSQC overlay, ∆pre-A chemical shift changes. This

material is available free of charge via the Internet at http://pubs.acs.org and partially described

in Annexe 3.

http://pubs.acs.org/

180

8.10. References

1. Brown, G. C., and Cooper, C. E. (1994) Nanomolar concentrations of nitric-oxide reversibly

inhibit synaptosomal respiration by competing with oxygen at cytochrome-oxidase, FEBS

Lett. 356, 295-298.

2. Brunori, M. (2001) Nitric oxide, cytochrome-c oxidase and myoglobin, Trends Biochem.Sci.

26, 21-23.

3. Brunori, M., Giuffre, A., Sarti, P., Stubauer, G., and Wilson, M. T. (1999) Nitric oxide and

cellular respiration, Cell. Mol. Life Sci. 56, 549-557.

4. Cleeter, M. W. J., Cooper, J. M., Darleyusmar, V. M., Moncada, S., and Schapira, A. H. V.

(1994) Reversible inhibition of cytochrome-c oxidase, the terminal enzyme of the

mitochondrial respiratory-chain, by nitric-oxide - Implications for the neurodegenerative

diseases, FEBS Lett. 345, 50-54.

5. Gardner, P. R., Costantino, G., and Salzman, A. L. (1998) Constitutive and adaptive

detoxification of nitric oxide in Escherichia coli - Role of nitric-oxide dioxygenase in the

protection of aconitase, Journal of Biological Chemistry 273, 26528-26533.

6. Gardner, P. R., Costantino, G., Szabo, C., and Salzman, A. L. (1997) Nitric oxide sensitivity

of the aconitases, Journal of Biological Chemistry 272, 25071-25076.

7. Stevanin, T. M., Ioannidis, N., Mills, C. E., Kim, S. O., Hughes, M. N., and Poole, R. K.

(2000) Flavohemoglobin hmp affords inducible protection for Escherichia coli respiration,

catalyzed by cytochromes bo ' or bd, from nitric oxide, Journal of Biological Chemistry 275,

35868-35875.

8. Ouellet, H., Ouellet, Y., Richard, C., Labarre, M., Wittenberg, B., Wittenberg, J., and

Guertin, M. (2002) Truncated hemoglobin HbN protects Mycobacterium bovis from nitric

oxide, Proc. Natl. Acad. Sci. U. S. A. 99, 5902-5907.

9. Via, L. E., Lin, P. L., Ray, S. M., Carrillo, J., Allen, S. S., Eum, S. Y., Taylor, K., Klein, E.,

Manjunatha, U., Gonzales, J., Lee, E. G., Park, S. K., Raleigh, J. A., Cho, S. N., McMurray,

D. N., Flynn, J. L., and Barry, C. E., 3rd. (2008) Tuberculous granulomas are hypoxic in

guinea pigs, rabbits, and nonhuman primates, Infect Immun 76, 2333-2340.

10. Daigle, R., Guertin, M., and Lague, P. (2009) Structural characterization of the tunnels of

Mycobacterium tuberculosis truncated hemoglobin N from molecular dynamics simulations,

Proteins 75, 735-747.

11. Milani, M., Pesce, A., Ouellet, Y., Dewilde, S., Friedman, J., Ascenzi, P., Guertin, M., and

Bolognesi, M. (2004) Heme-ligand tunneling in group I truncated hemoglobins, Journal of

Biological Chemistry 279, 21520-21525.

181

12. Daigle, R., Rousseau, J. A., Guertin, M., and Lague, P. (2009) Theoretical Investigations of

Nitric Oxide Channeling in Mycobacterium tuberculosis Truncated Hemoglobin N, Biophys.

J. 97, 2967-2977.

13. Golden, S. D., Olsen, K. W., and Robert, K. P. (2008) Identification of Ligand Binding

Pathways in Truncated Hemoglobins Using Locally Enhanced Sampling Molecular

Dynamics, Methods in Enzymology 437, 459-475.

14. Milani, M., Ouellet, Y., Ouellet, H., Guertin, M., Boffi, A., Antonini, G., Bocedi, A., Mattu,

M., Bolognesi, M., and Ascenzi, P. (2004) Cyanide binding to truncated hemoglobins: a

crystallographic and kinetic study, Biochemistry 43, 5213-5221.

15. Milani, M., Pesce, A., Ouellet, Y., Ascenzi, P., Guertin, M., and Bolognesi, M. (2001)

Mycobacterium tuberculosis hemoglobin N displays a protein tunnel suited for O2 diffusion

to the heme, EMBO J 20, 3902-3909.

16. Lama, A., Pawaria, S., Bidon-Chanal, A., Anand, A., Gelpi, J. L., Arya, S., Marti, M., Estrin,

D. A., Luque, F. J., and Dikshit, K. L. (2009) Role of Pre-A motif in nitric oxide scavenging

by truncated hemoglobin, HbN, of Mycobacterium tuberculosis, J Biol Chem 284, 14457-

14468.

17. Scott, N. L., and Lecomte, J. T. (2000) Cloning, expression, purification, and preliminary

characterization of a putative hemoglobin from the cyanobacterium Synechocystis sp. PCC

6803, Protein Sci. 9, 587-597.

18. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., and Bax, A. (1995) NMRPipe:

a multidimensional spectral processing system based on UNIX pipes, J Biomol NMR 6, 277-

293.

19. Johnson, B. A., and Blevins, R. A. (1994) NMRView: A computer program for the

visualization and analysis of NMR data, J Biomol NMR 4, 603-614.

20. Farrow, N. A., Muhandiram, R., Singer, A. U., Pascal, S. M., Kay, C. M., Gish, G., Shoelson,

S. E., Pawson, T., Forman-Kay, J. D., and Kay, L. E. (1994) Backbone dynamics of a free

and phosphopeptide-complexed Src homology 2 domain studied by 15

N NMR relaxation,

Biochemistry 33, 5984-6003.

21. Kay, L. E., Keifer, P., and Saarinen, T. (1992) Pure absorption gradient enhanced

heteronuclear single quantum correlation spectroscopy with improved sensitivity, Journal of

the American Chemical Society 114, 10663-10665.

22. Lipari, G., and Szabo, A. (1982) Model-free approach to the interpretation of nuclear

magnetic resonance relaxation in macromolecules. - Theory and range of validity, Journal of

the American Chemical Society 104, 4546-4559.

182

23. Lipari, G., and Szabo, A. (1982) Model-Free Approach to the Interpretation of Nuclear

Magnetic Resonance Relaxation in Macromolecules. - Analysis of Experimental Results,

Journal of the American Chemical Society 104, 4559-4570.

24. Morin, S., and Gagne, S. M. (2009) NMR dynamics of PSE-4 beta-lactamase: an interplay of

ps-ns order and µs-ms motions in the active site, Biophys J 96, 4681-4691.

25. MacKerell, A. D., Bashford, D., Bellott, M., Dunbrack, R. L., Evanseck, J. D., Field, M. J.,

Fischer, S., Gao, J., Guo, H., Ha, S., Joseph-McCarthy, D., Kuchnir, L., Kuczera, K., Lau, F.

T. K., Mattos, C., Michnick, S., Ngo, T., Nguyen, D. T., Prodhom, B., Reiher, W. E., Roux,

B., Schlenkrich, M., Smith, J. C., Stote, R., Straub, J., Watanabe, M., Wiorkiewicz-Kuczera,

J., Yin, D., and Karplus, M. (1998) All-Atom Empirical Potential for Molecular Modeling

and Dynamics Studies of Proteins, Journal of Physical Chemistry B 102, 3586-3616.

26. Brooks, B. R., Bruccoleri, R. E., Olafson, B. D., States, D. J., Swaminathan, S., and Karplus,

M. (1983) CHARMM: A program for macromolecular energy, minimization, and dynamics

calculations, Journal of Computational Chemistry 4, 187-217.

27. Fisette, O., Morin, S., Savard, P.-Y., Lague, P., and Gagne, S. M. (2010) TEM-1 backbone

dynamics-insights from combined molecular dynamics and nuclear magnetic resonance,

Biophys. J. 98, 637-645.

28. d'Auvergne, E. J., and Gooley, P. R. (2008) Optimisation of NMR dynamic models II. A new

methodology for the dual optimisation of the model-free parameters and the Brownian

rotational diffusion tensor, J Biomol NMR 40, 121-133.

29. d'Auvergne, E. J., and Gooley, P. R. (2008) Optimisation of NMR dynamic models I.

Minimisation algorithms and their performance within the model-free and Brownian

rotational diffusion spaces, J Biomol NMR 40, 107-119.

30. Assfalg, M., Banci, L., Bertini, I., Ciofi-Baffoni, S., and Barker, P. D. (2001) (15)N

backbone dynamics of ferricytochrome b(562): comparison with the reduced protein and the

R98C variant, Biochemistry 40, 12761-12771.

31. Volkman, B. F., Alam, S. L., Satterlee, J. D., and Markley, J. L. (1998) Solution structure

and backbone dynamics of component IV Glycera dibranchiata monomeric hemoglobin-CO,

Biochemistry 37, 10906-10919.

32. Bloom, M., Reeves, L. W., and Wells, E. J. (1965) Spin Echoes and Chemical Exchange, J.

Chem. Phys. 42.

33. Reddy, T., and Rainey, J. K. (2010) Interpretation of biomolecular NMR spin relaxation

parameters, Biochem Cell Biol 88, 131-142.

183

34. Goodman, J. L., Pagel, M. D., and Stone, M. J. (2000) Relationships between protein

structure and dynamics from a database of NMR-derived backbone order parameters, J Mol

Biol 295, 963-978.

35. Bidon-Chanal, A., Marti, M. A., Crespo, A., Milani, M., Orozco, M., Bolognesi, M., Luque,

F. J., and Estrin, D. A. (2006) Ligand-induced dynamical regulation of NO conversion in

Mycobacterium tuberculosis truncated hemoglobin-N., Proteins 64, 457-464.

36. Marsh, J. A., Singh, V. K., Jia, Z., and Forman-Kay, J. D. (2006) Sensitivity of secondary

structure propensities to sequence differences between alpha- and gamma-synuclein:

implications for fibrillation, Protein Sci 15, 2795-2804.

37. Baron, R., Riley, C., Chenprakhon, P., Thotsaporn, K., Winter, R. T., Alfieri, A., Forneris, F.,

van Berkel, W. J., Chaiyen, P., Fraaije, M. W., Mattevi, A., and McCammon, J. A. (2009)

Multiple pathways guide oxygen diffusion into flavoenzyme active sites, Proc Natl Acad Sci

U S A 106, 10603-10608.

38. Chen, L., Lyubimov, A., Brammer, L., Vrielink, A., and Sampson, N. (2008) The Binding

and Release of Oxygen and Hydrogen Peroxide Are Directed by a Hydrophobic Tunnel in

Cholesterol Oxidase., Biochemistry 47, 5368-5377.

39. Leroux, F., Dementin, S., Burlat, B., Cournac, L., Volbeda, A., Champ, S., Martin, L.,

Guigliarelli, B., Bertrand, P., Fontecilla-Camps, J., Rousset, M., and Léger, C. (2008)

Experimental approaches to kinetics of gas diffusion in hydrogenase., Proc Natl Acad Sci U

S A 105, 11188-11193.

40. Saam, J., Ivanov, I., Walther, M., Holzhütter, H.-G., and Kuhn, H. (2007) Molecular

dioxygen enters the active site of 12/15-lipoxygenase via dynamic oxygen access channels.,

Proc Natl Acad Sci U S A 104, 13319-13324.

41. Tomita, A., Sato, T., Ichiyanagi, K., Nozawa, S., Ichikawa, H., Chollet, M., Kawai, F., Park,

S. Y., Tsuduki, T., Yamato, T., Koshihara, S. Y., and Adachi, S. (2009) Visualizing

breathing motion of internal cavities in concert with ligand migration in myoglobin, Proc

Natl Acad Sci U S A 106, 2612-2616.

42. Ouellet, Y., Milani, M., Couture, M., Bolognesi, M., and Guertin, M. (2006) Ligand

interactions in the distal heme pocket of Mycobacterium tuberculosis truncated hemoglobin

N: roles of TyrB10 and GlnE11 residues, Biochemistry 45, 8770-8781.

43. Yeh, S. R., Couture, M., Ouellet, Y., Guertin, M., and Rousseau, D. L. (2000) A cooperative

oxygen binding hemoglobin from Mycobacterium tuberculosis. Stabilization of heme ligands

by a distal tyrosine residue, J Biol Chem 275, 1679-1684.

44. Trent, J. T., Kundu, S., Hoy, J. A., and Hargrove, M. S. (2004) Crystallographic analysis of

synechocystis cyanoglobin reveals the structural changes accompanying ligand binding in a

hexacoordinate hemoglobin., J Mol Biol 341, 1097-1108.

184

45. Vu, B. C., Nothnagel, H. J., Vuletich, D. A., Falzone, C. J., and Lecomte, J. T. (2004)

Cyanide binding to hexacoordinate cyanobacterial hemoglobins: hydrogen-bonding network

and heme pocket rearrangement in ferric H117A Synechocystis hemoglobin, Biochemistry

43, 12622-12633.

46. Falzone, C. J., Christie Vu, B., Scott, N. L., and Lecomte, J. T. (2002) The solution structure

of the recombinant hemoglobin from the cyanobacterium Synechocystis sp. PCC 6803 in its

hemichrome state, J Mol Biol 324, 1015-1029.

47. Hoy, J. A., Kundu, S., Trent, J. T., Ramaswamy, S., and Hargrove, M. S. (2004) The crystal

structure of Synechocystis hemoglobin with a covalent heme linkage., J Biol Chem 279,

16535-16542.

48. Morin, S., and S, M. G. (2009) Simple tests for the validation of multiple field spin relaxation

data, J Biomol NMR 45, 361-372.

185

Table 8.1 Average R1 and R2 relaxation rates (s-1

) and {1H}-

15N NOEs at 500, 600, and 800

MHz.

500 MHz 600 MHz 800 MHz

15N-R1 2.05 ± 0.16 1.22 ± 0.11 0.93 ± 0.12

15N-R2 12.43 ± 2.38 13.36 ± 2.76 Not Used*

{1H}-

15N NOE 0.66 ± 0.37 0.70 ± 0.31 0.77 ± 0.21

Mean value with associated SD for all the data available at three fields (101 residues).

*R2 acquired at 800 MHz were not used in our analysis as they were shown by consistency tests

(48) to be R2 from both 500 and 600 MHz. This was done in order to prevent artefactual

conclusions to be drawn from the data. More details are available in Supporting


186

Figure 8.1 Structure of TrHbN displaying the four tunnels: Long tunnel (LT), Short tunnel (ST),

EH tunnel, and BE tunnel.

187

Figure 8.2 (A) Assigned 1H-

15N HSQC spectrum of TrHbN cyanomet, 0.7 mM, pH 7.5, acquired

at 600 MHz, 26.4°C. (B) Zoom of the most crowded region of the spectrum.

188

Figure 8.3 (Legend on next page)

189

Figure 8.3 (A-D) NMR raw relaxation data (R1, R2, R2/R1, NOE) at 500 (black solid circles), 600

(grey solid squares), and 800 (black solid triangles) MHz. (E-G) Model-free parameters (S2, Rex,

and τe) for TrHbN cyanomet. (H-I) Comparision of S2 parameters obtained either from NMR

(white solid squares) or MD simulations (black solid circles). (J-K) NMR amide exchange data:

amide exchanges rates (kex) at pH 7.5 (blue circles) and 8.5 (red squares), and free energy for the

opening of the protecting structure (ΔGHX) at pH 7.5. (L-M) Molecular Dynamics data: Average

backbone ASA and Backbone amide hydrogen bond occupancy calculated with various cutoffs

for the N-H acceptor bond distance. Cutoffs used are 2.0 Å (red), 2.2 Å (green), and 2.4 Å

(blue). (N-P) Secondary structure of TrHbN calculated by NMR (HN, N, CA, CB chemical

shifts used in the program SSP), MD, or taken from the X-ray structure PDB 1S61B (11).

190

Figure 8.4. Mapping of NMR experimental results on the structure of TrHbN. (A) Mapping of

the S2 values. (Figure legend on next page)

191

Figure 8.4 Mapping of NMR experimental results on the structure of TrHbN. (A) Mapping of the

S2 values. (B) Mapping of residues exhibiting slow motions (Rex term, model-free models m3

and m4) (red), two timescale motions (S2

f and S2s, m5) (blue) or fitted with simple models (S

2 or

S2-τe) (white). Gray residues represent residues for which no data were avalaible. (C) Mapping of

the free energy for the opening of the protecting structure (ΔGHX). Residues for which data is

unavailable (N-terminus, Pro, unassigned, ambiguous and overlapped) are colored gray, while

residues with exchange rates too fast (kex > 10-3

s-1

) and too slow (kex < 10-9

s-1

) to be

characterized are colored either blue or red.

192

Figure 8.5 Mapping of the chemical shift differences between the wild type and the Δpre-A

mutant. The cleaved region is shown in green.

193

9.

Chapitre 9

Experimental and Theoretical Investigations Reveal that

Mycobacterium tuberculosis Truncated Hemoglobin N

contains Multiple Diffusion Routes to Sustain Rapid

Gaseous Ligand Entry and Exit.

9.1. Résumé

L’hémoglobine tronquée N de Mycobacterium tuberculosis (TrHbN) catalyse efficacement

l’oxydation du •NO en nitrate selon la constante bimoléculaire k´NOD ≈ 745 × 106 M

-1·s

-1, soit

près de 15 fois plus efficace que la même réaction catalysée par la Mb extraite du myocarde du

cheval. Nous avons tenté d’identifier quels sont les aspects de la structure et/ou de la dynamique

de TrHbN lui conférant une telle réactivité. De récentes simulations de dynamique moléculaire

sur TrHbN sous sa forme oxygénée ont montré que le •NO peut accéder le site actif à travers

trois tunnels hydrophobes nommés LT, ST et EHT. Par contre, il n’existe encore aucune preuve

expérimentale supportant le rôle des différents tunnels chez TrHbN. Dans le présent travail, nous

avons construit des mutants dans lesquels un ou plusieurs tunnels sont obstrués près de la surface

de la protéine par des acides aminés de grande taille. Nous avons mesuré l’activité NOD de ces

mutants. Nos résultats ont montré une baisse maximale de l’activité dans le triple mutant, lequel

ayant les trois routes obstruées, confirmant que chacune d’elle est fonctionnelle pour la diffusion

des ligands. De plus, des mesures cinétiques sur le produit photodissocié de la forme met-•NO

révèlent une phase de recombinaison très rapide (nsec) qui n’est pas observée chez la forme

sauvage et qui est maximale chez le triple mutant. Pour tenter d’expliquer l’activité NOD

résiduelle mesurée dans le triple mutant (43% de celle de wt-TrHbN), des simulations de

dynamique moléculaire ont été réalisées en absence ou en présence de •NO. Ces simulations

révèlent que 1) l’ouverture des tunnels survient à des endroits proches de ceux observés chez wt-

TrHbN; 2) des routes non détectées auparavant se sont formées et que 3) des molécules de •NO

peuvent modifier la dynamique des chaînes latérales de manière à favoriser leur entrée et sortie.

194

Ensemble, nos résultats mettent l’emphase sur le rôle crucial des chaînes latérales et de leur

dynamique dans la diffusion des ligands chez TrHbN.

9.2. Abstract

Oxygenated Mycobacterium tuberculosis truncated hemoglobin N (TrHbN) catalyzes the rapid

oxidation of •NO to innocuous nitrate with a second-order rate constant

k´NOD ≈ 745 × 106 M

-1·s

-1, which is 15-fold faster than the reaction of horse heart Mb. We asked

what aspects of TrHbN structure and/or dynamics give rise to this enhanced activity. Recent

molecular dynamics simulations of TrHbN in the oxygenated form showed that •NO can access

the active site trough three hydrophobic tunnels termed LT, ST and EHT. However, no

experimental evidence has been provided yet to support the role of the different tunnels in

TrHbN. In the present work we constructed mutants in which one or more tunnels were

obstructed with bulky amino acids. We measured the NOD activity of these mutants. Our results

showed a maximal decrease in NOD activity in the triple mutant, with the three tunnels blocked,

confirming that all tunnels constitute a functional route for ligand diffusion. Furthermore, kinetic

measurements of the photoproduct of the •NO derivative of met-TrHbN mutants revealed a ns

geminate binding phase not observed in the wild-type protein, with maximal amplitude in the

triple mutant. To explain the residual extent of NOD reaction measured in the triple mutant (43%

of wt-TrHbN), MD simulations were performed in absence or presence of •NO. Data revealed

that: 1) opening of the tunnels occurred at sites close to those observed in wt-TrHbN; 2)

previously undetected pathways were formed and 3) •NO molecules could alter side-chain

dynamics and so favor their entry/exit. Altogether our results emphasize the crucial role of side-

chains dynamics in ligand diffusion in TrHbN.

195

9.3. Introduction


inactivating key enzymes such as the terminal respiratory oxidases [1-5] and the iron/sulfur

protein aconitase [6,7].

•NO also combines at near diffusion-limited rate with superoxide

produced by respiring cells to form the highly oxidizing agent peroxynitrite [8, 9]. •NO-


The truncated hemoglobin TrHbN from the pathogenic bacterium Mycobacterium tuberculosis

has a potent ability to detoxify •NO to nitrate (nitric oxide dioxygenase (NOD reaction)) and to

protect aerobic respiration from inhibition by •NO in stationary phase cells of M. bovis BCG.

TrHbN catalyses the rapid oxidation of •NO to innocuous nitrate (TrHbN-FeII(O2) + •NO →

TrHbN-FeIII

+ NO3-), with a second-order rate constant k´NOD ≈ 745 µM

-1 s

-1 (23 °C) [10], which

is more than one order of magnitude faster than in horse heart Mb and sperm whale Mb [10,11]

and close to that of a diffusion-limited reaction [12]. In this context, the aspects of the structure

and/or dynamics giving rise to this enhanced reactivity become a critical issue. A first step is to

expose the molecular mechanisms controlling ligand/substrate access to the DHP.

The first indications came from crystallographic studies of TrHbN which revealed the presence

of two hydrophobic tunnels, termed the Short (ST) and the Long (LT) tunnels, connecting the

DHP to distinct protein surface sites [13,14] (Fig. 9.1). These tunnels were proposed to constitute

gas diffusion pathways. In support, saturation of TrHbNFeIII

(CN-) crystals with xenon led to the

identification of five xenon binding sites (Xe sites) along these pathways (PDB 1S56) [15]. Two

binding sites are located along the LT (Xe1, Xe5), one at the ST entrance (Xe3), one where LT

and ST are merging (Xe2) and finally one site (Xe4) located outside the protein core at 5.4 Å of

Xe3. Of interest, Xe2 is the only site communicating with the DHP.

Afterwards, MD simulations with the oxygenated and deoxygenated form of TrHbN showed that

tunnels are not permanently open nor static features, but rather result from the dynamical

reshaping of hydrophobic cavities that temporarily interconnect due to side-chain flexibility [16].

In addition, two other tunnels were observed; the EH (EHT) and BE (BET) tunnels.

196

More recently, different simulation strategies have been employed to map the complete network

and energy profile of gas migration inside TrHbN. Mishra and Meuwly performed 24×2-ns

simulations of TrHbN containing one •NO molecule placed in each of the Xe sites as well as

inside another cavity located on the proximal side of the heme [17]. In contrast, Daigle et al.

performed 10×20-ns simulations of TrHbN(FeIIO2) including 10 •NO molecules starting in the

bulk solvent [18]. The latter study revealed that •NO interacts with TrHbN at specific surface

sites at tunnel entrances, composed of hydrophobic residues. Once inside the protein, •NO

diffuses rapidly from one cavity to another. Of interest, most of these cavities correspond to Xe

sites. The trajectories also indicated that •NO alters the dynamics of Ile119(H11) residue,

favoring the adoption of rotamers promoting ST opening. Similarly, Mishra and Meuwly

reported •NO diffusion through the ST, LT and EHT [17]. They also identified a new route

located between G and H helices, allowing communications with various cavities and the

solvent. This route is hereafter referred as the GH tunnel (GHT) (Fig. 9.1).

Altogether, these observations suggest that TrHbN tunnels have been specifically tailored to

afford efficient gas capture from solvent as well as to provide direct and efficient diffusion of

gases to the DHP. As a consequence, the bound O2 may remain optimally oriented and stabilized

by the active site residues Tyr(B10) and Gln(E11) for reaction with •NO [16].

Apolar tunnels or cavities are also observed in several other structures of proteins that bind the

gaseous molecules O2, •NO or H2 or catalyze reactions involving these gases [19-28]. In some

cases, these tunnels have also been saturated with xenon to demonstrate that a gas can diffuse

through a specific tunnel or multiple tunnels.

To this day, few studies have combined computational and experimental approaches to study the

functionality of channels in proteins. In enzymes where a single specific channel is observed,

mutagenesis data support a functional role [22,23, 29-33]. However, in proteins with multiple

predicted channels the conclusions differ depending on the enzyme investigated. Notably, in

12/15-lipoygenase only one path out of three predicted by implicit ligand sampling (ILS)

calculations was found to be effective in O2 delivery to the active site, the other two being

occupied by the substrate linoleic acid [22]. In copper-containing amine oxidase (COA) two

possible major pathways have been identified by ILS calculation [21]. These two paths merge

197

near the active site. Mutants intended to block individual routes for O2 in COA, resulted in little

perturbation of the kcat/ km ratios. These results were taken as evidence of the existence of other

multiple dynamic pathways [21]. Several channels have been observed in Cytochrome c oxidase

(CcO). One is hydrophobic and was predicted to serve for ligand diffusion. The other ones are

hydrophilic and were proposed to serve for water diffusion and proton transfer. In contrast to

COA a single mutation gly→val intended to block the hydrophobic tunnel slowed O2 and CO

binding by several orders of magnitude [23]. To explain this effect, the authors postulated that

CcO structure should be highly rigid near to mutation site, preventing formation of alternative

routes. Furthermore, this work demonstrated that only one major gas diffusion route exits in

CcO, the hydrophilic routes being obstructed by water molecules.

In the present work we investigated the NOD reaction catalyzed by TrHbN mutants with the LT

(Ala24(B1)Leu), ST (Ala95(G9)Ile) or EHT (Ala65(E18)Ile) entrance obstructed. Kinetic data

indicated that each tunnel can deliver •NO to the active site. Attempts to completely block the

access to the DHP as in the triple mutant ST/LT/EHT, resulted in a 3-fold reduction of NOD

reaction rate. In agreement, kinetic measurements on the photoproduct of the •NO derivative of

met-TrHbN mutants revealed a geminate binding phase not observed in the wild-type protein,

that was greatest in the triple mutant.

To understand residual NOD activity in the triple mutant, theoretical investigations were

performed. MD simulations revealed that 1) LT, ST and EHT were partially blocked due to side-

chain flexibility; 2) other pathways were formed and 3) •NO could modify side-chains dynamics

so as to favor tunnels opening. Our results emphasize the importance of the flexibility of side

chain in ligand diffusion.

9.4. Experimental procedures

Design of the mutations – Mutations were performed to close the entrance of the ST, LT or EHT

tunnels. Substitutions were designed so as to preserve the hydrophobic character and the

topology of the entrances, and to minimize changes in the dynamics of tunnels and of the DHP.

198

For LT, the entrance is mainly defined by backbone from the A-B and G-H loops. The only

available residue for mutagenesis is Ala24(B1), which was replaced by leucine. Entrances of ST

and EHT share common characteristics. They are found between two helices and are shaped by

side-chains from four hydrophobic residues. For the ST, only Ala95(G9) was found correctly

oriented for efficient blocking while for the EHT, only Ala65(E18) was correctly oriented. For

the ST and EHT, the targeted residue was replaced by an isoleucine. According to the crystal

structure and preliminary MD simulations, these mutations were found to optimally fill the

tunnel space and to avoid clashes with other residues.

Mutagenesis, expression and purification – Amino acid substitutions were carried out using the

QuickChange Site-Directed Mutagenesis kit (Stratagene) following the recommended protocol. Details on

mutagenesis are provided in Supporting Material (Annexe 4). The expression and purification of the

recombinant proteins were performed in accordance with the previously published method [34].

NOD reaction – NOD reactions were measured by stopped-flow spectrophotometry under single

turnover conditions as previously described [10]. Complete protocol is available in Supporting

Material (Annexe 4). The results shown in figures 9.2, 9.3, S1 and S2 (Annexe 4) are

representative of at least two experiments.

Flash-photolysis experiments – Laser flash-photolysis studies of the ferric NO complexes of the

different proteins were carried out as previously described [34]. An average of at least ten kinetic

traces from at least two separate experiments were averaged and analyzed with the instrument

manufacturer software (Applied Photolysis, U.K.) to obtain the rate constants. The fraction of

geminate rebinding was calculated as described in [35]. Plots showing absorbance changes

following •NO photolysis were obtained using the KaleidaGraph software (Synergy Software,

USA).

MD Simulations - The structure and dynamics of the ST/LT/EHT mutant, under the FeIIO2 form,

was studied by performing a 30 ns MD simulation. Simulations were performed using

CHARMM [36]. Simulations were performed as described in [16]. A complete description of the

simulation protocol is given as Supporting Material (Annexe 4).

System setup - The coordinates of the mutant were built from an equilibrated MD frame of wild-

type oxygenated TrHbN. Missing coordinates were built using the internal coordinates definition

199

of CHARMM. The first 5 ns were considered as the equilibration phase giving 25 ns in

production mode. The analysis of this simulation was performed in comparisons with two 30 ns

MD simulations of TrHbN presented earlier [16].

Evaluation of tunnel entrance openings – The method used to evaluate the opening of the

tunnel entrances is given as Supporting Material (Annexe 4).

Locally Enhanced Sampling – The locally enhanced sampling method (LES) [37] was

employed in this work to find out the more likely NO diffusion pathways in TrHbN and the triple

mutant. At the same time, these simulations allow to verify if •NO themselves have local impacts

on protein dynamics. This simulation technique allows replication of a given part of the system

to improve sampling. In this work, a •NO molecule was replicated. Each replicate is invisible to

each other but all of them interact with the rest of the system. The latter interactions are scaled

by a factor 1/N, where N is the number of replicates used. This factor, easing barrier crossing,

coupled with the high number of replicates favor the sampling. Ten 10-ns simulations, each

including ten •NO replicates, were run for both TrHbN and the triple mutant. In these

simulations, •NO molecules started in the DHP. Each system was built using snapshots from MD

simulations at equilibrium of the TrHbN or the triple mutant. The number of water molecule 5 Å

around the center of mass of each •NO was used as a probe to detect entry and exit events along

the trajectories. Each event was then studied visually using PyMOL [38].

Implicit ligand sampling – The •NO potential of mean force (PMF) inside TrHbN and the triple

mutant was determined using the implicit ligand sampling (ILS) method [39]. Details about

ligand parameters are given in [18]. ILS methods tends to overestimate PMF levels as they

increase, a tendency than can be counterbalanced with a higher sampling (see equation 12 from

ref [39]). For this reason, high energy barrier (PMF levels > 5 kcal/mol) could not be evaluated

precisely. PMF levels were evaluated using a sampling of all 25 000 MD frames. All PMF plots

are available in Supporting Material (Annexe 4, Fig. S5).

200

9.5. Results

9.5.1. NOD reaction of TrHbN.

(a) Kinetics of the reaction and comparison with reaction of Myoglobin and Leghemoglobin.

No intermediate was detected when TrHbNFeII(O2) was reacted with equimolar •NO at 23ºC, pH

7.5, the oxy complex being converted to aquomet TrHbNFeIII

(OH2) during the mixing time.

Similar results have been reported when HbAFeII(O2), sperm whale MbFe

II(O2) or plant

leghemoglobin LbFeII(O2) are reacted with equimolar •NO at 23 ºC, pH 7.5 [40, 41]. In contrast,

an intermediate species with UV-Vis spectral properties similar to that for a ferric high-spin

species could be detected when HbAFeII(O2) or MbFe

II(O2) were reacted under alkaline

conditions (pH 9.5). Fig. S1a (Annexe 4)shows the optical changes using our data obtained with

horse heart MbFeII(O2) (5 M) reacted with 5 M •NO at 5 ºC, pH 9.5, while Fig. S1b

(Annexe 4) depicts the first spectrum (417, 544 and 580 nm) recorded after mixing (1.3 ms).

This spectrum has maxima that are very close to those of MbFeII(O2) under equilibrium

conditions (418.6 , 544 and 581 nm) indicating that only a small fraction of the protein had

reacted during mixing time. Singular value decomposition (SVD) and global analysis allowed

fitting of the kinetic data to the model ABC. (Fig. S1c (Annexe 4), Table 9.1) UV-Vis

spectrum for each species is shown in Fig S1d. Species A, corresponding to MbFeII(O2), decays

rapidly (145 ± 3.6 s-1

) to an intermediate species B showing maxima at 411.9, 502, 540, 580 and

634 nm. This spectrum is similar in the visible region to that reported for HbA [40] and indicates

a ferric species with a pronounced LS character distinctly different from that of ferric Mb at pH

7.5 under equilibrium conditions. The ferric HS intermediate decays to the FeIII

(OH-)-form

(species C) at a rate of 27.3 ± 0.54 s-1

. Recent resonance Raman analysis of rapid freeze

quenched samples of the reaction of MbFeII(O2) + •NO at pH 9.5 (3 °C) identified the HS-

intermediate as a ferric nitrato complex (FeIII

(ONOO-) [42]. It is worth mentioning that the

optical spectrum of the nitrato-intermediate in Mb is different from that observed under

equilibrium conditions with a large excess of nitrate at pH 7.5 and likely represents a transient

FeIII

(ONOO-) complex. In contrast, the reaction of oxy plant leghemoglobin with •NO yields the

201

FeIII

(OH-)-form without the detection of an intermediate even under alkaline conditions (pH

9.5) [41].

We also examined the NOD reaction in TrHbN under alkaline conditions. Fig. 9.2a depicts the

evolution of the optical spectra acquired during the first 250 ms after mixing TrHbNFeII(O2) with

equimolar •NO at 5 ºC and pH 9.5. In contrast to Mb, the first spectrum obtained 1.3 ms after

mixing bears the signature of a HS species with peaks at 630 and 500 nm (Fig. 9.2b) indicating

that most of the oxy complex has reacted during the mixing time. SVD and global analysis

indicated that three optical species were necessary to fit the kinetic data (ABC) where

species A is TrHbNFeII(O2) (Fig. 9.3a and 9.3b; Table 9.1). Accordingly, TrHbNFe

II(O2)

decays very rapidly (1912 ± 54.2 s-1

) to a ferric HS intermediate (species B) showing maxima at

406, 503, 533 (sh), 570 (sh) and 635 nm. The associated rate is 14.4 fold faster than that

measured for horse heart Mb and is close to the value (16.2-fold) calculated previously from the

overall rate constants measured for these two proteins under pseudo first-order at pH 7.5 and 23

°C [10]. The HS intermediate decays to the FeIII

(OH-)-form expected at this pH (species C) (411,

543, 576 and 607 (sh)) at a rate of 22.72 ± 0.76 s-1

(Fig. 9.4 and Table 9.1). The optical spectrum

of the HS intermediate is very similar to that of TrHbNFeIII

(OH2) or TrHbNFeIII

(ONOO-)

complexes at pH 7.5 under equilibrium conditions (406, 502, 542 (sh), 581 (sh) and 632 nm), the

latter’s being almost identical. We propose that the observed HS species at pH 9.5 may be either

the TrHbNFeIII

(ONOO-) as in Mb or TrHbNFe

III(OH2) complex. In agreement, theoretical

calculations by Mishra and Meuwly [43] predicted a reaction time too fast to be measured by

stopped-flow experiments. However, additional investigations are required to determine whether

the intermediate at pH 9.5 is the nitrato or the aquo form.

(b) The ST, EHT and LT all deliver •NO to the heme bound O2 in the NOD reaction. To

investigate the role of the tunnels in the NOD reaction TrHbN, mutants with the LT, ST and

EHT entrance obstructed were produced (see experimental procedures). UV-Vis optical spectra

and resonance Raman spectra of the different mutants were identical to those of the TrHbN,

which indicated that blocking the different entrances did not perturb the heme bound O2

complexes (unpublished results). Reactions took place upon turnover of one stoichiometric

equivalent of •NO under alkaline conditions at 5 ºC as for TrHbN.

202

Three optical species were found necessary to fit kinetic data of all TrHbN mutants with species

A being assigned to their respective oxy forms (Fig. 9.3 a, c and e, only WT, ST/LT and

ST/LT/EHT spectra and kinetic traces are depicted for the sake of clarity). In all cases, the

optical spectrum of species B was nearly identical to the intermediate in TrHbN indicating that

mutations did not perturb the reaction. As shown in Table 9.1, the rates associated with the

decay of the oxygenated complexes (k1) indicated that •NO access to the DHP was impeded in

all mutants and that the extent of inhibition increased as more tunnels were blocked. In

agreement, the spectrum at 1.3 ms in the ST/LT/EHT mutant had a significant LS character (409,

544 and 582 nm) relative to that of TrHbN (Fig. S2), indicating a mixture of TrHbNFeII(O2)

(416, 545 and 582 nm) and the HS intermediate. The remaining activity measured in the triple

mutant is noteworthy and suggests that tunnels are either partially blocked or that other pathways

have been used to access the DHP. The lower k2 values in LT mutants (Table 9.1), indicate that

LT could play a role in the exit of intermediate (H2O or NO3-) or an entryway for the hydroxyl

anion.

9.5.2. Mutants with obstructed tunnel entrance(s) show ns geminate rebinding

of the •NO.

We recently used flash photolysis of ferric TrHbN bound with •NO to study the kinetics of water

and •NO binding to the heme iron in TrHbN [34]. Photodissociation of the FeIII

(•NO) complex

leaves the heme distal site in a ferric dehydrated state 5c (FeIII

) [43]. After •NO escape, a water

molecule enters the DHP and binds very rapidly (≤ 10 ns) to the heme iron forming the 6c HS

aquomet state FeIII

(OH2). Subsequently, on a much longer time scale, a •NO coming from the

solvent displaces the bound water molecule and combines with the iron atom to form the

FeIII

(•NO) species. Fig. 9.4 shows the normalized kinetic traces obtained for TrHbN, ST, LT/ST

and ST/LT/EHT mutants when the reaction is monitored at 392 nm (23 °C) the wavelength

corresponding to the isosbestic point of the optical spectra of the 5c (FeIII

) and FeIII

(OH2). The

increase in absorbance corresponds to the formation of the 5c (FeIII

) species. In TrHbN, this

initial kinetic phase is followed by a plateau that is attributed to the long lived FeIII

(OH2) species

(the formation of the FeIII

(OH2) species cannot be measured at this wavelength). The bound

203

water is eventually displaced by •NO coming from the bulk solvent leading to a decrease in

absorbance (not shown; (see ref [34]). In contrast, all mutants, except the LT, showed an

additional kinetic phase (very rapid decrease in absorbance (≤ 10 ns)), which represents geminate

rebinding of •NO. The fraction of geminate recombination, Fgem increased markedly in going

from ST (0.43), ST/LT (0.49) to the ST/LT/EHT (0.75) mutant. The Fgem for TrHbN and LT

mutant could not be measured indicating that the rate of ligand escape must be much larger than

the rate of internal bond formation to the 5c (FeIII

). Similar observations were made when CO

recombination was measured subsequent to photodissociation of TrHbNFeII(CO) [34].

Altogether kinetic measurements on the photoproduct of the •NO derivative of met-TrHbN

mutants and on their NOD activity (Table 9.1) indicated that the ST, LT and EHT are routes for

gas entry and exit. Both also indicated that some route(s) for •NO diffusion still exist in the triple

mutant.

9.5.3. MD simulations emphasize the importance of side-chain flexibility on

ligand diffusion

The extent of inhibition of the NOD reaction measured in the triple mutant suggested that

obstruction of the tunnel entrances was either incomplete or that other pathways may have been

used to access DHP. To verify these hypotheses, theoretical calculations were performed on the

triple mutant. Two approaches were used. In the first approach, a 30-ns MD simulation of the

mutant in the FeII(O2) form was carried out. This simulation allowed us to measure the impacts

of the mutations on the conformation and dynamics of the tunnels and the DHP. This trajectory

also served to calculate the potential of mean force (PMF) for •NO diffusion inside the mutant

using the implicit ligand sampling method [39]. In the second approach, •NO diffusion inside the

mutant and TrHbN was investigated from 10×10-ns simulations using the locally enhanced

sampling method [37]. The latter method allowed to determine the more likely pathways used by

•NO to exit out of the protein matrix.

204

(I) Equilibrium MD simulations

(a) Effects of the mutations on the opening of the tunnel entrances - All mutated side chains

were found to adopt various rotamers along the trajectory (Table S1). Plots showing side-chain

dihedral fluctuations are available in Supporting Material (Annexe 4, Fig. S3). However, even

with this flexibility, tunnel-solvent communications were dramatically reduced for the triple

mutant compare to TrHbN (Table 9.2). For instance, no opening of the ST was observed.

Accordingly, PMF calculations revealed a high energy barrier at the ST entrance compare to

TrHbN (Fig. 9.5b). For the EHT, one opening event was detected, which was promoted by the

simultaneous displacement of the side-chain of Ile65(H18), Leu118(H10) and Val122(H14) (Fig.

9.6b). Because of this side-chain flexibility, PMF calculations revealed a weak energy barrier of

+4.6 kcal/mol +0.3/-2.4 kcal/mol. (Fig. 9.5c). Communications between LT and solvent were

more frequent (29 isolated MD frames). These opening events were caused by two distinct

movements of the protein. In the first movement, Leu24(B1) adopted rotamers mm and mt

(Table S1, Fig. S3), which caused Leu24(B1) to flip into the solvent resulting in the opening of

the LT entrance Fig. 9.6a). The second movement opening the LT entrance corresponds to the

displacement of the N-terminal end of the B-helix toward the solvent. This motion, modest

regarding Cα position (amplitude of about 2 Å),displaces Leu24(B1) side chain toward the

solvent sufficiently to open LT. Because of these open conformations, PMF calculations revealed

a small increase in PMF levels (up to +2.2 kcal/mol ±0.6 kcal/mol) near the surface. Limiting

PMF calculations to the “closed conformations”, which account for more than 92% of the time,

revealed that the mutation efficiently blocks LT in these circumstances (Fig. 9.5a, red curve).

(b) Impacts of mutations on residue forming tunnels and DHP – Dynamics of all tunnel

residues were studied and results are given in Supporting Material figures S3 and S4 (Annexe 4)

show plots of side chain dihedrals over time and give rotamer populations. Table S2 (Annexe 4)

summarizes interaction between all residues shaping the tunnels. Among these residues, six

showed a modified behavior in the mutant (side chain dynamics or rotamer populations)

(Ile15(A11), Ile19(A15), Ile25(B2), Phe62(E15), Val118(H10) and Ile119(H11)) and two

showed new rotamers (Ile15(A11) and Ile119(H11)). Only changes in behavior of residues

Phe62(E15) and Ile119(H11) resulted in a modification of tunnel dynamics.

205

Phe62(E15) - Previous theoretical investigations of TrHbN emphasized the importance of

Phe62(E15) on ligand diffusion in the LT and EHT [16, 18, 44, 45]. In the mutant, the change in

dynamics of Phe62(E15) is characterized by the more frequent t80 rotamer over rotamers m30

and m-85 (Table S1, Fig. S4 l). In TrHbN, this preference is inverted. This is due to changes in

Ile25(B2) dynamics, the latter residue being located between Leu24(B1) and Phe62(E15). The

main consequence of this change is a more frequent communication between Xe5↔Xe2 and

EHc↔Xe2 cavities. On the other hand, communications between Xe1 and Xe5 communications

are less frequent (Fig. 9.1, purple arrow). No other diffusion route was created by this

modification.

Ile119(H11) – Part of the ST and EHT are defined by Ile119(H11). Changes in dynamics and

conformations of Ile119(H11) side chain were characterized by the adoption of rotamers tt and tp

(5.6%), which are mostly unpopulated (< 0.1%) in TrHbNFeII(O2) simulations (Table S1,

Fig. S4w). These rotamers favor the EHc ↔ GHc communication (part of GHT tunnel)

(Fig. 9.1)

(c) Formation of additional diffusion pathways - In addition to ST, LT and EHT, two additional

diffusion pathways were observed in the triple mutant: the GHT and a new path hereafter termed

EH2 tunnel (EH2T). GHT extends from the EHc to an apolar cavity (GHc) located between G

and H helices (Fig. 9.1). Residues Ala99(G13), Ile112(H4) and Leu116(H8) separate GHc from

the solvent. Residues Leu98(G12), Leu102(G16), Ile115(H7) and Ile119(H11) separate GHc

from the protein core. For TrHbN as well as for the mutant, GHc is occasionally open to the

solvent (Table 9.2). Solvent exposure is promoted by the flexible Leu116(H8) side chain

(Fig. 9.6c). In TrHbN, PMF calculations revealed very high energy barriers between GHc and

the neighbouring cavities (Xe1, Xe3 and EHc) indicating that the communication with other

tunnels is limited (Fig. 9.5). As previously mentioned, energy barrier separating GHc and EHc is

dramatically lowered in the triple mutant (Fig. 9.5).

The second diffusion route, the EH2T, also originates from the EHc and passes through a cavity

located on the proximal side of the heme. Tunnel EH2T has its surface entry located between the

E and H helices and is defined by side chain from residues Phe61(E14), Met77(F4) and

Leu122(H14). Evaluation of the PMF levels along the EH2T revealed that this route is the less

206

favorable with higher energy barriers, especially between the heme and the solvent. However,

the heights of these barriers are lowered in the triple mutant (Fig. 9.5e).

(d) Impacts of mutations on the DHP – In agreement with UV-Vis optical and resonance

Raman spectroscopy analysis of the FeIIO2 complexes of the triple mutant and TrHbN, the

conformations and dynamics of the residues shaping the DHP were found unchanged relative to

TrHbN along the mutant trajectory (Fig. S4 f to j).

(II) Locally enhanced sampling

(a) ST, LT and EHT as the main diffusion routes in TrHbN - For all TrHbN simulations, all

•NO molecules rapidly migrated out of the protein matrix, with more than 80% of the •NO

molecules having left within the first 2 ns (Fig. S6, green). Trajectories of •NO diffusion inside

TrHbN are shown in Fig. S7. •NO molecules exited mainly by the ST > LT > EHT (Table 9.3).

Also, five •NO exited by GHT. In the latter case •NO reached GHc from the ST through a

limited opening between Leu98(G12), Leu116(H8) and Ile119(H11). These events were not

expected given the high energy barrier between ST and GHc. The LES method, lowering energy

barrier, may have favored these events. However, upon •NO passage from ST to GHc,

Leu116(H8) was displaced causing the enlargement of the passage indicating that •NO may have

favored this displacement.

Several entries inside the protein matrix were also observed (Fig. S6, blue). Interestingly, a

single entry event by the BET was observed (Fig. 9.1). However, the later molecule returned by

the same route after 80 ps. These results agreed well with our previous work, which predicted the

ST, LT and EHT as the main diffusion routes [18]. In the latter work, simulations showed a

similar tunnel usage for •NO entry and exit (Occurrences from a total of 200 ns simulation time:

Entries : ST (20) > LT (13) > EHT (9), Exits : ST (13) > LT (7) > EHT (3) [18].

(b) Triple mutant partial blockade of the tunnels, other diffusion routes and the influence of

•NO – While more than 80% of the •NO molecules had reached the solvent within the first 2 ns

in TrHbN, only 47% exited the mutant within 10 ns indicating that although significantly

blocked the tunnels still allowed entry and exit of •NO (Table 9.3, Fig. S6 and S7b). As shown

in Table 9.3, the LT (15×), GHT (15x) and EHT (12×) were the most used routes. The ST was

207

used only three times. Exits by ST were not expected giving the very high energy barrier

calculated at the ST entrance (Fig. 9.5) and the absence of open conformation in the equilibrium

MD simulation (Table 9.2). However, detailed analysis of LES simulations (100 000 frames)

showed twelve MD isolated frames with an open ST (Table 9.2). These openings occurred

following the simultaneous displacement of the side chains defining ST entrance and without

these adopting non-previously observed rotamers (Fig. 9.6, d). Remarkably, eleven opening

events took place simultaneously with either an •NO exit or happened while a •NO was in

contact with ST surface residues (Table 9.3) indicating that the presence of •NO may impact

tunnel opening. The more frequent usage of LT and EHT was consistent with PMF calculations,

which had revealed lower energy barriers at the entrance relatively to ST (Fig. 9.5). For the EHT,

•NO diffusion was constrained by a narrow passage between Phe61(E14) and Leu122(H14). The

side-chain flexibility of Ile65(E18) and Leu122(H14) contributed to these exits. In the case of the

LT, transient displacements of Leu24(B1) side chain allowed •NO diffusion as in TrHbN. This

open conformation was also observed in the simulation in absence of •NO molecule (rotamers

mm and mt, Fig. 9.6, a and Table 9.2). As for ST, opening events at LT or EHT entrances were

more frequent in LES simulations than in MD simulations highlighting the impact of ligand

molecules on side-chain dynamics. A high percentage of these opening events were observed

while at least one •NO was near the mutated side chain (Table 9.2).

As mentioned before, frequent exits occurred via GHT in the triple mutant (Table 9.3). In all

cases •NO transited through the GHc before leaving the protein. In eleven events out of fifteen

•NO came from the EHc, through a passage enlarged by the displacement of Ile119(H11) side

chain (rotamers tt and tp). In the four other cases, •NO reached GHc from the Xe1 cavity through

a path opened by a transient displacement of Leu102(G16) (rotamer mt→tp). This motion of

Leu102(G16) is absent in equilibrium simulations without •NO (wild-type and mutant)

suggesting that •NO triggered the displacement of this residue. Solvent-GHc communications

were much more frequent in LES simulations (wt and mutant) than in equilibrium MD

simulations (Table 9.2) and were dependent on the conformation of Leu116(H8) side chain,

which is enhanced upon adoption of rotamers tt or tp (Fig. 9.6). In agreement, when a •NO is

docked in GHc, these rotamers account for 64.7% of the time compare to 5.6% in absence of

•NO.

208

9.6. Conclusions

Prior to this work, i.e. about 10 years following trHbN discovery and the elucidation of its

structure, there was no experimental proof demonstrating the functional role of trHbN tunnels. In

the other hand, many theoretical works were conducted by several teams around the world and

all of them share a common major conclusion: ligands diffuse through the tunnels. In the present

work, the modest effects of the mutations on geminate recombination kinetics as well as NOD

reactions are challenging our precedent theoretical understanding. However, the new

experimental data do not lead us to the conclusion that trHbN tunnels do no support ligand

diffusion. Intead, they surely told us that there is still much to discover and they cleary warns

that trHbN is a complexier protein than expected. To support this view, several hypotheses are

proposed to reconcile experiments in the discussion section of this thesis.

209

9.7. References

1. Brunori, M., Nitric oxide, cytochrome-c oxidase and myoglobin. Trends in Biochemical

Sciences, 2001. 26(1): p. 21-23.

2. Brunori, M., et al., Nitric oxide and cellular respiration. Cell Mol Life Sci, 1999. 56(7-8):

p. 549-57.

3. Stevanin, T.M., et al., Flavohemoglobin Hmp affords inducible protection for Escherichia

coli respiration, catalyzed by cytochromes bo' or bd, from nitric oxide. J Biol Chem,

2000. 275(46): p. 35868-75.

4. Cleeter, M.W., et al., Reversible inhibition of cytochrome c oxidase, the terminal enzyme

of the mitochondrial respiratory chain, by nitric oxide. Implications for neurodegenerative

diseases. FEBS Lett, 1994. 345(1): p. 50-4.

5. Brown, G.C. and C.E. Cooper, Nanomolar concentrations of nitric oxide reversibly

inhibit synaptosomal respiration by competing with oxygen at cytochrome oxidase. FEBS

Lett, 1994. 356(2-3): p. 295-8.

6. Gardner, P.R., G. Costantino, and A.L. Salzman, Constitutive and Adaptive

Detoxification of Nitric Oxide in Escherichia coli. Role of nitric-oxide dioxygenase in

the protection of actonitase. J. Biol. Chem., 1998. 273(41): p. 26528-26533.

7. Gardner, P.R., et al., Nitric Oxide Sensitivity of the Aconitases. J. Biol. Chem., 1997.

272(40): p. 25071-25076.

8. Kissner, R., et al., Formation and Properties of Peroxynitrite as Studied by Laser Flash

Photolysis, High-Pressure Stopped-Flow Technique, and Pulse Radiolysis. Chem. Res.

Toxicol., 1997. 10(11): p. 1285-1292.

9. Pfeiffer, S., et al., Metabolic Fate of Peroxynitrite in Aqueous Solution. Reaction with

nitric oxide and pH-dependent decomposition to nitrite and oxygen in a 2:1 stoichimetry.

J. Biol. Chem., 1997. 272(6): p. 3465-3470.

10. Ouellet, H., et al., Truncated hemoglobin HbN protects Mycobacterium bovis from nitric

oxide. Proc Natl Acad Sci U S A, 2002. 99(9): p. 5902-7.

11. Scott, E.E., Q.H. Gibson, and J.S. Olson, Mapping the pathways for O2 entry into and exit

from myoglobin. Journal of Biological Chemistry, 2001. 276(7): p. 5177-5188.

12. Huie, R.E. and S. Padmaja, The Reaction of NO with Superoxide. Free Radical Research

Communications, 1993. 18(4): p. 195-199.

13. Milani, M., et al., Mycobacterium tuberculosis hemoglobin N displays a protein tunnel

suited for O2 diffusion to the heme. EMBO J, 2001. 20(15): p. 3902-9.

210

14. Milani, M., et al., Cyanide binding to truncated hemoglobins: a crystallographic and

kinetic study. Biochemistry, 2004. 43(18): p. 5213-21.

15. Milani, M., et al., Heme-ligand tunneling in group I truncated hemoglobins. J Biol Chem,

2004. 279(20): p. 21520-5.

16. Daigle, R., M. Guertin, and P. Lague, Structural characterization of the tunnels of

Mycobacterium tuberculosis truncated hemoglobin N from molecular dynamics

simulations. Proteins-Structure Function and Bioinformatics, 2009. 75(3): p. 735-747.

17. Mishra, S. and M. Meuwly, Nitric oxide dynamics in truncated hemoglobin: docking

sites, migration pathways, and vibrational spectroscopy from molecular dynamics

simulations. Biophys J, 2009. 96(6): p. 2105-2018.

18. Daigle, R., et al., Theoretical Investigations of Nitric Oxide Channeling in

Mycobacterium tuberculosis Truncated Hemoglobin N. Biophysical Journal, 2009.

97(11): p. 2967-2977.

19. Baron, R., et al., Multiple pathways guide oxygen diffusion into flavoenzyme active sites.

Proceedings of the National Academy of Sciences of the United States of America, 2009.

106(26): p. 10603-10608.

20. Montet, Y., et al., Gas access to the active site of Ni-Fe hydrogenases probed by X-ray

crystallography and molecular dynamics. Nat Struct Biol, 1997. 4(7): p. 523-526.

21. Johnson, B.J., et al., Exploring molecular oxygen pathways in Hansenula polymorpha

copper-containing amine oxidase. J Biol Chem, 2007. 282(24): p. 17767-17776.

22. Saam, J., et al., Molecular dioxygen enters the active site of 12/15-lipoxygenase via

dynamic oxygen access channels. Proc Natl Acad Sci U S A, 2007. 104(33): p. 13319-

13324.

23. Salomonsson, L., et al., A single-amino-acid lid renders a gas-tight compartment within a

membrane-bound transporter. Proceedings of the National Academy of Sciences of the

United States of America, 2004. 101(32): p. 11617-11621.

24. Igarashi, R.Y. and L.C. Seefeldt, Nitrogen fixation: the mechanism of the Mo-dependent

nitrogenase. Crit Rev Biochem Mol Biol, 2003. 38(4): p. 351-384.

25. Volbeda, A. and J.C. Fontecilla-Camps, Crystallographic evidence for a CO/CO2 tunnel

gating mechanism in the bifunctional carbon monoxide dehydrogenase/acetyl coenzyme

A synthase from Moorella thermoacetica. J Biol Inorg Chem, 2004. 9(5): p. 525-532.

26. Pesce, A., et al., The 109 residue nerve tissue minihemoglobin from Cerebratulus lacteus

highlights striking structural plasticity of the alpha-helical globin fold. Structure, 2002.

10(5): p. 725-735.

211

27. Hiromoto, T., et al., Crystal structure of 3-hydroxybenzoate hydroxylase from

Comamonas testosteroni has a large tunnel for substrate and oxygen access to the active

site. J Mol Biol, 2006. 364(5): p. 878-896.

28. Pesce, A., et al., Reversible hexa- to penta-coordination of the heme Fe atom modulates

ligand binding properties of neuroglobin and cytoglobin. IUBMB Life, 2004. 56(11-12):

p. 657-664.

29. Chen, L., et al., The Binding and Release of Oxygen and Hydrogen Peroxide Are

Directed by a Hydrophobic Tunnel in Cholesterol Oxidase. Biochemistry, 2008. 47(19):

p. 5368-5377.

30. Tan, X.S., et al., The tunnel of acetyl-coenzyme a synthase/carbon monoxide

dehydrogenase regulates delivery of CO to the active site. Journal of the American

Chemical Society, 2005. 127(16): p. 5833-5839.

31. Tan, X., et al., Function of the tunnel in acetylcoenzyme A synthase/carbon monoxide

dehydrogenase. J Biol Inorg Chem, 2006. 11(3): p. 371-378.

32. Riistama, S., et al., Channelling of dioxygen into the respiratory enzyme. Biochim

Biophys Acta, 1996. 1275(1-2): p. 1-4.

33. Riistama, S., et al., Binding of O2 and its reduction are both retarded by replacement of

valine 279 by isoleucine in cytochrome c oxidase from Paracoccus denitrificans.

Biochemistry, 2000. 39(21): p. 6365-6372.

34. Ouellet, Y.H., et al., Ligand binding to truncated hemoglobin N from Mycobacterium

tuberculosis is strongly modulated by the interplay between the distal heme pocket

residues and internal water. J Biol Chem, 2008. 283(40): p. 27270-8.

35. Salter, M.D., et al., The apolar channel in cerebratulus lacteus hemoglobin is the route for

O2 entry and exit. J Biol Chem, 2008. 283(51): p. 35689-35702.

36. Brooks, B.R., et al., CHARMM: A program for macromolecular energy, minimization,

and dynamics calculations. Journal of Computational Chemistry, 1983. 4(2): p. 187-217.

37. Elber, R. and M. Karplus, Enhanced sampling in molecular-dynamics - Use of the time-

dependent Hartree approximation for a simulation of carbon-monoxide diffusion through

myoglobin. Journal of the American Chemical Society, 1990. 112(25): p. 9161-9175.

38. DeLano, W.L., The PyMOL Molecular Graphics System. 2002.

39. Cohen, J., et al., Imaging the migration pathways for O2, CO, NO, and Xe inside

myoglobin. Biophys J, 2006. 91(5): p. 1844-57.

40. Eich, R.F., et al., Mechanism of NO-induced oxidation of myoglobin and hemoglobin.

Biochemistry, 1996. 35(22): p. 6976-6983.

212

41. Herold, S. and A. Puppo, Oxyleghemoglobin scavenges nitrogen monoxide and

peroxynitrite: a possible role in functioning nodules? Journal of Biological Inorganic

Chemistry, 2005. 10(8): p. 935-945.

42. Yukl, E.T., S. de Vries, and P. Moenne-Loccoz, The Millisecond Intermediate in the

Reaction of Nitric Oxide with Oxymyoglobin is an Iron(III)-Nitrato Complex, Not a

Peroxynitrite. Journal of the American Chemical Society, 2009. 131(21): p. 7234-7235.

43. Mishra, S. and M. Meuwly, Atomistic Simulation of NO Dioxygenation in Group I

Truncated Hemoglobin. Journal of the American Chemical Society, 2010. 132(9): p.

2968-2982.

44. Crespo, A., et al., Theoretical study of the truncated hemoglobin HbN: exploring the

molecular basis of the NO detoxification mechanism. J Am Chem Soc, 2005. 127(12): p.

4433-44.

45. Bidon-Chanal, A., et al., Ligand-induced dynamical regulation of NO conversion in

Mycobacterium tuberculosis truncated hemoglobin-N. Proteins, 2006. 64(2): p. 457-64.

213

9.8. Footnotes

We are grateful to Dr. Beatrice A. Wittenberg and Dr. Jonathan B. Wittenberg from the Albert

Einstein College of Medicine (NY, USA) for insightful discussions. This work was supported by

the National Sciences and Engineering Research Council (NSERC) grant 46306-01 and the

Fonds Québécois de la Recherche sur la Nature et les Technologies (FQRNT) grant 104897 to

Dr. Michel Guertin. Patrick Lagüe is supported by the Canadian Foundation for Innovation (CFI)

grant 12428 and the Fonds Québécois de la Recherche sur la Nature et les Technologies

(FQRNT) grant 104897.

The abbreviations used are: BCG, bacillus Calmette-Guérin; DHP, distal heme pocket, Hb,

hemoglobin; HS, high spin; LS, low spin; ILS, implicit ligand sampling; LES, locally enhanced

sampling; Mb, myoglobin; TrHb, truncated hemoglobin; TrHbN, Mycobacterium tuberculosis

truncated hemoglobin N; ST. short tunnel; LT, long tunnel; EHT, EH tunnel, BET, BE tunnel,

GHT, GH tunnel; MD, molecular dynamics; QM-MM, quantum mechanics-molecular

mechanics.

214

Table 9.1 Kinetics constants for the NOD reactions of TrHbN and its tunnel mutants.

Protein

A → B B → C

k1 k2

s-1

± s-1

±

TrHbN 1902 54 22.7 0.2

LT 1751 55 15.2 0.1

ST 1383 25 19.0 0.1

LT/ST 1034 20 12.5 0.1

LT/ST/EHT 814 56 13.7 0.3

Horse heart Mb 145 4 27.3 0.5

215

Table 9.2 Number of MD snapshots showing a tunnel open at its entrance from simulations

of TrHbN and the triple mutant.

Tunnel

Opening Events †

wt-TrHbN Wt-TrHbN + •NO Triple mutant Triple mutant +

•NO†

LT 34116 21719 (29%) 29 811 (65%)

ST 1452 1276 (74%) 0 12 (92%)

EHT 2550 1485 (48%) 1 177 (96%)

GHT 30 422 (32%) 8 1678 (41%)

† Based on a 100 ns sampling for the sake of comparisons. Numbers for snapshots without

•NO are doubled for wt-TrHbN and quadrupled for the mutant

216

Table 9.3 Exit and entry events observed in LES simulations of the wild type and triple

mutant.

Tunnel

Wild type Triple mutant

Exit Entry Exit Entry

ST 68 29 3 0

LT 67 26 15 4

EHT 22 7 12 1

BET 1 1 0 0

GHT 5 2 15 0

EH2T 0 0 2 0

217

Figure 9.1 TrHbN tunnel system. Xenon binding sites and cavities are identified by pale

yellow spheres. Communications between solvent and tunnels are identified by cyan

arrows. Helices are identified by corresponding letters. Cavity communications forming

along each tunnels are summarized below the picture. The picture was produced using

PyMOL (38)

218

Figure 9.2 Reaction of TrHbNFeII(O2) (5 µM) with one equivalent of •NO at 5 ºC, pH 9.5.

(a) Evolution of the optical spectra acquired during the first 250 ms and collected on time

scales ranging from 1.3 ms (red line) to 250 ms (blue line) with an integration time of 2.5

ms. Abs, absorbance units. (b) UV-V spectrum of first spectrum obtained 1.3 ms after

mixing bearing the signature of a HS species with peaks at 500 and 630 nm.

219

Figure 9.3 Reaction of the FeII(O2) form (5 µM) of TrHbN (a, b) and tunnel mutants LT/ST

(c, d) and LT/ST/EHT (e, f) with one equivalent of •NO at 5 ºC, pH 9.5. (a, c and e), optical

spectra of the species obtained by singular value decomposition and global analysis of the

rapid scan data from (a): Species A (oxygenated form, red), species B (intermediate

species, black) and species C (hydroxyl form, blue). Abs: absorbance. (b, d and f), the

reaction of oxidation of the different proteins by •NO was well described using a double

exponential function (ABC). The kinetics at 433 nm (black) and the fit (red) are

shown. Note the appearance of peaks at 500 and 630 nm for c and e panels.

220

Figure 9.4 Kinetic traces illustrating the absorbance changes after photo-dissociation of (a)

TrHbNFeIII

(•NO), (b) ST-FeIII

(•NO), (c) LT/ST-FeIII

(•NO), and (d) LT/ST/EHT-FeIII

(•NO)

at 23 °C. The reaction was monitored at 392 nm, this wavelength corresponds to the

isosbestic point of the optical spectra of the 5c (FeIII

) and FeIII

(OH2) species of the different

proteins. The rapid decrease in absorbance observed in the mutants corresponds to

geminate •NO rebinding.

221

Figure 9.5 PMF profiles for •NO diffusion in the different tunnels for the TrHbN (blue) and

mutant (black). Shaded gray zones correspond to tunnel region filled by the mutations.

PMF profiles were calculated for tunnels LT (a), ST (b), EHT (c), GHT (d) and EH2 (e).

For LT (a), profile in red was calculated for the mutant without Ala24(B1) mm and mt

rotamers. For tunnel passing by GHc (d), profile in red was calculated mutant without

Ile119(H11) rotamers tt and tp. To highlight Xe1, Xe2, Xe5 cavities along LT, PMF was

calculated depending on Phe62(E15) conformations (rotamers t80 or m30/m-85) and only

rotamer t80 is shown here for picture clarity. PMF profiles calculated with m-85 and m30

rotamers are available in Supporting Material. Errorbars, depending on calculated PMF

levels, are not shown for picture clarity (see method).

222

Figure 9.6 Typical closed (left) and open (right) tunnels at the surface observed for the

triple mutant protein. The mutated residues are colored in green. (a) LT : Opening coupled

with displacement of Leu24(B1) (green). (b) EHT : Opening coupled with displacement of

Ile65(E18) (yellow), Val118(H10) and Leu122(H14) (c) GHc : Opening coupled with

displacement of Leu116(H8) (yellow). (d) ST : Opening coupled with displacement of all

surface residues and observed while a •NO was nearby.

223

10.

Chapitre 10

Discussion

Les travaux concernant la présente thèse de doctorat visaient à étudier la structure et la

dynamique de TrHbN à l’aide d’outils bio-informatiques. Ces travaux ont permis d’étudier

plusieurs aspects de TrHbN, notamment la structure et la dynamique du site actif, du

squelette et des tunnels de l’enzyme. Également, le rôle des tunnels dans la diffusion des

substrats gazeux a été étudié. Ces travaux théoriques ont bénéficié d’un appui expérimental

grâce à des travaux en spectroscopie RMN en solution et à l’étude cinétique de mutants.

Les principales avancées scientifiques réalisées au cours de cette thèse sont discutées dans

les pagraphes qui suivent.

10.1. Structure et dynamique de la poche distale de l’hème

Le site distal de TrHbN est formé par la chaîne latérale de quelques résidus, soient la

Phe32(B9), la Tyr33(B10), la Phe46(CD1), la Gln58(E11) et la Val94(G8). Les résidus en

position 33 et 58 sont polaires et suffisamment près du fer pour former des interactions

avec le ligand (Chapitres 5 et 6). En absence de l’O2 lié au fer, les résidus distaux

Try33(B10) et Gln58(E11) jouent un rôle dans le maintien et le positionnement d’une

molécule d’eau au site actif de la forme deoxy-TrHbN (Chapitre 5). Cette molécule d’eau

se positionne alors près du fer limitant l’accès des ligands à la poche distale de l’hème.

En présence de cette molécule d’eau, le positionnement des résidus distaux et la dynamique

des tunnels de deoxy-TrHbN sont comparables à ceux de la forme oxygénée. Sans

molécule d’eau, le positionnement de la Gln58(E11) et la dynamique de la Phe62(E15)

diffèrent (Chapitre 5). Lorsque la Tyr33(B10) et Gln58(E11) sont remplacés par des résidus

apolaires, il a été démontré que la molécule d’eau est instable et celle-ci s’éloigne et/ou

quitte vers le solvant. En accord, les constantes bimoléculaires d’association de l’O2 et CO

pour les mutants apolaires sont largement augmentées, en particulier lorsque la Tyr33(B10)

224

est substituée; la constante d’association de l’O2 au fer dépassant même la constante

bimoléculaire de la réaction NOD (Chapitre 5). La présence de cette molécule d’eau permet

donc d’expliquer pourquoi la constante bimoléculaire de la réaction NOD catalysée par

TrHbN est 15x plus élevée que la fixation de l’oxygène moléculaire à deoxy-TrHbN.

Il demeure encore certaines interrogations quant à cette molécule d’eau. D’où vient-elle?

Entre-t-elle via un des tunnels apolaires ou emprunte-t-elle une route différente? Est-elle

déjà présente dans la structure de TrHbN? Les simulations de DM menées sur la forme

deoxy n’ont pas montré d’entrée de molécule d’eau. De plus, la structure tridimensionnelle

de la forme deoxy est encore inconnue. Il n’est pas impossible qu’un changement structural

survienne lors de la formation de deoxy-TrHbN. En appui, la spectroscopie de résonance

Raman menée sur la forme deoxy du double mutant Tyr33(B10)Phe/Gln58(E11)Val a

révélé la présence d’un mélange 5C et 6C au lieu d’une seule forme 5C, le 6e ligand étant

de nature endogène [175]. L’absence de chaîne latérale pouvant se lier au fer dans la

proximité immédiate du fer suggère un réarrangement plutôt important de la structure de

TrHbN. Comme hypothèse, il est possible que ce réarrangement, du moins une certaine

partie, survienne également pour la forme sauvage de l’enzyme après la dissociation du

ligand. Cependant, chez TrHbN ce changement favoriserait l’entrée d’une molécule d’eau

et sa diffusion vers le site distal, celle-ci étant attirée puis maintenue par les résidus polaires

Tyr33(B10) et Gln58(E11). Une fois stabilisée près du fer, cette molécule d’eau jouerait un

rôle structural en favorisant le maintien d’une poche distale dans une configuration proche

de celle de la forme oxygénée, propice à l’arrivée d’un substrat gazeux apolaire arrivant via

un tunnel apolaire. Cette hypothèse pourrait éventuellement être testée à l’aide de

simulations de DM. Avec l’avènement de nouveaux supercalculateurs, il serait possible de

produire des trajectoires de dynamique moléculaire (>1 µs) possiblement suffisamment

longues pour observer ces mouvements.

225

10.2. Structure et dynamique des tunnels de TrHbN

Le site distal de TrHbN est relié à la surface de l’enzyme via un réseau de tunnels

hydrophobes contenu dans la matrice protéique. Jusqu’à quatre tunnels distincts ont été

observés au cours des simulations (Chapitre 6), soit deux de plus que ceux observé dans la

structure cristalline [174]. Cette caractéristique structurale distingue TrHbN des TrHbs des

groupes II et III qui possèdent une cavité distale très compacte et des cavités apolaires

isolées et de volumes plus modestes. La communication entre le site distal et les tunnels est

grandement due à la taille du résidu en position G8 qui est occupée par une valine chez

TrHbN alors qu’on y retrouve un tryptophane chez les TrHbs des groupes II et III.

Les simulations de DM ont révélé que les tunnels prenant place au sein de la structure de

TrHbN sont très dynamiques. En effet, il peut s’y former différents tunnels lors de la fusion

momentanée de différentes cavités hydrophobes (Chapitre 6). Cette caractéristique de la

structure de TrHbN est sans doute un facteur clef permettant à cette enzyme de catalyser la

réaction NOD à une vitesse s’approchant des réactions limitées que par la diffusion des

molécules.

Il est intéressant de rappeler que les tunnels sont contenus au centre de quatre hélices alpha

rigides (Chapitres 6 et 8). Au Chapitre 6, il a été proposé que cette rigidité permette le

maintien du volume vide interne. Ces dernières observations s’opposent à une autre théorie

générale chez les enzymes où une réduction de la mobilité conformationelle est associée à

une réduction de l’activité [176, 177]. Par contre, de récentes études d’évolution dirigée ont

démontré que rigidité et activité peuvent être découplées [178, 179]. TrHbN, étant à la fois

rigide et un catalyseur très efficace, représente donc un cas particulier. L’espace vide

interne et la mobilité des chaînes latérales semblent alors la clef pour garantir une haute

activité. Cette rigidité joue possiblement un rôle dans un contexte physiologique en

permettant la diffusion des ligands interne plus indépendante des conditions physico-

chimiques pouvant subsister dans l’environnement intracellulaire. Enfin, les mouvements

lents (µs-ms) localisés le long des hélices B et G et révélés par spectroscopie RMN en

solution peuvent avoir des répercussions non négligeables sur le volume vide interne et son

organisation dans l’espace.

226

10.3. Rôle des tunnels dans la diffusion des substrats entre le

solvant et le site actif

Comme mentionné précédemment, de nombreuses simulations de DM ont mené à

différents postulats quant aux mécanismes de la diffusion des ligands à l’intérieur de

TrHbN [170, 171, 180-183]. Ces derniers ont en particulier identifié les tunnels Court et

Long comme les routes de diffusion privilégiées. En support, les structures de TrHbN

obtenues à partir de cristaux placés sous haute pression de xénon ont montré des atomes de

xénon localisés le long de ces deux routes [184].

Avant les travaux menés en laboratoire et présentés au Chapitre 9, cette observation était le

seul support expérimental suggérant un rôle pour les tunnels de TrHbN. Dans les travaux

présentés dans ce chapitre, nous avons tenté de confirmer expérimentalement l’existence de

plusieurs routes fonctionnelles pour la diffusion des ligands à l’intérieur de TrHbN. Pour ce

faire, nous avons construit des mutants ayant les tunnels Court, Long et EHT bloqués

individuellement ou en combinaison et tenté ensuite de mettre en évidence leur importance

pour l’accès du •NO vers le site actif de TrHbN. Tous les mutants ont démontré une

réaction NOD ralentie indiquant que ceux-ci constituent des routes pour la diffusion du

•NO (Table 9.1, Fig. 9.3). Cette conclusion a été renforcée par des expériences de

photolyse avec la forme ferrique-•NO (Fig. 9.4).

Pendant la rédaction de cette thèse, d’autres travaux sur TrHbN et le triple mutant des

tunnels étaient réalisés en collaboration dans un laboratoire étranger, celui du Dr. Marten

Vos. Ce chercheur de calibre international a développé une expertise technique permettant

l’étude de la recombinaison géminée dans l’échelle de temps de la picoseconde [185].

Étonnamment, des essais menés sur la forme ferrique-•NO de TrHbN ont montré des tracés

de recombinaison identiques pour le mutant et la protéine sauvage. De plus, des essais

menés sur la forme CO n’ont pas révélé de recombinaison géminée 4 ns après la photolyse,

un résultat cohérent avec la sortie facile du ligand à l’extérieur des deux protéines. La

situation opposée a été obtenue pour la protéine mt-TrHbO laquelle n’a aucun tunnel

227

apparent et une poche distale beaucoup plus compacte que celle de TrHbN. Ces derniers

résultats viennent mêler les cartes quant au rôle des tunnels de TrHbN. Viennent-ils

invalider les modèles précédemment proposés sur le rôle et le fonctionnement des tunnels

de TrHbN? Pour tenter de répondre à cette question, les prochains paragraphes mettront en

perspectives l’ensemble des travaux théoriques et expérimentaux menés sur TrHbN.

Les mutations créées dans le but de bloquer les tunnels de TrHbN (Chapitre 9) n’ont pas eu

d’impacts dramatiques tels que ceux observés chez la mini-Hb du nématode Cerebratulus

lacteus [186]. Par surcroît, comme mentionné précédemment, les résultats obtenus par le

laboratoire du Dr. Marten Vos montrent que les mutations n’ont aucun impact sur les

cinétiques de recombinaison. Ensemble, ces résultats expérimentaux contrastent avec ceux

obtenus via les simulations de DM qui prédisent un rôle fonctionnel des tunnels dans la

diffusion des ligands (Chapitres 7, 9). En se basant sur les travaux de DM, les mutations

auraient dû avoir causé des effets significatifs in vitro. D’un autre côté, les simulations de

DM présentées dans cette thèse contiennent également plusieurs observations pouvant

réconcilier les résultats expérimentaux et théoriques. Au Chapitre 9, les simulations ont

montré :

que les ligands peuvent entrer/sortir près de l’entrée obstruée malgré

l’augmentation de l’encombrement stérique.

d’autres routes de diffusion, non observées dans la protéine sauvage. Ces

dernières sont vraisemblablement la conséquence des mutations elles-mêmes

sur la structure et la dynamique de la protéine.

Que l’espace vide considérablement important qui est contenu dans la

structure de TrHbN est favorable à la réorganisation des chaînes latérales.

Que les interactions protéine:ligand engendrent le déplacement de chaînes

latérales spécifiques menant à la formation de route de diffusion non

observée dans les simulations sans ligand.

Toutes ces observations montrent clairement que l’espace vide interne de TrHbN est

hautement dynamique. Par conséquent, cette caractéristique de la structure de TrHbN

228

pourrait expliquer pourquoi ses tunnels sont très difficiles à étudier et à mettre en évidence

par le biais de mutations. Pour ces raisons, au lieu d’invalider les précédents modèles

proposés sur le fonctionnement et le rôle des tunnels de TrHbN, les toutes dernières

données expérimentales précisent notre compréhension. Celles-ci soulèvent également de

nouvelles questions et ouvrent d’autres avenues de recherches. Il est clair que TrHbN est

une protéine plus complexe qu’anticipé et que notre compréhension est encore partielle.

En plus des observations énumérées précédemment, d’autres explications peuvent aider à

réconcilier les données théoriques et expérimentales. D’abord, il n’est pas exclu que

d’autres conformations structurales de TrHbN existent en solution. Si tel est le cas,

l’organisation des tunnels et donc la diffusion des ligands peuvent différer de ce que

proposent les modèles actuels. Pour supporter cette hypothèse, les travaux de spectroscopie

RMN en solution menés sur TrHbN sous sa forme cyanomet (Chapitre 8) ont révélé que la

pre-A est désordonnée contrairement à l’hélice alpha observée dans la structure cristalline

et maintenue lors des simulations de DM. De plus, ces travaux ont révélé l’existence de

mouvements lents se produisant sur l’échelle de temps µs-ms pour plusieurs résidus situés

le long des hélices B et G. Parmi les résidus concernés, il y a la Tyr33(B10), résidu clef

étant donné sa chaîne latérale formant une partie de la poche distale de l’hème et

interagissant avec le ligand lié au fer. En même temps, les atomes de la chaîne principale de

ce résidu délimitent en partie l’ouverture du tunnel BET présenté au Chapitre 6. Ce tunnel

n’a pas été ciblé par les mutations puisque ce tunnel est beaucoup plus étroit près de la

surface de la protéine par rapport aux trois autres ciblés (Chapitre 6). Des mouvements des

hélices B et G pourraient avoir des effets significatifs sur l’ouverture du tunnel BET et/ou

permettre la formation d’autres routes de diffusion.

Enfin, il est possible que les mutations créées pour bloquer les tunnels aient causé des effets

significatifs sur la structure de TrHbN causant la formation de nouvelles routes de

diffusion. De telles modifications n’auraient pas pu être observées au cours des simulations

de DM puisque celles-ci impliquent des mouvements de grande amplitude s’effectuant sur

des échelles de temps encore non accessibles aux simulations de DM.

229

10.4. Perspective de recherche sur les relations structure-

fonction des tunnels de TrHbN

Différents travaux pourraient être réalisés pour pousser notre compréhension de TrHbN, en

particulier sur le rôle des tunnels dans la diffusion des substrats. Les prochains paragraphes

en proposent quelques-uns.

D’abord, déterminer la structure 3D du triple mutant des tunnels permettrait de révéler leurs

impacts sur la configuration des cavités et des tunnels de TrHbN. Si des différences sont

observées, les nouvelles coordonnées pourraient servir à initier de nouveaux calculs de DM

et/ou à mener de nouveaux travaux en laboratoire (par exemple tester de nouvelles

mutations).

Étudier TrHbN et le triple mutant des tunnels en utilisant la spectroscopie de Laue résolue

en temps réel pourrait mettre en évidence expérimentalement la ou les routes de diffusion

utilisée(s) ainsi que les fluctuations structurales s’effectuant au cours de la diffusion du

ligand photodissocié. En complément, ces travaux pourraient fournir une description

structurale des mouvements lents observés le long des hélices B et G et mis en évidence par

la spectroscopie RMN en solution (Chapitre 8).

Étudier la dynamique des chaînes latérales de TrHbN et le triple mutant par spectroscopie

RMN en solution pourrait mettre en évidence des mouvements non observés au cours des

simulations de DM. Ces derniers pourraient avoir un impact significatif sur la diffusion des

ligands. Si de tels mouvements étaient identifiés, ceux-ci pourraient être par la suite

modélisés pour évaluer leurs impacts sur l’organisation de l’espace interne vide et sur la

dynamique des autres résidus.

10.5. Routes de diffusions multiples et pertinence fonctionnelle

L’ensemble des travaux présentés lors de cette thèse démontre clairement que la structure

de TrHbN permet la diffusion rapide des ligands vers son site actif. Cette propriété pourrait

230

être garantie par l’espace interne vide considérablement grand lequel est connecté à la

surface de la protéine en plusieurs endroits. La pertinence fonctionnelle à maintenir de

nombreuses voies de diffusion chez une protéine présente un intérêt fondamental certain.

D’abord, la présence de plusieurs routes apolaires chez TrHbN pourrait augmenter la

chance de capter une molécule de •NO ou d’O2 et ainsi contribuer à une haute efficacité de

catalyse de la réaction NOD sous une faible tension d’O2 (~1-3 µM) [187], telle que celle

prévalant durant l’infection de Mtb [188]. Le calcul d’affinité des tunnels pour le •NO

présenté au Chapitre 7 a prédit une faible occupation de ce ligand dans les tunnels (Kd = 6.2

mM) [171]. Dans des conditions physiologiques où la concentration du •NO est ~0.1-1.0

µM, une molécule de TrHbN à toutes les ~62000 à 6200 devrait contenir une molécule de

•NO. Pour maintenir une vitesse de réaction de 745 s-1

, la formation du produit de la

réaction de ne devrait pas pouvoir durer plus de 0.2 µs. En accord, des travaux théoriques

menés par Mishra et Meuwly suggèrent que la réaction entre TrHbNFeII(O2) et le •NO

s’effectue sur quelques dizaines de picosecondes incluant (i) la liaison du •NO à

TrHbNFeII(O2), (ii) le réarrangement de l’intermédiaire peroxynitrite (Fe

III(OONO

-)) et

finalement (iii) la dissociation du complexe nitrato-TrHbN [180]. Ainsi, la réaction NOD

ne serait limitée que par la diffusion du •NO du solvant vers le site actif comme le suggère

la constante bimoléculaire k´NOD ≈ 745 µM-1

s-1

(23 °C) [189]. Dans ce contexte, les

tunnels hydrophobes de TrHbN garantiraient un accès rapide du •NO vers le site actif.

10.6. Comparaisons de TrHbN avec d’autres protéines liant des

gaz et perspectives

À notre connaissance, étant donné les données cinétiques du triple mutant des tunnels

(Chapitre 9), TrHbN pourrait représenter la première protéine confirmée possédant

plusieurs routes fonctionnelles. De nombreux travaux théoriques ont suggéré l’existence de

routes de diffusion multiple chez d’autres protéines dont la COA [93], la flavoenzyme

oxydase et la mono-oxygénase [190] et la Mb [88-90]. Chez la Mb, l’étude de plusieurs

mutants a suggéré que ceux-ci ne sont pas fonctionnels ou négligeables [191]. Au contraire,

231

l’ensemble des travaux expérimentaux pointe vers une route majeure contrôlée par

l’His(E7) [191]. De manière similaire, plusieurs routes de diffusion à l’intérieur de la mini-

Hb de Cerebratulus lacteus (CerHb) ont été observées par échantillonnage amélioré de

ligands [192], mais une seule a été confirmée expérimentalement [186]. Il est intéressant de

noter que comme TrHbN, la structure de CerHb contient un volume interne vide très

significatif. Des mutations créées dans le but de bloquer un seul tunnel correspondant au

tunnel EHT de TrHbN ont causé des effets très importants [186]. Par surcroît, les mutations

créées pour bloquer ce tunnel ont été réalisées sur le résidu en position E18, soit la même

position topologique que celle ciblée avec TrHbN (Chapitre 9).

Une étude théorique comparative de la structure et de la dynamique de TrHbN avec celles

de CerHb devrait révéler d’importantes différences entre les deux protéines. Comme

hypothèse, il serait possible d’envisager que la structure de CerHb, en particulier les

chaînes latérales internes formant le cœur de sa structure, soit plus rigide que TrHBN. Ceci

limiterait la réorganisation du volume interne vide de CerHb empêchant donc la formation

de routes alternatives. De plus, l’impact des ligands eux-mêmes sur ces mêmes chaînes

latérales, c.-à-d. leur conformation et leur dynamique, serait également davantage restreint.

Une explication similaire a été proposée pour le cytochrome c oxydase chez laquelle une

seule mutation a causé un blocage dramatique de la diffusion des ligands [129].

Il est également d’intérêt de comparer la dynamique des tunnels de TrHbN à celle d’autres

TrHbs du groupe I dans le contexte des travaux de Cohen et al [89]. Ces derniers travaux

théoriques ont conclu que différentes globines, malgré leur repliement structural similaire,

présentent différents patrons de routes de diffusion. Ainsi, comparer la dynamique des

tunnels de TrHbN à celle d'autres TrHbs du groupe I, lesquelles présentent toutes un

volume vide interne important [78, 184, 193], suscite un fort intérêt. Ce type d’étude

pourrait permettre de mieux comprendre les déterminants structuraux affectant la diffusion

des ligands à l’intérieur de cette famille de protéines et des protéines en général. La

flexibilité des chaînes latérale délimitant l’espace interne vide tout comme la rigidité du

squelette pourraient s’avérer des caractéristiques évolutives de la famille de TrHbs du

groupe I.

232

Les travaux de Cohen [89] tout comme les nôtres suggèrent que plusieurs routes

alternatives peuvent être contenues dans une structure tertiaire générale donnée. Cependant,

certaines routes sont privilégiées dues à l’emboîtement des chaînes latérales, leur degré de

liberté et l’impact des ligands sur celles-ci. L’impact des ligands sur la dynamique de la

protéine est également un facteur important chez TrHbN (Chapitres 7 et 9), une observation

également soulevée pour la Mb par Tomita et al [194]. Ces derniers, en illuminant

continuellement des cristaux de CO-Mb sous des températures cryogéniques ont observé

que la migration des molécules de CO à l’intérieur de sites Xe cause l’expansion du volume

des cavités laquelle cause la formation de passage entre les cavités et la diffusion du CO.

Tout comme le mouvement des chaînes latérales, la mobilité de la chaîne principale peut

être un facteur important. Enfin, la surface hydrophobe invaginée en forme d’entonnoir

favorise la capture et l’entrée de ligands apolaires est notée pour le tunnel Court

(Chapitre 7) [171]. Un arrangement structural similaire a également été noté chez la 12/15-

lipoxygénase et ce dernier présente, comme l’entrée du tunnel Court de TrHbN, une plus

haute affinité pour les ligands apolaires que celle du solvant [96].

10.7. Localisation de TrHbN au niveau des membranes –

nouvelles perpectives de recherche

Il existe un certain nombre d’observations qui suggèrent que TrHbN puisse être localisée

près ou à l’intérieur les membranes. D’abord, l’extraction de TrHbN après sa surexpression

dans E. coli a révélé qu’environ 15 à 20% des molécules de TrHbN étaient associées aux

membranes (résultats non publiés). Également, il a été démontré que TrHbN interagit avec

l’ADN, en particulier via la région pre-A. En effet, des expériences de retardement de

migrations sur gel ont montré que TrHbN cause un retard dans la migration d’un plasmide.

À l’opposé, la protéine mutante TrHbN_Δpre-A ne cause pas ce retardement (résultats non

publiés). De plus, la spectroscopie de résonance RMN a montré des modifications dans le

déplacement chimique de résidus de la pre-A lorsque TrHbN est incubée en présence

d’ADN. La présence de quatre résidus chargés positivement dans la pre-A (trois arginines

et une lysine) pourrait jouer un rôle dans ces interactions par le biais d’interactions ioniques

avec des groupements phosphate de l’ADN, lesquels sont chargés de négativement.

233

Indirectement, ces observations suggèrent que des interactions semblables pourraient

également se former avec la tête de phospholipides chargés négativement. En support, de

récents travaux de pression maximale d’insertion en monocouche avec TrHbN menés au

laboratoire du Pr. Christan Salesse ont montré que TrHbN interagit avec les lipides

anioniques cardiolipine et phosphatidylinositol et, dans une moindre mesure, avec les

lipides zwitterioniques phosphatidylcholine et phophatidyléthanolamine (résultats non

publiés).

Afin d’étudier les interactions entre TrHbN et les membranes, des travaux théoriques ont

été menés par Julie-Anne Rousseau, étudiante à la maîtrise au laboratoire du Pr. Patrick

Lagüe. D'abord, des calculs d’hydrophathie menés avec l’outil MPEx [195] sur la structure

primaire de TrHbN prédisent que deux segments (résidus 32 à 50 et 85 à 103), formant

ensemble une vaste zone de la surface de TrHbN, sont membranaires. Différentes

trajectoires de TrHbN en présence d’une membrane explicite a permis d’étudier différents

degrés d’enfoncement de TrHbN à l’intérieur des membranes, les intéractions TrHbN :

phospholides et l’impact de la membrane sur la dynamique de TrHbN et ses tunnels. Ces

simulations ont permis de montrer, entre autres, que les phospholipides peuvent se placer à

l’intérieur des tunnels ce qui pourrait contribuer à l’encrage de TrHbN aux membranes. La

rédaction d’un mémoire de maîtrise décrivant ces travaux est en cours. D’autres travaux,

actuellement menés par une étudiante à la maîtrise au laboratoire du Pr. Michèle Auger,

visent à étudier par spectroscopie RMN solide les interactions entre les membranes et

TrHbN ainsi que le mutant TrHbN_Δpre-A. En accord avec les précédentes observations,

ces derniers travaux ont montré que TrHbN interagit avec les membranes et que la pre-A

joue un rôle important dans ces interactions (résultats non publiés).

La surface de TrHbN présente plusieurs régions apolaires. En particulier, on retrouve une

vaste région englobant l’entrée du tunnel court. En plus de l’implication de la pre-A, ces

surfaces hydrophobes pourraient contribuer à l’insertion membranaire. L’insertion de

TrHbN dans les membranes pourrait présenter certains avantages. En effet, les substrats

apolaires sont plus solubles dans les membranes que dans des solvants aqueux [196]. De ce

fait, TrHbN pourrait, via ses tunnels, puiser plus efficacement ses substrats dans les

membranes. Également, d’autres protéines membranaires dont certaines hémoglobines se

234

trouvent associées aux membranes tel que la cytochrome c oxydase [197]. La présence de

TrHbN pourrait permettre de jouer certains rôles en éliminant le •NO et empêchant

l’inhibition de certaines globines, en particulier certaines globines de la chaîne respiratoire.

10.8. Conclusion

Les travaux présentés dans cette thèse de doctorat visaient à étudier la structure et la

dynamique de TrHbN à l’aide de méthodes théoriques. En particulier, le rôle des tunnels de

TrHbN a fait l’objet d’une attention très pointue. Ces travaux ont permis de mettre en

évidence plusieurs éléments clefs de la dynamique de TrHbN et de ses tunnels. Ces

éléments renferment i) la présence de tunnels dynamiques grâce à la flexibilité des chaînes

latérales, ii) des routes de diffusions additionnelles à celles observées dans la structure

cristalline, iii) un squelette rigide, iv) l’affinité accrue des tunnels par rapport au solvant

pour les substrats gazeux et v) le tunnel Court est prédit comme la route de diffusion la plus

favorable. De plus, ces travaux théoriques ont permis démontrer que les substrats diffusent

à travers ces tunnels, de cavité en cavité, pour atteindre le site actif de TrHbN. Ces travaux

ont bénéficié d’un support expérimental important par l’étude de TrHbN et plusieurs

mutants caractérisés en cinétique, en spectroscopie de résonance Raman et en spectroscopie

RMN en solution. Avant les travaux présentés dans cette thèse, soit près de 10 ans après la

découverte et l’élucidation de la structure de TrHbN, il n’y avait aucune preuve

expérimentale démontrant le rôle fonctionnel des tunnels de TrHbN. L’ensemble des

travaux réalisés a permis plusieurs avancées scientifiques, mais a également soulevé de

nouvelles questions. Parmi ces interrogations, notons l’effet modeste inattendu des

mutations créées dans le but de bloquer les tunnels ainsi que la découverte des mouvements

lents (µs-ms) le long des hélices B et G mis en évidence par la spectroscopie RMN en

solution.

235

Bibliographie

1. World Health Orgnization report 2009, Global tuberculosis control. Epidemiology,

Strategy, Financing: Geneva.

2. Dye, C., Global epidemiology of tuberculosis. Lancet, 2006. 367(9514): p. 938-940.

3. Raviglione, M.C., The TB epidemic from 1992 to 2002. Tuberculosis, 2003. 83(1-

3): p. 4-14.

4. Zignol, M., et al., Global incidence of multidrug-resistant tuberculosis. Journal of

Infectious Diseases, 2006. 194(4): p. 479-485.

5. Russell, D.G., Who puts the tubercle in tuberculosis? Nat Rev Microbiol, 2007.

5(1): p. 39-47.

6. Deretic, V., et al., Mycobacterium tuberculosis inhibition of phagolysosome

biogenesis and autophagy as a host defence mechanism. Cellular Microbiology,

2006. 8(5): p. 719-727.

7. Russell, D.G., Mycobacterium tuberculosis: Here today, and here tomorrow. Nature

Reviews Molecular Cell Biology, 2001. 2(8): p. 569-577.

8. Russell, D.G., et al., Mycobacterium tuberculosis and the Four-Minute Phagosome

- By arresting the maturation of phagosomes, M-tuberculosis avoids being delivered

to lysosomes. Asm News, 2005. 71(10): p. 459-463.

9. Ulrichs, T. and S.H.E. Kaufmann, New insights into the function of granulomas in

human tuberculosis. Journal of Pathology, 2006. 208(2): p. 261-269.

10. Flynn, J.L. and J. Chan, What's good for the host is good for the bug. Trends in

Microbiology, 2005. 13(3): p. 98-102.

11. Algood, H.M.S., P.L. Lin, and J.L. Flynn, Tumor necrosis factor and chemokine

interactions in the formation and maintenance of granulomas in tuberculosis.

Clinical Infectious Diseases, 2005. 41: p. S189-S193.

12. Tsai, M.C., et al., Characterization of the tuberculous granuloma in murine and

human lungs: cellular composition and relative tissue oxygen tension. Cellular

Microbiology, 2006. 8(2): p. 218-232.

13. Akaki, T., et al., Effector molecules in expression of the antimicrobial activity of

macrophages against Mycobacterium avium complex: roles of reactive nitrogen

intermediates, reactive oxygen intermediates, and free fatty acids. Journal of

Leukocyte Biology, 1997. 62(6): p. 795-804.

14. Adams, L.B., et al., Comparison of the roles of reactive oxygen and nitrogen

intermediates in the host response to Mycobacterium tuberculosis using transgenic

mice. Tubercle and Lung Disease, 1997. 78(5-6): p. 237-246.

15. Rich, E.A., et al., Mycobacterium tuberculosis (MTB)-stimulated production of

nitric oxide by human alveolar macrophages and relationship of nitric oxide

production to growth inhibition of MTB. Tubercle and Lung Disease, 1997. 78(5-6):

p. 247-255.

16. Stewart, G.R., B.D. Robertson, and D.B. Young, Tuberculosis: A problem with

persistence. Nature Reviews Microbiology, 2003. 1(2): p. 97-105.

17. Samuni, U., et al., Ligand recombination and a hierarchy of solvent slaved

dynamics: The origin of kinetic phases in hemeproteins. Gene, 2007.

236

18. Gardner, P.R., et al., Nitric oxide sensitivity of the aconitases. Journal of Biological

Chemistry, 1997. 272(40): p. 25071-25076.

19. Stevanin, T.M., et al., Flavohemoglobin hmp affords inducible protection for

Escherichia coli respiration, catalyzed by cytochromes bo ' or bd, from nitric oxide.

Journal of Biological Chemistry, 2000. 275(46): p. 35868-35875.

20. Dedon, P.C. and S.R. Tannenbaum, Reactive nitrogen species in the chemical

biology of inflammation. Archives of Biochemistry and Biophysics, 2004. 423(1): p.

12-22.

21. Patel, R.P., et al., Biological aspects of reactive nitrogen species. Biochimica Et

Biophysica Acta-Bioenergetics, 1999. 1411(2-3): p. 385-400.

22. Ouellet, H., et al., Truncated hemoglobin HbN protects Mycobacterium bovis from

nitric oxide. Proc Natl Acad Sci U S A, 2002. 99(9): p. 5902-7.

23. Perutz, M.F., Regulation of Oxygen-Affinity of Hemoglobin - Influence of Structure

of the Globin on the Heme Iron. Annual Review of Biochemistry, 1979. 48: p. 327-

386.

24. Gardner, P.R., Nitric oxide dioxygenase function and mechanism of

flavohemoglobin, hemoglobin, myoglobin and their associated reductases. Journal

of Inorganic Biochemistry, 2005. 99(1): p. 247-266.

25. Tsai, A., How does NO activate hemeproteins? FEBS Lett, 1994. 341(2-3): p. 141-

5.

26. Doyle MP, H.J., Oxidation Of Nitrogen-Oxides By Bound Dioxygen In

Hemoproteins. Journal of Inoganic Biochemistry, 1981. 14(4): p. 351-358.

27. DeLano, W.L., The PyMOL Molecular Graphics System. 2002.

28. Vinogradov, S.N., et al., Three globin lineages belonging to two structural classes

in genomes from the three kingdoms of life. Proceedings of the National Academy

of Sciences of the United States of America, 2005. 102(32): p. 11385-11389.

29. Wittenberg, J.B., et al., Truncated hemoglobins: a new family of hemoglobins

widely distributed in bacteria, unicellular eukaryotes, and plants. J Biol Chem,

2002. 277(2): p. 871-4.

30. Vuletich, D.A. and J.T.J. Lecomte, A phylogenetic and structural analysis of

truncated hemoglobins. J Mol Evol, 2006. 62(2): p. 196-210.

31. Lecomte, J.T.J., D.A. Vuletich, and A.M. Lesk, Structural divergence and distant

relationships in proteins: evolution of the globins. Curr Opin Struct Biol, 2005.

15(3): p. 290-301.

32. Vinogradov, S.N. and L. Moens, Diversity of globin function: enzymatic, transport,

storage, and sensing. J Biol Chem, 2008. 283(14): p. 8773-8777.

33. Singel, D.J. and J.S. Stamler, Blood traffic control. Nature, 2004. 430(6997): p. 297.

34. Cossins, A. and M. Berenbrink, Physiology - Myoglobin's new clothes. Nature,

2008. 454(7203): p. 416-417.

35. Lundberg, J.O., E. Weitzberg, and M.T. Gladwin, The nitrate-nitrite-nitric oxide

pathway in physiology and therapeutics. Nature Reviews Drug Discovery, 2008.

7(2): p. 156-167.

36. Pesce, A., et al., A novel two-over-two alpha-helical sandwich fold is characteristic

of the truncated hemoglobin family. EMBO J, 2000. 19(11): p. 2424-34.

37. Hofacker, I. and K. Schulten, Oxygen and proton pathways in cytochrome c

oxidase. Proteins, 1998. 30(1): p. 100-107.

237

38. Milani, M., et al., Mycobacterium tuberculosis hemoglobin N displays a protein

tunnel suited for O2 diffusion to the heme. EMBO J, 2001. 20(15): p. 3902-9.

39. Pesce, A., et al., The 109 residue nerve tissue minihemoglobin from Cerebratulus

lacteus highlights striking structural plasticity of the alpha-helical globin fold.

Structure, 2002. 10(5): p. 725-735.

40. Nardini, M., et al., Archaeal protoglobin structure indicates new ligand diffusion

paths and modulation of haem-reactivity. Embo Reports, 2008. 9(2): p. 157-163.

41. Vallone, B., et al., The structure of murine neuroglobin: Novel pathways for ligand

migration and binding. Proteins, 2004. 56(1): p. 85-92.

42. Monod, J., J. Wyman, and J.P. Changeux, On nature of allosteric transitions - A

plausible model. Journal of Molecular Biology, 1965. 12(1): p. 88-&.

43. Huie, R.E. and S. Padmaja, The Reaction of NO with Superoxide. Free Radical

Research Communications, 1993. 18(4): p. 195-199.

44. Pryor, W.A. and G.L. Squadrito, The Chemistry of Peroxynitrite - A Product from

The Reaction of Nitric-Oxide with Superoxide. American Journal of Physiology-

Lung Cellular and Molecular Physiology, 1995. 268(5): p. L699-L722.

45. Tsai, J.H.M., et al., Role of conformation of peroxinitrite anion (ONOO-) in its

stability and toxicity. Journal of the American Chemical Society, 1994. 116(9): p.

4115-4116.

46. Miranda, K.M., M.G. Espey, and D.A. Wink, A discussion of the chemistry of

oxidative and nitrosative stress in cytotoxicity. Journal of Inorganic Biochemistry,

2000. 79(1-4): p. 237-240.

47. Pacher, P., J.S. Beckman, and L. Liaudet, Nitric Oxide and Peroxynitrite in Health

and Disease. Physiological Reviews, 2007. 87(1): p. 315-424.

48. Beckman, J.S., Oxidative damage and tyrosine nitration from peroxynitrite.

Chemical Research in Toxicology, 1996. 9(5): p. 836-844.


Truncated Hemoglobin. Journal of the American Chemical Society, 2010.

50. Gardner, P.R., et al., Nitric-oxide dioxygenase activity and function of

flavohemoglobins - Sensitivity to nitric oxide and carbon monoxide inhibition.

Journal of Biological Chemistry, 2000. 275(41): p. 31581-31587.

51. Frey, A.D., et al., Bacterial hemoglobins and flavohemoglobins for alleviation of

nitrosative stress in Escherichia coli. Applied and Environmental Microbiology,

2002. 68(10): p. 4835-4840.

52. Ouellet, H., et al., Reactions of Mycobacterium tuberculosis truncated hemoglobin

O with ligands reveal a novel ligand-inclusive hydrogen bond network.

Biochemistry, 2003. 42(19): p. 5764-74.

53. Minning, D.M., et al., Ascaris haemoglobin is a nitric oxide-activated deoxygenase.

Nature, 1999. 401(6752): p. 497-502.

54. Eich, R.F., et al., Mechanism of NO-induced oxidation of myoglobin and

hemoglobin. Biochemistry, 1996. 35(22): p. 6976-6983.

55. Takano, T., Structure of Myoglobin Refined at 2.0 A Resolution: I. Crystallographic

Refinement of Metmyoglobin from Sperm Whale. Journal of Molecular Biology,

1977. 110(3): p. 537-568.

56. Ouellet, Y., et al., Ligand interactions in the distal heme pocket of Mycobacterium

tuberculosis truncated hemoglobin N: roles of TyrB10 and GlnE11 residues.

Biochemistry, 2006. 45(29): p. 8770-81.

238

57. Pesce, A., et al., Thr-E11 regulates O2 affinity in Cerebratulus lacteus mini-

hemoglobin. J Biol Chem, 2004. 279(32): p. 33662-72.

58. Gibson, Q.H. and M.H. Smith, Rates of Reaction of Ascaris Haemoglobins with

Ligands. Proceedings of the Royal Society of London Series B-Biological Sciences,

1965. 163(991): p. 206-&.

59. Salter, M.D., et al., The apolar channel in cerebratulus lacteus hemoglobin is the

route for O2 entry and exit. J Biol Chem, 2008.

60. Das, T.K., et al., Ligand binding in the ferric and ferrous states of Paramecium

hemoglobin. Biochemistry, 2000. 39(46): p. 14330-14340.

61. Lu, C., et al., Structural and functional properties of a truncated hemoglobin from a

food-borne pathogen Campylobacter jejuni. J Biol Chem, 2007. 282(18): p. 13627-

13636.

62. Trent, J.T., et al., Crystallographic analysis of synechocystis cyanoglobin reveals

the structural changes accompanying ligand binding in a hexacoordinate

hemoglobin. J Mol Biol, 2004. 341(4): p. 1097-108.

63. Hamdane, D., et al., The Redox State of the Cell Regulates the Ligand Binding

Affinity of Human Neuroglobin and Cytoglobin. Journal of Biological Chemistry,

2003. 278(51): p. 51713-51721.

64. Perutz, M.F., Mechanisms Regulating the Reactions of Human Hemoglobin with

Oxygen and Carbon-Monoxide. Annual Review of Physiology, 1990. 52: p. 1-25.

65. Quillin, M.L., et al., Structural and Function-Effets of Apolar Mutations of the

Distal Valine in Myoglobin. Journal of Molecular Biology, 1995. 245(4): p. 416-

436.

66. Kendrew, J.C., et al., A Three-Dimensional Model of the Myoglobin Molecule

Obtained by X-Ray Analysis. Nature, 1958. 181(4610): p. 662-666.

67. Perutz, M.F., et al., Structure of Haemoglobin: A Three-Dimensional Fourier

Synthesis at 5.5-[angst]. Resolution, Obtained by X-Ray Analysis. Nature, 1960.

185(4711): p. 416-422.

68. Milani, M., et al., Cyanide binding to truncated hemoglobins: a crystallographic

and kinetic study. Biochemistry, 2004. 43(18): p. 5213-21.

69. Milani, M., et al., Heme-ligand tunneling in group I truncated hemoglobins. J Biol

Chem, 2004. 279(20): p. 21520-5.

70. Falzone, C.J., et al., The solution structure of the recombinant hemoglobin from the

cyanobacterium Synechocystis sp. PCC 6803 in its hemichrome state. J Mol Biol,

2002. 324(5): p. 1015-29.

71. Hoy, J.A., et al., The crystal structure of Synechocystis hemoglobin with a covalent

heme linkage. J Biol Chem, 2004. 279(16): p. 16535-42.

72. Milani, M., et al., A TyrCD1/TrpG8 hydrogen bond network and a TyrB10TyrCD1

covalent link shape the heme distal site of Mycobacterium tuberculosis hemoglobin

O. Proc Natl Acad Sci U S A, 2003. 100(10): p. 5766-5771.

73. Giangiacomo, L., et al., The truncated oxygen-avid hemoglobin from Bacillus

subtilis: X-ray structure and ligand binding properties. J Biol Chem, 2005. 280(10):

p. 9192-202.

74. Ilari, A., et al., Crystal structure and ligand binding properties of the truncated

hemoglobin from Geobacillus stearothermophilus. Arch Biochem Biophys, 2007.

457(1): p. 85-94.

239

75. Bonamore, A., et al., A novel thermostable hemoglobin from the actinobacterium

Thermobifida fusca. FEBS J, 2005. 272(16): p. 4189-201.

76. Pesce, A., et al., Structural characterization of a group II 2/2 hemoglobin from the

plant pathogen Agrobacterium tumefaciens. Biochimica et Biophysica Acta (BBA)

- Proteins & Proteomics. 1814(6): p. 810-816.

77. Nardini, M., et al., Structural determinants in the group III truncated hemoglobin

from campylobacter jejuni. J Biol Chem, 2006.

78. Nardini, M., et al., Protein fold and structure in the truncated (2/2) globin family.

Gene, 2007. 398(1-2): p. 2-11.

79. Richards, F.M., Interpretation of protein structures - Toal volume, group, volume

distributions and packing density. Journal of Molecular Biology, 1974. 82(1): p. 1-

14.

80. McCammon, J.A., B.R. Gelin, and M. Karplus, Dynamics of Folded Proteins.

Nature, 1977. 267(5612): p. 585-590.

81. Srajer, V., et al., Photolysis of the carbon monoxide complex of myoglobin:

Nanosecond time-resolved crystallography. Science, 1996. 274(5293): p. 1726-

1729.

82. Srajer, V. and W.E. Royer, Time-resolved x-ray crystallography of heme proteins.

Methods Enzymol, 2008. 437: p. 379-395.

83. Bourgeois, D., et al., Extended subnanosecond structural dynamics of myoglobin

revealed by Laue crystallography. Proc Natl Acad Sci U S A, 2006. 103(13): p.

4924-4929.

84. Carlson, M.L., R.M. Regan, and Q.H. Gibson, Distal cavity fluctuations in

myoglobin: Protein motion and ligand diffusion. Biochemistry, 1996. 35(4): p.

1125-1136.

85. Bossa, C., et al., Extended molecular dynamics simulation of the carbon monoxide

migration in sperm whale myoglobin. Biophysical Journal, 2004. 86(6): p. 3855-

3862.

86. Tilton, R.F., et al., Protein-ligand dynamics : A 96 picosecond simulation of a

myoglobin-xenon complex. Journal of Molecular Biology, 1988. 199(1): p. 195-211.

87. Kuczera, K., J. Kuriyan, and M. Karplus, Temperature dependence of the structure

and dynamics of myoglobin. A simulation approach. J Mol Biol, 1990. 213(2): p.

351-373.

88. Elber, R. and M. Karplus, Enhanced sampling in molecular-dynamics - Use of the

time-dependent Hartree approximation for a simulation of carbon-monoxide

diffusion through myoglobin. Journal of the American Chemical Society, 1990.

112(25): p. 9161-9175.

89. Cohen, J., et al., Imaging the migration pathways for O2, CO, NO, and Xe inside

myoglobin. Biophys J, 2006. 91(5): p. 1844-57.

90. Ruscio, J.Z., et al., Atomic level computational identification of ligand migration

pathways between solvent and binding site in myoglobin. Proceedings of the

National Academy of Sciences, 2008. 105(27): p. 9204-9209.

91. Amara, P., et al., Ligand diffusion in the catalase from Proteus mirabilis: a

molecular dynamics study. Protein Sci, 2001. 10(10): p. 1927-35.

92. Anselmi, M., et al., Molecular dynamics simulation of deoxy and carboxy murine

neuroglobin in water. Biophys J, 2007. 93(2): p. 434-441.

240

93. Johnson, B.J., et al., Exploring molecular oxygen pathways in Hansenula

polymorpha copper-containing amine oxidase. J Biol Chem, 2007. 282(24): p.

17767-17776.

94. Cohen, J., et al., Finding gas diffusion pathways in proteins: application to O2 and

H2 transport in CpI [FeFe]-hydrogenase and the role of packing defects. Structure,

2005. 13(9): p. 1321-1329.

95. Furse, K.E., et al., Molecular dynamics simulations of arachidonic acid-derived

pentadienyl radical intermediate complexes with COX-1 and COX-2: insights into

oxygenation regio- and stereoselectivity. Biochemistry, 2006. 45(10): p. 3206-3218.

96. Saam, J., et al., Molecular dioxygen enters the active site of 12/15-lipoxygenase via

dynamic oxygen access channels. Proc Natl Acad Sci U S A, 2007. 104(33): p.

13319-13324.

97. Vojtechovsky, J., et al., Crystal structures of myoglobin-ligand complexes at near-

atomic resolution. Biophysical Journal, 1999. 77(4): p. 2153-2174.

98. Miles, E.W., S. Rhee, and D.R. Davies, The molecular basis of substrate

channeling. Journal of Biological Chemistry, 1999. 274(18): p. 12193-12196.

99. Raushel, F.M., J.B. Thoden, and H.M. Holden, Enzymes with molecular tunnels.

Accounts of Chemical Research, 2003. 36(7): p. 539-548.

100. Hyde, C.C., et al., 3-Dimensional structure of the tryptophan synthase alpha-2-

beta-2-multienzyme complex from salmonella-typhimurium. Journal of Biological

Chemistry, 1988. 263(33): p. 17857-17871.

101. Darnault, C., et al., Ni-Zn-[Fe-4-S-4] and Ni-Ni-[Fe-4-S-4] clusters in closed and

open subunits of acetyl-CoA synthase/carbon monoxide dehydrogenase. Nature

Structural Biology, 2003. 10(4): p. 271-279.

102. Thoden, J.B., et al., Structure of Carbamoyl Phosphate Synthetase: A Journey of 96

A from Substrate to Productâ€ ,â€¡. Biochemistry, 1997. 36(21): p. 6305-6316.

103. Muchmore, C.R.A., et al., Crystal structure of glutamine

phosphoribosylpyrophosphate amidotransferase from Escherichia coli. Protein

Science, 1998. 7(1): p. 39-51.

104. Larsen, T.M., et al., Three-Dimensional Structure of Escherichia coli Asparagine

Synthetase B: A Short Journey from Substrate to Productâ€ ,â€¡. Biochemistry,

1999. 38(49): p. 16146-16157.

105. Binda, C., et al., Cross-talk and ammonia channeling between active centers in the

unexpected domain arrangement of glutamate synthase. Structure, 2000. 8(12): p.

1299-1308.

106. Chaudhuri, B.N., et al., Crystal structure of imidazole glycerol phosphate synthase:

A tunnel through a (beta/alpha)(8) barrel joins two active sites. Structure, 2001.

9(10): p. 987-997.

107. Teplyakov, A., et al., Channeling of ammonia in glucosamine-6-phosphate

synthase. Journal of Molecular Biology, 2001. 313(5): p. 1093-1102.

108. Li, R.B., J.P. Klinman, and F.S. Mathews, Copper amine oxidase from Hansenula

polymorpha: the crystal structure determined at 2.4 angstrom resolution reveals the

active conformation. Structure, 1998. 6(3): p. 293-307.

109. Fontecilla-Camps, J.C., et al., Structure/function relationships of [NiFe]- and

[FeFe]-hydrogenases. Chemical Reviews, 2007. 107(10): p. 4273-4303.

110. Boyington, J.C., B.J. Gaffney, and L.M. Amzel, The three-dimensional structure of

an arachidonic acid 15-lipoxygenase. Science, 1993. 260(5113): p. 1482-1486.

241

111. Furse, K.E., et al., Molecular dynamics simulations of arachidonic acid complexes

with COX-1 and COX-2: insights into equilibrium behavior. Biochemistry, 2006.

45(10): p. 3189-3205.

112. Minor, W., et al., Crystal structure of soybean lipoxygenase L-l at 1.4 angstrom

resolution. Biochemistry, 1996. 35(33): p. 10687-10701.

113. Lario, P.I., N. Sampson, and A. Vrielink, Sub-atomic resolution crystal structure of

cholesterol oxidase: What atomic resolution crystallography reveals about enzyme

mechanism and the role of the FAD cofactor in redox activity. Journal of Molecular

Biology, 2003. 326(5): p. 1635-1650.

114. Coulombe, R., et al., Oxygen access to the active site of cholesterol oxidase through

a narrow channel is gated by an Arg-Glu pair. J Biol Chem, 2001. 276(32): p.

30435-30441.

115. Petrek, M., et al., CAVER: a new tool to explore routes from protein clefts, pockets

and cavities. BMC Bioinformatics, 2006. 7: p. 316.

116. Pesce, A., et al., Crystallization and preliminary X-ray analysis of neural

haemoglobin from the nemertean worm Cerebratulus lacteus. Acta

Crystallographica Section D-Biological Crystallography, 2001. 57: p. 1897-1899.

117. Couture, M., et al., A cooperative oxygen-binding hemoglobin from Mycobacterium

tuberculosis. Proc Natl Acad Sci U S A, 1999. 96(20): p. 11223-8.

118. Thorsteinsson, M.V., et al., A cyanobacterial hemoglobin with unusual ligand

binding kinetics and stability properties. Biochemistry, 1999. 38(7): p. 2117-2126.

119. Brunori, M., et al., The role of cavities in protein dynamics: crystal structure of a

photolytic intermediate of a mutant myoglobin. Proc Natl Acad Sci U S A, 2000.

97(5): p. 2058-2063.

120. Lakowicz, J.R. and G. Weber, Quenching of protein fluorescence by oxygen.

Detection of structural fluctuations in proteins on nanosecond time scale.

Biochemistry, 1973. 12(21): p. 4171-4179.

121. Perutz, M.F. and F.S. Mathews, An X-ray study of azide methaemoglobin. Journal of

Molecular Biology, 1966. 21(1): p. 199-202.

122. Tilton, R.F., I.D. Kuntz, and G.A. Petsko, Cavities in proteins: structure of a

metmyoglobin-xenon complex solved to 1.9 A. Biochemistry, 1984. 23(13): p. 2849-

2857.

123. Scott, E.E., Q.H. Gibson, and J.S. Olson, Mapping the pathways for O-2 entry into

and exit from myoglobin. Journal of Biological Chemistry, 2001. 276(7): p. 5177-

5188.

124. Olson, J.S., J. Soman, and G.N. Phillips, Ligand pathways in myoglobin: a review

of Trp cavity mutations. IUBMB Life, 2007. 59(8-9): p. 552-562.

125. Leroux, F., et al., Experimental approaches to kinetics of gas diffusion in

hydrogenase. Proc Natl Acad Sci U S A, 2008.

126. Gibson, Q.H., Hemoproteins, Ligands, and Quanta. Journal of Biological

Chemistry, 1989. 264(34): p. 20155-20158.

127. Tan, X.S., et al., The tunnel of acetyl-coenzyme a synthase/carbon monoxide

dehydrogenase regulates delivery of CO to the active site. Journal of the American

Chemical Society, 2005. 127(16): p. 5833-5839.

128. Chen, L., et al., The Binding and Release of Oxygen and Hydrogen Peroxide Are

Directed by a Hydrophobic Tunnel in Cholesterol Oxidase. Biochemistry, 2008.

47(19): p. 5368-5377.

242

129. Salomonsson, L., et al., A single-amino-acid lid renders a gas-tight compartment

within a membrane-bound transporter. Proceedings of the National Academy of

Sciences of the United States of America, 2004. 101(32): p. 11617-11621.

130. Tan, X., et al., Function of the tunnel in acetylcoenzyme A synthase/carbon

monoxide dehydrogenase. Journal Biological Inorganic Chemistry, 2006. 11(3): p.

371-378.

131. Riistama, S., et al., Channelling of dioxygen into the respiratory enzyme. Biochimia

Biophysica Acta, 1996. 1275(1-2): p. 1-4.

132. Riistama, S., et al., Binding of O2 and its reduction are both retarded by

replacement of valine 279 by isoleucine in cytochrome c oxidase from Paracoccus

denitrificans. Biochemistry, 2000. 39(21): p. 6365-6372.

133. Schlichting, I., et al., Crystal-structure of photolyzed carbonmonoxy-myoglobin.

Nature, 1994. 371(6500): p. 808-812.

134. Schlichting, I. and K. Chu, Trapping intermediates in the crystal: ligand binding to

myoglobin. Current Opinion in Structural Biology, 2000. 10(6): p. 744-752.

135. Ostermann, A., et al., Ligand binding and conformational motions in myoglobin.

Nature, 2000. 404(6774): p. 205-208.

136. Srajer, V., et al., Protein conformational relaxation and ligand migration in

myoglobin: A nanosecond to millisecond molecular movie from time-resolved Laue

X-ray diffraction. Biochemistry, 2001. 40(46): p. 13802-13815.

137. Palmer, A.G., NMR characterization of the dynamics of biomacromolecules.

Chemical Reviews, 2004. 104(8): p. 3623-3640.

138. Case, D.A. and M. Karplus, Dynamics of Ligand-Binding ti Heme-Proteins. Journal

of Molecular Biology, 1979. 132(3): p. 343-368.

139. Karplus, M. and J.A. McCammon, Molecular dynamics simulations of

biomolecules. Nature Structural Biology, 2002. 9(9): p. 646-652.

140. Alder, B.J. and T.E. Wainwright, Phase transition for a hard sphere system. Journal

of Chemical Physics, 1957. 27(5): p. 1208-1209.

141. Alder, B.J. and T.E. Wainwright, Studies in molecular dynamics .I. General

method. Journal of Chemical Physics, 1959. 31(2): p. 459-466.

142. Rahman, A., Correlations in the Motion of Atoms in Liquid Argon. Physical

Review, 1964. 136(2A): p. A405.

143. Stilling.Fh and A. Rahman, Improved simulation of liquid water by molecular

dynamics. Journal of Chemical Physics, 1974. 60(4): p. 1545-1557.

144. Deisenhofer, J. and W. Steigemann, Crystallographic Refinement of Structure of

Bovine Pancreatic Trypsin-Inhibithor at 1.5 A Resolution. Acta Crystallographica

Section B-Structural Science, 1975. B 31(JAN15): p. 238-250.

145. Brooks, B. and M. Karplus, Normal-Modes for Specific Motions of Macromolecules

- Application to the Hinge-Bending Mode of Lysozyme. Proceedings of the National

Academy of Sciences of the United States of America, 1985. 82(15): p. 4995-4999.

146. Colonna-Cesari, F., et al., Interdomain Motion in Liver Alcohol-Dehydrogenase -

Structural and Energetic Analysis of the Hinge Bending Mode. Journal of

Biological Chemistry, 1986. 261(32): p. 5273-5280.

147. Levy, R.M., M. Karplus, and P.G. Wolynes, NMR Relaxation Parameters in

Molecules With Internal Motion - Exact Langevin Trajectory Results Compared

With Simplified Relaxation Models. Journal of the American Chemical Society,

1981. 103(20): p. 5998-6011.

243

148. Olejniczak, E.T., et al., Motional Averaging of Proton Nuclear Overhauser Effects

in Proteins - Predictions from a Molecular-Dynamics Simulations of Lysozyme.

Journal of the American Chemical Society, 1984. 106(7): p. 1923-1930.

149. Brunger, A.T., C.L. Brooks, and M. Karplus, Active-Site Dynamics of Ribonuclease.

Proceedings of the National Academy of Sciences of the United States of America,

1985. 82(24): p. 8458-8462.

150. Nadler, W., et al., Molecular and Stochastic Dynamics of Proteins. Proceedings of

the National Academy of Sciences of the United States of America, 1987. 84(22): p.

7933-7937.

151. Frauenfelder, H., et al., Thermal-Expansion of a Protein. Biochemistry, 1987.

26(1): p. 254-261.

152. Schlick, T., Molecular Modeling and Simulation. 2002: Springer.

153. Kettering, C.F., L.W. Shutts, and D.H. Andrews, A representation of the dynamic

properties of molecules by mechanical models. Physical Review, 1930. 36(3): p.

0531-0543.

154. Born, M. and R. Oppenheimer, Zur Quantentheorie der Molekeln. Annalen der

Physik, 1927. 389(20): p. 457-484.

155. MacKerell, A.D., et al., All-Atom Empirical Potential for Molecular Modeling and

Dynamics Studies of Proteins. Journal of Physical Chemistry B, 1998. 102(18): p.

3586-3616.

156. Brooks, B.R., et al., CHARMM: A program for macromolecular energy,

minimization, and dynamics calculations. Journal of Computational Chemistry,

1983. 4(2): p. 187-217.

157. Lovell, S.C., et al., The penultimate rotamer library. Proteins, 2000. 40(3): p. 389-

408.

158. Steinbach, P.J. and B.R. Brooks, New spherica-cutoff methods for long-range forces

in macromolecular simulation. Journal of Computational Chemistry, 1994. 15(7): p.

667-683.

159. Koehl, P., Electrostatics calculations: latest methodological advances. Current

Opinion in Structural Biology, 2006. 16(2): p. 142-151.

160. Darden, T., D. York, and L. Pedersen, Particle mesh Ewald: An N·log(N) method

for Ewald sums in large systems. The Journal of Chemical Physics, 1993. 98(12): p.

10089-10092.

161. Sagui, C. and T.A. Darden, Molecular dynamics simulations of biomolecules: Long-

range electrostatic effects. Annual Review of Biophysics and Biomolecular

Structure, 1999. 28: p. 155-179.

162. Ryckaert, J.-P., G. Ciccotti, and H.J.C. Berendsen, Numerical integration of the

cartesian equations of motion of a system with constraints: molecular dynamics of

n-alkanes. Journal of Computational Physics, 1977. 23(3): p. 327-341.

163. Verlet, L., Computer experiments on Classical Fluids .I. Thermodynamical

Properties of Lennard-Jones Molecules. Physical Review, 1967. 159(1): p. 98-&.

164. Petrek, M., et al., MOLE: a Voronoi diagram-based explorer of molecular channels,

pores, and tunnels. Structure, 2007. 15(11): p. 1357-1363.

165. Smart, O.S., et al., HOLE: A program for the analysis of the pore dimensions of ion

channel structural models. Journal of Molecular Graphics & Modelling, 1996.

14(6): p. 354-&.

244

166. Dijkstra, E.W., A note on two problems in connexion with graphs. Numerische

Mathematik, 1959. 1(1): p. 269-271.

167. Kleywegt, G.J. and T.A. Jones, Detection, delineation, measurement and display of

cavities in macromolecular structures. Acta Crystallogr D Biol Crystallogr, 1994.

50(Pt 2): p. 178-85.

168. Humphrey, W., A. Dalke, and K. Schulten, VMD: Visual molecular dynamics.

Journal of Molecular Graphics, 1996. 14(1): p. 33-&.

169. Anselmi, M., A. Di Nola, and A. Amadei, The Kinetics of Ligand Migration in

Crystallized Myoglobin as Revealed by Molecular Dynamics Simulations. Biophys.

J., 2008. 94(11): p. 4277-4281.

170. Daigle, R., M. Guertin, and P. Lague, Structural characterization of the tunnels of


simulations. Proteins-Structure Function and Bioinformatics, 2009. 75(3): p. 735-

747.

171. Daigle, R., et al., Theoretical Investigations of Nitric Oxide Channeling in

Mycobacterium tuberculosis Truncated Hemoglobin N. Biophysical Journal, 2009.

97(11): p. 2967-2977.

172. Golden, S.D., K.W. Olsen, and K.P. Robert, Identification of Ligand-Binding

Pathways in Truncated Hemoglobins Using Locally Enhanced Sampling Molecular

Dynamics, in Methods in Enzymology. 2008, Academic Press. p. 459-475.

173. Milani, M., et al., Cyanide binding to truncated hemoglobins: a crystallographic

and kinetic study. Biochemistry, 2004. 43(18): p. 5213-21.

174. Milani, M., et al., Mycobacterium tuberculosis hemoglobin N displays a protein

tunnel suited for O2 diffusion to the heme. EMBO Journal, 2001. 20(15): p. 3902-9.

175. Ouellet, Y., et al., Ligand interactions in the distal heme pocket of Mycobacterium

tuberculosis truncated hemoglobin N: roles of TyrB10 and GlnE11 residues.

Biochemistry, 2006. 45(29): p. 8770-81.

176. Cowan, D.A., Thermophilic proteins: Stability and function in aqueous and organic

solvents. Comparative Biochemistry and Physiology a-Molecular & Integrative

Physiology, 1997. 118(3): p. 429-438.

177. Liang, X.Y., et al., Apparent Tradeoff of Higher Activity in MMP-12 for Enhanced

Stability and Flexibility in MMP-3. Biophysical Journal, 2010. 99(1): p. 273-283.

178. Walters, J., et al., Thermodynamic, enzymatic and structural effects of removing a

salt bridge at the base of loop 4 in (pro)caspase-3. Archives of Biochemistry and

Biophysics, 2011. 508(1): p. 31-38.

179. Tian, J., et al., Enhanced thermostability of methyl parathion hydrolase from

Ochrobactrum sp. M231 by rational engineering of a glycine to proline mutation.

FEBS Journal, 2010. 277(23): p. 4901-4908.

180. Mishra, S. and M. Meuwly, Nitric oxide dynamics in truncated hemoglobin:

docking sites, migration pathways, and vibrational spectroscopy from molecular

dynamics simulations. Biophysical Journal, 2009. 96(6): p. 2105-2018.

181. Bidon-Chanal, A., et al., Ligand-induced dynamical regulation of NO conversion in

Mycobacterium tuberculosis truncated hemoglobin-N. Proteins, 2006. 64(2): p. 457-

64.

182. Crespo, A., et al., Theoretical study of the truncated hemoglobin HbN: exploring the

molecular basis of the NO detoxification mechanism. Journal American Chemistry

Society, 2005. 127(12): p. 4433-44.

245


Truncated Hemoglobin. Journal of the American Chemical Society, 2010. 132(9): p.

2968-2982.

184. Milani, M., et al., Heme-ligand tunneling in group I truncated hemoglobins. J Biol

Chem, 2004. 279(20): p. 21520-5.

185. Vos, M.H., Ultrafast dynamics of ligands within heme proteins. Biochimica Et

Biophysica Acta-Bioenergetics, 2008. 1777(1): p. 15-31.

186. Salter, M.D., et al., The apolar channel in cerebratulus lacteus hemoglobin is the

route for O2 entry and exit. Journal Biological Chemistry, 2008. 283(51): p. 35689-

35702.

187. Via, L.E., et al., Tuberculous granulomas are hypoxic in guinea pigs, rabbits, and

nonhuman primates. Infection and Immunity, 2008. 76(6): p. 2333-2340.

188. Shi, L., et al., Transcriptional characterization of the antioxidant response of

Mycobacterium tuberculosis in vivo and during adaptation to hypoxia in vitro.

Tuberculosis, 2008. 88(1): p. 1-6.

189. Ouellet, H., et al., Truncated hemoglobin HbN protects Mycobacterium bovis from

nitric oxide. Proceedings of the National Academy of Sciences of the United States

of America, 2002. 99(9): p. 5902-7.

190. Baron, R., et al., Multiple pathways guide oxygen diffusion into flavoenzyme active

sites. Proceedings of the National Academy of Sciences of the United States of

America, 2009. 106(26): p. 10603-10608.

191. Scott, E.E., Q.H. Gibson, and J.S. Olson, Mapping the pathways for O2 entry into

and exit from myoglobin. Journal of Biological Chemistry, 2001. 276(7): p. 5177-

5188.

192. Orlowski, S. and W. Nowak, Oxygen diffusion in minihemoglobin from

Cerebratulus lacteus: a locally enhanced sampling study. Theoretical Chemistry

Accounts, 2007. 117(2): p. 253-258.

193. Pesce, A., et al., Protein structure in the truncated (2/2) hemoglobin family.

IUBMB Life, 2007. 59(8-9): p. 535-541.

194. Tomita, A., et al., Visualizing breathing motion of internal cavities in concert with

ligand migration in myoglobin. Proceedings of the National Academy of Sciences

of the United States of America, 2009. 106(8): p. 2612-2616.

195. Snider, C., et al., MPEx: A tool for exploring membrane proteins. Protein Science,

2009. 18(12): p. 2624-2628.

196. Wang, Y., et al., Exploring gas permeability of cellular membranes and membrane

channels with molecular dynamics. J Struct Biol, 2007. 157(3): p. 534-544.

197. Sinibaldi, F., et al., Extended cardiolipin anchorage to cytochrome c: a model for

protein-mitochondrial membrane binding. Journal of Biological Inorganic

Chemistry. 15(5): p. 689-700.

246

Annexe 1

Matériel supplémentaire du chapitre 6

Figure S1. CHARMM atom types for the iron-porphyrin ring used in the force field

parameters optimization. The heme-bound oxygen molecule, which uses the OM atom type

for both oxygen atoms, is omitted for clarity. The parameters in Table S1 and Table S2 are

given in terms of atom types defined here.

247

Figure S2. Various Gln58(E11) rotamers from typical MD frames for oxy-TrHbN (top) and

deoxyTrHbN (bottom). The (a) tp-100º, (b) mm100º, (c) mt-30º and the (d) tt0º rotamers

are shown. The heme, Phe62(E15), Gln58(E11), Tyr33(B10) and the proximal histidine

His81(F8) are represented with balls and sticks. Hydrogen bonds are represented by dashed

lines with their corresponding length. Pictures were generated using PyMOL (41).

248

Figure S3. Stereo view of the active site configurations observed from typical MD frames

for oxy-TrHbN (top) and deoxy-TrHbN (bottom). The heme, Phe62(E15), Gln58(E11),

Tyr33(B10) and the proximal histidine His81(F8) are represented with balls and sticks.

Hydrogen bonds are represented by dashed lines. Additionnal pictures from different

Gln58(E11) rotamers are available in the supplemental figure S2. Pictures were generated

using PyMOL (41).

249

Table S1. Atomic charges according to the CHARMM atom type

CHARMM atom type CHARMM22 Optimized charges Cytochrome c*

FE 0.24 1.42 1.20

NPH -0.18 -0.65 -0.76

CPA 0.12 0.42 0.32

CPB -0.06 -0.12 0.07

CPM -0.10 -0.27 -0.28

OM (proximal) 0.02 -0.18 -

OM (distal) -0.02 -0.32 -

* Reduced heme with methionine as bound substrate, see Chapter 6, reference #12.

Table S2. Optimized and CHARMM22 OM-OM-FE angle parameter.

Angle (º) Force constant

Optimized 122.22 13.18

CHARMM22 180.00 0.00

250

Annexe 2


Simulation details

MD simulation were performed using CHARMM (2) with the CHARMM22 all-atom

potential energy parameter set (3) with phi, psi cross term map correction CMAP (4), and

modified TIP3P waters (5). Optimized oxygenated heme atomic charges and Fe-O-O angle

parameters were used (1). A 3-point model was used for free •NO molecule and the force

field parameters were derived from ref. (6). This model takes into account the dipole of the

molecule, and accurately reproduces solvation energy (More details on •NO and water

force fields are given in the next section). Electrostatic interactions were calculated via the

Particle Mesh Ewald method (7), using a sixth-order spline interpolation for

complementary function, with κ=0.34 Å-1

, and a fast-Fourier grid density of ≈1 Å-1

. Cutoffs

for the real space portion of the Particle Mesh Ewald calculation and the truncation of the

Lennard-Jones interactions were 10 Å, with the latter smoothed via a shifting function over

the range of 8 Å to 10 Å. The SHAKE algorithm (8) was used to constrain all covalent

bonds involving hydrogen atoms. All simulations employed the leapfrog algorithm and an

integration step of 1 femtosecond. Coordinates were saved every picosecond (ps) for

analysis. Nonbond and image lists were updated heuristically. All simulations were

performed at constant pressure and temperature (NPT ensemble) using Hoover algorithm

for temperature control (9). The mass of the thermal piston was 20,000 kcal•mol-1

•ps2 and

the mass of the pressure piston equaled 1000 amu. All simulations were carried out at 298

K and 1 atm. The net translation and rotation of the systems were removed every 10,000

steps.

251

Ligand binding affinities

The binding affinities of the ligands were estimated from the PMF maps obtained from the

ILS calculations as follows. The equilibrium constant Kb of the binding reaction L+P ⇌ LP

is defined as Kb = [LP]/([L][P]), where [L], [P] and [LP] are the equilibrium concentrations

of the unbound ligand, unbound protein, and bound complex, respectively. The standard

binding free energy is defined from the equilibrium constant by ΔGbº= -kBT ln (CºKb),

where kB is the Boltzmann constant, T the temperature, and Cº is a standard concentration.

Here, the equilibrium association constants were calculated using the PMFs grid

maps obtained from ILS calculations. The relation between the equilibrium constant and

the PMF is given by (10, 11):

∫ [ ( ) ( )]

where r represents the positions of the grid maps, β=1/kBT, w the PMF between the ligand

and the protein, and r' is a reference position far away in the bulk. In our calculations, w(r')

was calculated as the average PMF of the ligand in a water box. The sums were executed

using Simpson's rule. Errors on affinities were estimated from 5 different ILS maps. Local

binding affinities were calculated at the protein surface and inside the tunnels by integrating

over selected grids points. For the surface binding affinity, a 4.5 Å layer of solvent was

considered.

Unless otherwise noted in the main text, all affinity numbers given in this study refer to

Kd=1/Kb.

252

Table S1: Force field parameters for nitric oxide molecule

Molecules Electrostatic Van der Waals

•NO three points model qN = -0.250e

qO = -0.345e

qcom = 0.595e

εN = 0.20 kcal/mol

εO = 0.16 kcal/mol

rN = 2.00 Å

rO = 2.05 Å

The •NO solvation free energies was determined by free energy perturbation molecular

dynamics (FEP-MD) (12). Simulations details are given in the preceding section. The

simulated system consists of cubic box composed of 500 water molecules and one •NO

molecule. Parameters for •NO are given in table S1. 8 FEP-MD simulations were

performed; 4 for both forward and reverse reactions (where NO disappears and appears,

respectively). These calculations returned •NO solvation free energies of 1.6±0.2 kcal/mol,

in agreement with the experimentally measured free energy of 1.53 kcal/mol at 298.15 K

and 1 ATM (13).

Table S2 : •NO and O2 parameters used for implicit ligand sampling and resulting solvation

free energies

Ligands Van der Waals Bond length

(Å)

NO solvation energies

•NO εN = -0.20 rmin/2 = 1.85 Å

εO = -0.12 rmin/2 = 1.70 Å

1.15 1.37±0.03

O2 εO = -0.12 rmin/2 = 1.70 Å 1.21*

1.91±0.03

* In the original paper of Cohen et al. presenting ILS method (14), the O2 bond length

parameter suggested is 1.12 Å. Using this setting, we were unable to reproduce published

value of 1.97 kcal/mol. We contacted Klaus Schulten group (developers of the ILS method)

about this discrepancy and after few communications, they suggested us to use a bond

length of 1.21 Å. This length is close to the experimental value of 1.208 Å published in the

CRC Handbook of Chemistry (pages 9-82 and 9-98) and allows to obtain O2 solvations

energy close to the expected value. The small differences arise from simulation conditions,

like to the use (or not) of rattle to constrain covalent bonds involving hydrogen.

253

Figure S1. Gradient of affinity for nitric oxide calculated as function of the position at the

surface of wild type TrHbN (top) and polar entrance mutant (bottom). Location of the

surface entrances are indicated by magenta arrows. NO affinity was estimated using the

PMF maps obtained from the ILS calculations (see methods section). First, about 5100

coordinates were selected from ILS grid maps. These points surround the entire protein

surface at a distance of about 2.25 Å. For each point, affinity is calculated by integrating

over all PMF grid points located closer than 2.25 Å (spheres of ~48 Å3). Each point is

colored according to the calculated affinity.

254

Three addition molecular movies built from MD simulations snapshots are accessible at the

following internet link:

http://www.cell.com/biophysj/supplemental/S0006-3495(09)01450-7

Movie 1: xe1_ehc_xe2.wmv

Animation presenting a •NO molecule diffusing from LT entrance to Xe2 via EHc.

(LT entrance → Xe1 → EHc → Xe2).

Movie 2 : NO_entering_ST.wmv

Animation presenting a •NO molecule diffusing from the ST entrance to Xe2.

Ile119(H11) side chain is shown with sticks.

Movie 3 : NO_entering_EHT.wmv

Animation presenting a •NO molecule entering EHT and reaching EHc.

255

References

1. Daigle, R., M. Guertin, and P. Lagüe. 2009. Structural characterization of the

tunnels of Mycobacterium tuberculosis truncated hemoglobin N from molecular

dynamics simulations. Proteins-Structure Function and Bioinformatics 75:735-747.

2. Brooks, B. R., R. E. Bruccoleri, B. D. Olafson, D. J. States, S. Swaminathan, and M.

Karplus. 1983. CHARMM: A program for macromolecular energy, minimization,

and dynamics calculations. Journal of Computational Chemistry 4:187-217.

3. MacKerell, A. D., D. Bashford, M. Bellott, R. L. Dunbrack, J. D. Evanseck, M. J.

Field, S. Fischer, J. Gao, H. Guo, S. Ha, D. Joseph-McCarthy, L. Kuchnir, K.

Kuczera, F. T. K. Lau, C. Mattos, S. Michnick, T. Ngo, D. T. Nguyen, B. Prodhom,

W. E. Reiher, B. Roux, M. Schlenkrich, J. C. Smith, R. Stote, J. Straub, M.

Watanabe, J. Wiorkiewicz-Kuczera, D. Yin, and M. Karplus. 1998. All-Atom

Empirical Potential for Molecular Modeling and Dynamics Studies of Proteins.

Journal of Physical Chemistry B 102:3586-3616.

4. Mackerell, A. D., M. Feig, and C. L. Brooks. 2004. Extending the treatment of

backbone energetics in protein force fields: limitations of gas-phase quantum

mechanics in reproducing protein conformational distributions in molecular

dynamics simulations. J Comput Chem 25:1400-1415.

5. Price, D. J., and C. L. B. III. 2004. A modified TIP3P water potential for simulation

with Ewald summation. The Journal of Chemical Physics 121:10096-10103.

6. Meuwly, M., O. M. Becker, R. Stote, and M. Karplus. 2002. NO rebinding to

myoglobin: a reactive molecular dynamics study. Biophys Chem 98:183-207.

7. Darden, T., D. York, and L. Pedersen. 1993. Particle mesh Ewald: An N·log(N)

method for Ewald sums in large systems. The Journal of Chemical Physics

98:10089-10092.

8. Ryckaert, J.-P., G. Ciccotti, and H. J. C. Berendsen. 1977. Numerical integration of

the cartesian equations of motion of a system with constraints: molecular dynamics

of n-alkanes. Journal of Computational Physics 23:327-341.

9. Feller, S. E., Y. Zhang, R. W. Pastor, and B. R. Brooks. 1995. Constant pressure

molecular dynamics simulation: The Langevin piston method. The Journal of

Chemical Physics 103:4613-4621.

10. Chandler, D., and L. R. Pratt. 1976. Statistical mechanics of chemical equilibria and

intramolecular structures of nonrigid molecules in condensed phases. The Journal of


11. Simonson, B. R. a. T. 1999. Implicit solvent models. Biophys Chem 78:1-20.

256

12. Kollman, P. 1993. Free-Energy Calculations - Applications to chemical and

biochemical phenomena. Chemical Reviews 93:2395-2417.

13. Scharlin, P., R. Battino, E. Silla, I. Tunon, and J. L. Pascual-Ahuir. 1998. Solubility

of gases in water: Correlation between solubility and the number of water molecules

in the first solvation shell. Blackwell Science Ltd. 1895-1904.

14. Cohen, J., A. Arkhipov, R. Braun, and K. Schulten. 2006. Imaging the migration

pathways for O2, CO, NO, and Xe inside myoglobin. Biophys J 91:1844-1857.

257

Annexe 3


Methods

NMR

Protein expression, labeling, and purification - Construction of pET3a plasmid

containing the glbN gene lacking the pre-A helix. We used the PCR to amplify the coding

region lacking residues 1 to 12 from the pET3a plasmid containing the glbN gene (1). The

DNA primers used were 5'-GACCCATATGATCAGCATCTACGACAAGATCGG-3'

(upper primer) and 5'-GATGGATCCTCAGACTGGTGCCGTGGTG-3' (lower primer).

Protein expression and labeling. E. coli BL21(DE3) cells were transformed with pET3a

plasmid containing glbN (1) or ΔpreA-glbN gene. Transformed cells were spread on LB

agar plates containing 200 μg/ml ampicilin (Roche Diagnostic) and grown overnight at 37

°C. Ten colonies were used to inoculate 125 ml LB medium supplemented with 200 μg/ml

ampicillin. The culture was incubated at 37°C until OD600 nm reached 0.5-0.6. Ten milliliters

of the preculture was used to inoculate 1L of M9 medium (48 mM Na2HPO4, 22 mM

KH2PO4, 8,5 mM NaCl, 18,7 mM NH4Cl 20 mM glucose, 2 mM MgSO4, 2 μM FeCl3, 100

μM CaCl2, 50 μM ZnSO4, 7.8 μM Thiamin-HCl, BME vitamin solution (Sigma Aldrich)

diluted 1/2000 and 200 μg/ml ampicillin) in a 3L Vessel (BioFlo 110 fermentor, New

Brunswick Scientific CO., Edison, NJ). Cells were grown at 37 °C with agitation.

Dissolved oxygen was maintained at 30 % and pH was maintained between 7.2 and 7.5.

When OD600 nm reached 0.9, cells were induced with 0.5 mM IPTG for 16-18h at 25 °C

with same conditions for dissolved oxygen and pH 15

NH4Cl (Cambridge isotope laboratory,

Andover, MA) was added to obtain 15

N-labeled protein. For 13

C15

N labeled protein, (13

C)-

Glucose (Cambridge isotope laboratory, Andover, MA) was used.

Purification of apo-TrHbN. Cell suspension was centrifuged 15 min at 5,000 x g, 4 °C. Cell

pellets were suspended in lysis buffer (50 mM Tris, 1 mM PMSF, pH 7.5) supplemented

with DNase I (1 μg/g wet cell) (Roche Diagnostic) and broken by passing them twice

through a French pressure cell operated at 20, 000 psi. The cell lysate was centrifuged 15

258

min at 4 °C at 16,000 x g. The supernatant was discarded and the pellet was washed twice

with 60 ml wash buffer (5mM Tris, 50 μM EDTA, 1mM PMSF and 0.5% triton X-100) at 4

°C and centrifuged 15 min at 4 °C at 16,000 x g. Inclusion bodies containing apo-TrHbN

were solubilized at room temperature with 40 mL urea solution (8 M deionized urea, 50

mM Tris, 50 μM EDTA, pH 7.5) by gentle shaking until solution was clear and centrifuged

at 31,000 x g, 15 min at room temperature. Supernatant was dialyzed overnight at 4 °C

against reconstitution buffer (50 mM tris, 50 μM EDTA, pH 7,5) (6−8 kDa molecular mass

cutoff membranes from Spectrapor). ApoTrHbN solution was centrifuged at 31,000 x g for

10 min at 4 °C before reconstitution.

Reconstitution of holo-TrHbN. Reconstitution was performed as described in

reference (2). Protein and hemin were stirred gently for 30 min in the dark for complete

reconstitution of holo-TrHbN. Hemin excess was removed by fractionation with

ammonium sulfate (40-80%). The precipitate was centrifuged at 10,000 x g for 20 min at 4

°C and resuspended in small amount of 50 mM tris, 50 μM EDTA and 300 mM NaCl

buffer and dialyzed over night against the same buffer at 4 °C. The protein mixture was

loaded onto a Hiload 26/60 Superdex 75 gel-filtration column (GE Healthcare) equilibrated

with the same buffer and monitored with AKTA FPLC (GE Healthcare) at 4 °C. Fraction

with a 412 nm/218 nm ratio between 5 and 7 were pooled and concentrated by ammonium

sulfate precipitation (75% saturation). After centrifugation (10,000 x g, 20 min, 4 °C) the

precipitate was dissolved in 20mM KH2PO4, 50 μM EDTA, pH 7.5 and dialyzed overnight

against same buffer. Purified protein was concentrated with centricon Ultracel YM-

10 (Amicon) and heme concentration was determined by pyridine-hemochrome method (3).

15N NMR relaxation experiments - Relaxation delays for R1 were 10.9, 21.8 (x2), 43.6,

87.2 (x2), 174.4, 348.9 (x2), 697.7 (x2), 11395.4, and 1994.5 ms. For R2, relaxation delays

were 10, 30 (x2), 50, 70 (x2), 90 (x2), 130 (x2), 170, 210, and 250 ms. For both R1 and R2

experiments, eight transients were recorded per FID with a recycle delay of 2s and these

experiments were recorded in an interleaved manner in order to avoid effect of field or

sample variation as a function of time (4). For {1H}-

15N NOE measurements at 500 and

600 MHz, spectra were recorded with and without 1H saturation, with 44 transients. A

saturation time of 4s was used and recycle delays were of 5s for experiments with and

259

without saturation (4s of saturation + 1s of blank delay, or 5s of blank delay, respectively).

For the NOE experiments recorded at 800 MHz, we had to set the saturation time to 10s in

order to obtain reliable values. Indeed, longer delays are sometimes needed at higher

magnetic field because the cross-correlation between 1H-

15N dipole and

15N chemical shift

anisotropy becomes significant for small proteins (5). R1 and R2 rates were extracted with

the program relax (6, 7) and errors were initially estimated from 500 Monte Carlo

simulations according to experiments acquired in duplicata. Errors were doubled in order to

avoid over fitting during model-free analysis. NOE values were obtained from the ratio of

the intensities of peaks from experiments recorded with and without proton saturation and

errors were derived from the noise. To verify integrity of the sample during the time needed

for acquisition of the three relaxation parameters (R1, R2, and NOE), R2 data were

recorded both at the beginning and at the end of acquisition at each magnetic field. As R2

values are very sensitive to changes in the homogeneity of the sample, we assumed that our

samples were stable when R2 values were not significantly different. This was as in Morin

and Gagné (8).

Model-free analysis - The model-free analysis (9, 10) was performed using an axially

symmetric diffusion tensor with the program MODELFREE 4.20 (A.G. Palmer III,

Columbia University, New York, NY). Validation tests were performed with our multiple

field data (11) and it appeared that the R2 values recorded at 800 MHz were slightly

inconsistent with other datasets, so we chose not to use these. Only residues with data

available at the three magnetic fields were used in this analysis, in order to avoid under- or

overfitting (8). An initial estimate of the diffusion tensor was obtained by using the inertia

and diffusion tensors package from A.G. Palmer lab (pdbinertia 1.11, R2R1_diffusion 1.11,

and Quadric 1.13), with the structure PDB 1S61 chain B. This chain was used because it

contains all the C-terminus residues, which is not the case for chain A, where the structure

does not contain residues 130-136. For this estimation, only residues located in secondary

structures were used and residues with NOE values ≤ 0.65 were removed as well as

residues with high values of R2 (unless their R1 values were low) (12). For the selection of

model-free models, data were best fit to the five models and 8 iterations of model selection

were performed in order to obtain stable diffusion tensor parameters. After 3 iterations of

global optimization of the tensor, all residues were included and a final run was performed

260

with fixed diffusion tensor parameters. A 1.02 Å 1H-

15N bond length and a -172 ppm 15N

chemical shift anisotropy were assumed.

Amide exchange experiments - NMR spectra were recorded at the following delays

following solubilization of lyophilized protein in D2O: 40; 50; 60; 70; 80; 100; 110; 120;

130; 140; 150; 164; 184; 243; 262; 282; 302; 322; 341; 371; 410; 449; 489; 528; 567; 617;

646; 686; 725; 764; 803; 843; 882; 1138; 1257; 1398; 1467; 1844; 2818; 3120; 6073;

11,385; 27,322; 37,515; and 60,478 min. At pH 8.5, 42 spectra were recorded at these

delays: 39; 48; 58; 68; 78; 88; 98; 108; 118; 128; 162; 182; 202; 221; 241; 280; 300; 320;

340; 369; 408; 448; 487; 526; 566; 605; 644; 684; 723; 762; 802; 841; 880; 1210; 1348;

1496; 2633; 6138; 6959; 8710; 12,716; and 18,773 min. At the beginning, spectra with a

low number of transients were recorded; as time passed, a higher number of transients were

recorded to ensure a higher S/N. Thus, depending on when the spectra were recorded, 2, 4,

or 8 transients were used for NMR data averaging. Amide exchange rates were extracted

from the peak intensities using the following equation:

It = I0 exp(-kex • t) + I∞,

where I is the intensity at time 0, t or infinity (offset, accounting for residual 1H in

solution), and kex is the exchange rate. Fits were performed using the program CURVEFIT

(A. G. Palmer, Columbia University, New York, NY) and errors were estimated from either

500 Monte Carlo simulations or the Jackknife approach, using the method which yielded

the highest errors in all cases. To avoid inclusion of noise into the fit of fast exchanged NH

groups, datasets were manually trimmed in order to include approximately five data points

when exchange was completed (i.e. when amplitude yielded a plateau of value ~I∞). From

the kex values at pH 7.5, ΔGHX (the free energy for the opening of the structure protecting

the NH group from exchange), Kop (the opening rate for exposure of the NH group) and SF

(the slowing factor for the NH group exchange) were calculated within Excel using a

spreadsheet from S. W. Englander (University of Pennsylvania, Philadelphia, PA), and kc

values (intrinsic rates of exchange for the free amino acids in solution) were from (13, 14).

261

Molecular dynamics simulations

Force field optimization of the cyanide-bound heme atomic charges and Fe-C-N angle

parameter - CHARMM22 lacks parameters for the cyanide-bound heme. In this work, to

simulate CN-bound TrHbN, the atomic charges of heme prosthetic group as well as the

cyanide and the Fe-C-N angle parameter were optimized following the standard

parametrization protocol for the CHARMM22 force field (15). Ab initio quantum

mechanical (QM) calculations were performed using the program Gaussian 03 (16). The

B3LYP/6-31G* level of theory was used for the initial geometry optimization and

subsequent single point calculations. This level of theory was applied successfully to the

parametrization of the CHARMM force field of iron–porphyrin systems (17, 18). The

atoms included for this procedure are those of the central iron–porphyrin ring and those of

the linking molecules CN- and imidazole. The heme side groups were omitted. Initial atom

coordinates were taken from the oxygenated TrHbN crystal structure (PDB accession code

1RTE (A chain)). The atomic charges were obtained from the optimized geometry

Mulliken charges, adjusted consistently with the electrostatic fitting procedure of the

parametrization protocol (15). The potential energy of interaction between the CN ligand

and a water molecule was calculated from the difference in QM energies of a gas-phase

system-water complex and the isolated molecules. For this calculation, the single point

calculations were performed while keeping the intramolecular geometry of each molecule

kept fixed.

The optimum Fe-C-N angle was determined at 180° while a zero force constant was found

through optimization procedure. The heme atom repulsion on cyanide N atom suffices to

favor a linear Fe-C-N angle. In wild type TrHbN, Fe-C-N angle is slightly bent, varying

from 165° to 171° (PDB accession 1RTE and 1S61). In agreement, using this set of

parameters, an equilibrium angle of 166.9° (standard error below 0.1° and standard

deviation of ±6.8°) was obtained in the MD simulations presented here.

Systems and simulation setup - Initial coordinates were taken from wild type cyanomet

TrHbN crystallographic structure (PDB 1RTE) (19). Crystallographic water and sulfate

262

ions were ignored. Hydrogen atoms were added using CHARMM’s HBUILD facility (20).

All ionisable residues were considered in their standard protonation state at pH 7.0 with

neutral histidine protons placed at the ND1 position. The crystal unit cell contained two

TrHbN molecules (A and B chains). Both chains were used individually to start two

independent MD simulations in order to increase sampling (trajectories identified hereafter

as A-TrHbN and B-TrHbN). For each chain, the missing C-terminal end (8 residues) was

built using internal coordinates definition of CHARMM. The C-terminal end was optimally

positioned by performing 3 ns Langevin dynamics using CHARMM with a 1 fs time step

and a friction coefficient FBETA of 5 ps-1 while keeping constrained all coordinates from

the crystal structure. The converged structures were immersed in a cubic box containing

preequilibrated TIP3P waters. This water box had a volume of ~ 4.4×105 Å3 (box edge of

76 Å) and contained 15,137 water molecules. Five sodium ions were added to neutralize

the charges of the system. Water molecules within 2.8 Å of any protein atom were deleted,

the resulting system contained 44,672 and 44,648 atoms for A-TrHbN and B-TrHbN,

respectively. The energy of the systems was minimized with 5,000 conjugate gradient steps

while keeping fixed all protein atoms at their crystallographic position. Two independent

85 ns trajectories of these systems were produced. In the equilibrium phase, all protein

atoms were fixed at their crystallographic position during the first 500 ps. In the following

500 ps, only the protein backbone was constrained. After the first nanosecond (ns), all

atoms were unconstrained. The first 5 ns were not considered for analyses giving a total of

160 ns simulation time in production mode.

MD simulations were performed using NAMD 2.6 software (21)¸with the CHARMM22

all-atom potential energy parameter set (15) with φ, ψ cross term map correction (CMAP)

(22) and modified TIP3P waters (23). Electrostatic interactions were calculated via the

Particle Mesh Ewald method (24) using a sixth-order spline interpolation for

complementary function and a grid spacing of 1.0 Å. Water molecules were kept rigid

using the SETTLE algorithm (25). For non-water molecules, covalent bonds involving

hydrogen atoms were kept at their equilibrium length using the SHAKE algorithm (26).

The Lennard-Jones interactions were smoothed over the distance range of 8 Å to 10 Å.

Long range electrostatics were computed every two steps. The non-bonded pair list was

updated every 10 steps. Simulations were performed at constant pressure and temperature

263

(NPT ensemble) at 298.15 K and 1 ATM and an integration timestep of 1 fs. Temperature

was controlled using a Lanvegin dynamics and the pressure using the Langevin piston

method. The damping coefficient was set to 1 ps-1, the Langevin piston period was set 100

fs and the Langevin piston decay to 50 fs. Coordinates were saved every ps for analysis.

264

Table S1. Optimized CN-bound heme atomic charges

CHARMM atom type Optimized charges

(E)

FE +1.29

NPH -0.72

CPA +0.36

CPB -0.04

CPM -0.23

C1 -0.10

N1 -0.51

265

Table S2: 15

N spin relaxation data

Residues 500 MHz 600 MHz 800 MHz

# AA R1 (s-1) δR1 (-1) R2 (-1) δR2 (-1) NOE δNOE R1 (s

-1) δR1 (-1) R2 (-1) δR2 (-1) NOE δNOE R1 (s-1) δR1 (-1) NOE δNOE

1 M

2 G

3 L

4 L

5 S

6 R

7 L 1.962 0.095 5.719 0.188 -0.122 0.037 1.474 0.043 5.632 0.158 0.022 0.028 1.522 0.148 0.315 0.023

8 R

9 K

10 R

11 E 1.866 0.053 7.082 0.121 0.182 0.028 1.387 0.025 6.753 0.102 0.244 0.020 1.188 0.065 0.365 0.016

12 P

13 I 1.648 0.023 7.708 0.062 0.318 0.019 1.308 0.016 8.180 0.070 0.350 0.016 1.086 0.034 0.471 0.012

14 S 1.689 0.035 11.388 0.138 0.530 0.028 1.346 0.024 12.576 0.190 0.540 0.021 1.025 0.057 0.619 0.020

15 I 1.844 0.046 12.462 0.183 0.797 0.031 1.370 0.029 13.221 0.256 0.810 0.025 1.005 0.063 0.811 0.021

16 Y 1.767 0.040 12.654 0.173 0.817 0.030 1.286 0.025 13.008 0.193 0.793 0.022 0.972 0.052 0.813 0.019

17 D 1.723 0.030 12.348 0.132 0.805 0.021 1.343 0.020 13.684 0.179 0.748 0.019 0.975 0.039 0.844 0.016

18 K

19 I 1.672 0.032 12.332 0.143 0.801 0.026 1.251 0.021 12.924 0.160 0.696 0.020 0.928 0.041 0.850 0.017

20 G 1.680 0.028 12.012 0.112 0.787 0.020 1.249 0.014 13.395 0.139 0.757 0.016 0.964 0.037 0.809 0.014

21 G 1.705 0.025 11.416 0.102 0.727 0.018 1.342 0.016 12.351 0.122 0.784 0.016 1.033 0.036 0.836 0.015

22 H 1.741 0.032 13.476 0.149 0.750 0.022 1.324 0.020 13.882 0.186 0.748 0.019 0.962 0.045 0.870 0.018

23 E 1.537 0.022 14.057 0.113 0.687 0.017 1.163 0.012 15.397 0.177 0.789 0.014 0.941 0.034 0.771 0.013

24 A 1.540 0.015 12.369 0.072 0.757 0.014 1.203 0.010 13.553 0.094 0.744 0.012 0.886 0.021 0.789 0.010

25 I 1.571 0.023 13.254 0.132 0.747 0.022 1.178 0.016 14.674 0.188 0.755 0.018 0.818 0.036 0.792 0.016

26 E 1.499 0.025 14.563 0.154 0.716 0.024 1.194 0.019 16.169 0.234 0.778 0.021 0.818 0.051 0.811 0.020

27 V

28 V 1.534 0.020 12.736 0.119 0.772 0.020 1.199 0.015 14.108 0.154 0.764 0.018 0.890 0.033 0.823 0.015

29 V

30 E 1.468 0.019 15.043 0.127 0.705 0.018 1.095 0.012 16.205 0.162 0.815 0.015 0.822 0.028 0.904 0.014

31 D 1.485 0.021 14.534 0.133 0.756 0.020 1.137 0.014 15.876 0.183 0.807 0.016 0.862 0.029 0.832 0.013

32 F 1.542 0.023 13.628 0.124 0.785 0.018 1.226 0.015 15.113 0.175 0.786 0.017 0.856 0.027 0.841 0.013

33 Y 1.560 0.026 14.136 0.137 0.730 0.021 1.178 0.015 15.742 0.189 0.777 0.017 0.859 0.033 0.865 0.015

34 V 1.433 0.020 14.412 0.128 0.789 0.019 1.078 0.013 15.444 0.140 0.811 0.015 0.813 0.028 0.809 0.012

35 R 1.524 0.027 14.009 0.161 0.788 0.023 1.135 0.018 15.321 0.219 0.787 0.018 0.885 0.042 0.817 0.016

(Table continued on next page)

Table S2: 15


266




36 V 1.523 0.024 13.910 0.152 0.759 0.021 1.155 0.015 14.581 0.179 0.777 0.017 0.824 0.035 0.806 0.015

37 L 1.521 0.023 14.830 0.150 0.775 0.021 1.172 0.016 16.246 0.188 0.755 0.016 0.823 0.035 0.867 0.016

38 A 1.494 0.021 14.255 0.119 0.767 0.017 1.128 0.012 15.632 0.165 0.793 0.014 0.834 0.028 0.880 0.013

39 D 1.682 0.023 11.602 0.085 0.783 0.016 1.258 0.013 12.348 0.103 0.793 0.013 0.940 0.026 0.841 0.011

40 D 1.792 0.027 12.014 0.102 0.680 0.017 1.295 0.015 12.790 0.121 0.792 0.013 1.035 0.037 0.793 0.012

41 Q 1.545 0.021 13.671 0.099 0.721 0.016 1.147 0.012 14.468 0.128 0.794 0.012 0.862 0.027 0.803 0.011

42 L 1.639 0.028 12.181 0.144 0.695 0.021 1.261 0.018 12.953 0.151 0.763 0.018 0.907 0.037 0.796 0.014

43 S 1.746 0.022 11.757 0.078 0.764 0.016 1.230 0.014 12.436 0.102 0.735 0.012 0.974 0.027 0.816 0.011

44 A 1.943 0.027 12.115 0.092 0.758 0.015 1.442 0.014 12.894 0.097 0.753 0.012 1.095 0.035 0.802 0.011

45 F 1.718 0.022 11.935 0.095 0.797 0.017 1.310 0.013 12.742 0.097 0.809 0.013 0.957 0.029 0.866 0.012

46 F 1.753 0.033 11.953 0.123 0.775 0.022 1.288 0.019 12.607 0.154 0.757 0.018 0.942 0.037 0.848 0.016

47 S 1.669 0.017 12.302 0.064 0.775 0.012 1.253 0.009 13.193 0.071 0.771 0.009 0.932 0.018 0.825 0.008

48 G

49 T 1.759 0.024 11.419 0.086 0.760 0.016 1.284 0.012 12.009 0.089 0.784 0.012 0.972 0.025 0.810 0.010

50 N

51 M 1.997 0.102 13.653 0.390 0.791 0.042 1.298 0.040 14.625 0.503 0.785 0.030 0.934 0.101 0.822 0.026

52 S 1.758 0.085 13.490 0.327 0.673 0.035 1.231 0.033 15.050 0.389 0.791 0.027 0.883 0.082 0.817 0.024

53 R 1.620 0.027 13.988 0.139 0.789 0.021 1.166 0.014 14.387 0.151 0.768 0.016 0.904 0.035 0.792 0.014

54 L

55 K 1.659 0.031 12.742 0.139 0.720 0.023 1.271 0.018 13.776 0.169 0.749 0.018 0.911 0.038 0.801 0.016

56 G 1.549 0.025 13.217 0.135 0.779 0.022 1.182 0.016 14.638 0.169 0.775 0.018 0.846 0.033 0.832 0.015

57 K

58 Q 1.673 0.025 12.301 0.121 0.797 0.021 1.338 0.018 13.079 0.149 0.782 0.017 0.977 0.043 0.876 0.016

59 V 1.581 0.025 12.686 0.136 0.757 0.022 1.229 0.016 13.848 0.159 0.799 0.017 0.910 0.037 0.780 0.015

60 E 1.610 0.022 13.395 0.123 0.741 0.019 1.151 0.013 14.581 0.137 0.795 0.015 0.895 0.029 0.824 0.013

61 F 1.657 0.024 13.129 0.126 0.717 0.021 1.267 0.015 14.161 0.163 0.776 0.015 0.913 0.034 0.881 0.015

62 F 1.720 0.028 12.763 0.134 0.720 0.020 1.314 0.017 13.881 0.176 0.763 0.016 0.977 0.037 0.869 0.016

63 A 1.651 0.019 13.244 0.108 0.754 0.016 1.243 0.012 14.303 0.126 0.774 0.013 0.920 0.030 0.800 0.013

64 A 1.553 0.018 13.512 0.103 0.723 0.017 1.193 0.011 14.608 0.120 0.729 0.013 0.885 0.026 0.831 0.013

65 A

66 L 1.666 0.026 11.819 0.114 0.783 0.023 1.263 0.019 12.536 0.151 0.791 0.019 0.930 0.035 0.815 0.015

67 G 1.520 0.025 14.286 0.142 0.796 0.022 1.142 0.017 15.914 0.192 0.777 0.018 0.861 0.036 0.822 0.016

68 G 1.496 0.027 14.130 0.152 0.720 0.023 1.188 0.017 14.973 0.187 0.785 0.018 0.855 0.037 0.832 0.015

69 P

70 E 1.688 0.025 10.935 0.088 0.706 0.020 1.323 0.017 10.992 0.122 0.747 0.018 0.959 0.038 0.765 0.014


Table S2: 15


267




71 P

72 Y 1.529 0.023 10.609 0.083 0.687 0.019 1.179 0.014 11.343 0.105 0.684 0.014 0.893 0.028 0.725 0.012

73 T 1.862 0.053 12.031 0.198 0.686 0.029 1.343 0.027 12.621 0.232 0.795 0.023 1.071 0.080 0.750 0.020

74 G

75 A 1.747 0.026 11.244 0.105 0.727 0.020 1.306 0.017 12.020 0.127 0.710 0.016 1.034 0.046 0.759 0.015

76 P

77 M 1.616 0.045 14.465 0.241 0.795 0.034 1.306 0.027 15.628 0.337 0.801 0.026 0.980 0.064 0.865 0.022

78 K 1.625 0.044 13.845 0.217 0.808 0.027 1.286 0.028 14.130 0.259 0.765 0.024 0.901 0.066 0.790 0.022

79 Q 1.391 0.016 11.652 0.085 0.590 0.016 1.083 0.011 12.467 0.102 0.632 0.013 0.857 0.024 0.669 0.012

80 V 1.474 0.022 13.925 0.136 0.597 0.020 1.159 0.015 15.360 0.168 0.712 0.016 0.858 0.035 0.773 0.017

81 H 1.670 0.042 12.280 0.217 0.618 0.031 1.286 0.033 12.946 0.333 0.755 0.033 0.993 0.121 0.756 0.040

82 Q 1.559 0.017 12.073 0.084 0.726 0.016 1.164 0.011 12.966 0.106 0.754 0.013 0.904 0.026 0.770 0.012

83 G 1.506 0.025 12.933 0.122 0.723 0.021 1.131 0.014 13.350 0.151 0.721 0.017 0.881 0.031 0.776 0.015

84 R 1.928 0.048 11.901 0.160 0.778 0.025 1.424 0.023 12.767 0.188 0.776 0.019 1.063 0.055 0.858 0.018

85 G 1.717 0.026 11.232 0.105 0.778 0.019 1.342 0.017 11.866 0.125 0.810 0.017 0.977 0.035 0.830 0.013

86 I 1.670 0.039 11.584 0.168 0.762 0.029 1.262 0.026 12.166 0.201 0.761 0.026 0.974 0.064 0.851 0.023

87 T 1.547 0.024 13.303 0.111 0.786 0.018 1.129 0.013 14.167 0.123 0.750 0.013 0.862 0.026 0.828 0.011

88 M

89 H

90 H 1.595 0.028 13.901 0.141 0.745 0.021 1.141 0.017 14.949 0.161 0.806 0.017 0.868 0.036 0.822 0.015

91 F 1.506 0.018 14.115 0.105 0.767 0.016 1.104 0.012 15.139 0.119 0.786 0.013 0.824 0.023 0.851 0.011

92 S 1.569 0.025 13.958 0.142 0.760 0.019 1.100 0.016 14.837 0.157 0.803 0.016 0.853 0.036 0.884 0.014

93 L 1.565 0.026 13.040 0.131 0.744 0.019 1.210 0.016 14.413 0.173 0.830 0.016 0.784 0.033 0.822 0.014

94 V 1.544 0.019 14.040 0.113 0.746 0.017 1.169 0.013 15.170 0.124 0.782 0.014 0.863 0.024 0.785 0.012

95 A 1.484 0.021 14.327 0.144 0.779 0.020 1.161 0.013 16.343 0.180 0.821 0.015 0.837 0.028 0.856 0.014

96 G 1.534 0.019 13.473 0.110 0.820 0.017 1.107 0.012 14.868 0.130 0.758 0.013 0.862 0.025 0.843 0.011

97 H 1.611 0.023 12.938 0.108 0.714 0.019 1.202 0.015 14.659 0.161 0.784 0.017 0.869 0.033 0.844 0.014

98 L 1.493 0.016 13.836 0.091 0.717 0.014 1.096 0.010 14.731 0.099 0.754 0.011 0.854 0.020 0.820 0.010

99 A 1.489 0.019 14.801 0.155 0.738 0.017 1.164 0.011 15.711 0.140 0.774 0.012 0.841 0.027 0.880 0.013

100 D 1.612 0.020 13.710 0.106 0.763 0.017 1.224 0.012 15.016 0.145 0.779 0.013 0.923 0.029 0.831 0.012

101 A 1.495 0.025 13.898 0.143 0.756 0.023 1.156 0.017 14.912 0.193 0.759 0.018 0.850 0.037 0.820 0.016

102 L 1.521 0.026 14.547 0.141 0.713 0.021 1.110 0.013 15.717 0.176 0.774 0.015 0.836 0.034 0.831 0.015

103 T 1.573 0.019 14.442 0.112 0.821 0.015 1.165 0.011 15.685 0.129 0.778 0.012 0.896 0.025 0.872 0.010

104 A 1.542 0.017 12.616 0.080 0.739 0.015 1.172 0.010 13.852 0.102 0.787 0.011 0.867 0.022 0.838 0.010

105 A 1.477 0.019 13.341 0.113 0.784 0.018 1.106 0.014 15.203 0.157 0.767 0.015 0.808 0.031 0.769 0.013


Table S2: 15


268




106 G 1.692 0.033 12.584 0.165 0.793 0.024 1.237 0.021 13.656 0.194 0.733 0.019 0.911 0.050 0.779 0.017

107 V

108 P

109 S

110 E

111 T

112 I 1.591 0.018 14.721 0.095 0.808 0.015 1.178 0.011 14.539 0.103 0.797 0.012 0.907 0.025 0.795 0.010

113 T

114 E 1.559 0.017 12.792 0.088 0.800 0.015 1.249 0.012 13.691 0.111 0.784 0.012 0.883 0.023 0.875 0.010

115 I 1.637 0.022 12.181 0.116 0.769 0.019 1.333 0.017 12.705 0.138 0.768 0.016 0.932 0.036 0.787 0.014

116 L 1.680 0.027 13.555 0.127 0.782 0.022 1.282 0.015 14.762 0.154 0.798 0.016 0.926 0.033 0.835 0.016

117 G 1.494 0.019 13.493 0.107 0.815 0.017 1.131 0.013 14.863 0.134 0.803 0.014 0.852 0.028 0.843 0.012

118 V 1.578 0.020 11.835 0.103 0.786 0.018 1.303 0.014 13.147 0.114 0.772 0.014 0.936 0.027 0.850 0.012

119 I 1.596 0.029 12.881 0.163 0.732 0.024 1.248 0.020 14.101 0.188 0.747 0.019 0.891 0.043 0.849 0.018

120 A 1.499 0.019 13.287 0.108 0.741 0.017 1.201 0.013 13.844 0.124 0.751 0.014 0.819 0.030 0.857 0.013

121 P

122 L 1.515 0.019 12.488 0.116 0.750 0.018 1.231 0.014 13.716 0.127 0.810 0.015 0.863 0.027 0.822 0.013

123 A 1.561 0.015 13.208 0.084 0.743 0.014 1.221 0.011 14.370 0.104 0.797 0.012 0.852 0.025 0.902 0.012

124 V 1.417 0.016 13.605 0.107 0.729 0.016 1.055 0.010 14.992 0.123 0.678 0.013 0.825 0.024 0.757 0.012

125 D 1.422 0.017 11.964 0.095 0.676 0.016 1.074 0.011 12.967 0.108 0.659 0.014 0.836 0.024 0.739 0.012

126 V 1.558 0.030 13.161 0.156 0.733 0.026 1.217 0.019 14.656 0.199 0.725 0.021 0.877 0.043 0.819 0.021

127 T 1.548 0.028 12.408 0.142 0.710 0.022 1.178 0.017 14.072 0.172 0.743 0.017 0.862 0.034 0.802 0.015

128 S

129 G 2.050 0.126 7.635 0.271 0.433 0.042 1.513 0.047 7.633 0.240 0.548 0.031 1.243 0.126 0.605 0.026

130 E 1.946 0.042 6.535 0.072 0.236 0.018 1.563 0.017 6.146 0.069 0.325 0.013 1.366 0.048 0.480 0.012

131 S

132 T

133 T 1.532 0.096 3.631 0.173 -0.823 0.049 1.201 0.039 3.375 0.145 -0.374 0.028 1.258 0.121 0.010 0.022

134 A 1.337 0.015 3.115 0.024 -1.099 0.010 1.166 0.006 2.789 0.019 -0.706 0.007 1.251 0.022 -0.200 0.006

135 P

136 V 0.825 0.002 1.656 0.007 -1.783 0.004 0.753 0.002 1.117 0.004 -1.536 0.004 0.800 0.005 -0.658 0.002

Table S3: Model-free analysis results

Residue Model S2 δS2 S2f δS2

f S2s δS2

f τe δτe Rex 600MHz δRex 600MHz

1 MET

269

2 GLY

3 LEU

4 LEU

5 SER

6 ARG

7 LEU 5 0.324 0.021 0.862 0.035 0.377 0.029 887.501 41.513

8 ARG

9 LYS

10 ARG

11 GLU 5 0.439 0.014 0.829 0.020 0.530 0.020 901.448 40.467

12 PRO

13 ILE 5 0.492 0.010 0.812 0.014 0.606 0.011 987.453 35.949

14 SER 2 0.857 0.016 109.277 33.890

15 ILE 1 0.923 0.022

16 TYR 1 0.917 0.020

17 ASP 1 0.928 0.018

18 LYS

19 ILE 2 0.891 0.018 17.020 9.685

20 GLY 1 0.889 0.015

21 GLY 2 0.873 0.014 17.378 7.190

22 HIS 1 0.928 0.018

23 GLU 2 0.913 0.016 42.615 12.639

24 ALA 5 0.820 0.013 0.881 0.012 0.931 0.008 1279.948 216.280

25 ILE 2 0.902 0.017 27.594 9.830

26 GLU 1 0.947 0.018

27 VAL

28 VAL 1 0.881 0.015

29 VAL

30 GLU 3 0.876 0.019 2.029 0.363

31 ASP 1 0.927 0.016

32 PHE 1 0.927 0.017

33 TYR 3 0.894 0.022 1.450 0.416

34 VAL 2 0.900 0.015 13.505 6.691

35 ARG 1 0.917 0.019




f S2s δS2

f τe δτe Rex

600MHz

δRex

600MHz

36 VAL 1 0.909 0.018

37 LEU 1 0.952 0.017

270

38 ALA 1 0.918 0.016

39 ASP 1 0.853 0.014

40 ASP 1 0.897 0.016

41 GLN 2 0.900 0.016 21.828 7.240

42 LEU 2 0.884 0.018 26.409 8.412

43 SER 1 0.872 0.014

44 ALA 5 0.888 0.016 0.957 0.016 0.928 0.011 1460.318 357.782

45 PHE 1 0.887 0.015

46 PHE 1 0.889 0.017

47 SER 2 0.889 0.013 17.176 4.434

48 GLY

49 THR 1 0.858 0.014

50 ASN

51 MET 1 0.963 0.030

52 SER 1 0.940 0.028

53 ARG 2 0.904 0.016 28.335 9.000

54 LEU

55 LYS 1 0.913 0.018

56 GLY 1 0.905 0.017

57 LYS

58 GLN 1 0.896 0.017

59 VAL 1 0.899 0.016

60 GLU 2 0.904 0.016 17.387 7.860

61 PHE 1 0.921 0.016

62 PHE 1 0.931 0.017

63 ALA 1 0.927 0.016

64 ALA 2 0.903 0.016 29.534 7.735

65 ALA

66 LEU 1 0.872 0.016

67 GLY 1 0.933 0.017

68 GLY 1 0.912 0.016

69 PRO

70 GLU 5 0.781 0.016 0.862 0.015 0.906 0.014 1266.666 254.080




f S2s δS2

f τe δτe Rex

600MHz

δRex

600MHz

71 PRO

72 TYR 2 0.788 0.014 26.339 3.448

73 THR 1 0.907 0.022

271

74 GLY

75 ALA 5 0.809 0.017 0.891 0.016 0.908 0.015 1146.893 220.982

76 PRO

77 MET 1 0.974 0.021

78 LYS 1 0.932 0.022

79 GLN 5 0.747 0.014 0.832 0.014 0.897 0.010 665.041 83.421

80 VAL 5 0.889 0.018 0.940 0.022 0.947 0.015 415.945 205.281

81 HIS 2 0.875 0.023 60.397 55.779

82 GLN 2 0.842 0.013 23.639 4.417

83 GLY 4 0.779 0.019 16.257 3.983 2.749 0.338

84 ARG 5 0.742 0.032 0.913 0.019 0.813 0.034 5535.414 2044.161

85 GLY 1 0.858 0.014

86 ILE 1 0.859 0.021

87 THR 4 0.789 0.019 8.004 3.032 3.234 0.285

88 MET

89 HIS

90 HIS 3 0.857 0.021 2.143 0.400

91 PHE 3 0.821 0.018 3.028 0.289

92 SER 3 0.865 0.022 1.408 0.418

93 LEU 1 0.895 0.016

94 VAL 4 0.857 0.018 19.191 4.864 2.172 0.291

95 ALA 3 0.860 0.019 2.753 0.360

96 GLY 2 0.889 0.015 11.061 5.553

97 HIS 2 0.898 0.015 17.563 8.429

98 LEU 4 0.817 0.016 12.400 2.901 2.514 0.239

99 ALA 3 0.877 0.017 2.035 0.327

100 ASP 2 0.922 0.015 27.115 9.998

101 ALA 3 0.836 0.021 2.645 0.431

102 LEU 3 0.837 0.019 3.093 0.392

103 THR 1 0.940 0.015

104 ALA 2 0.872 0.014 12.134 4.239

105 ALA 4 0.779 0.017 12.727 3.157 3.722 0.320




f S2s δS2

f τe δτe Rex

600MHz

δRex

600MHz

106 GLY 5 0.828 0.021 0.904 0.020 0.916 0.016 1619.513 660.151

107 VAL

272

108 PRO

109 SER

110 GLU

111 THR

112 ILE 4 0.855 0.017 12.916 4.120 2.624 0.255

113 THR

114 GLU 1 0.893 0.015

115 ILE 2 0.882 0.015 26.375 8.153

116 LEU 1 0.952 0.017

117 GLY 1 0.899 0.015

118 VAL 1 0.880 0.016

119 ILE 1 0.912 0.019

120 ALA 1 0.882 0.015

121 PRO

122 LEU 1 0.886 0.015

123 ALA 1 0.907 0.015

124 VAL 2 0.862 0.014 31.918 5.426

125 ASP 5 0.780 0.014 0.838 0.014 0.931 0.010 635.625 134.955

126 VAL 1 0.914 0.018

127 THR 2 0.874 0.017 23.347 7.201

128 SER

129 GLY 5 0.434 0.030 0.813 0.034 0.534 0.036 1537.863 149.921

130 GLU 5 0.323 0.009 0.825 0.015 0.391 0.011 1255.365 36.660

131 SER

132 THR

133 THR 5 0.166 0.018 0.749 0.034 0.222 0.026 702.565 25.944

134 ALA 5 0.132 0.003 0.752 0.008 0.176 0.004 601.909 8.469

135 PRO

136 VAL 5 0.030 0.001 0.545 0.003 0.055 0.001 470.089 6.849

273

Table S4. MD-derived dynamics parameter

Residue

M1a M2

b

A-TrHbN B-TrHbN Global

S2 Conv.

c S

2 Conv. Conv.

d S

2

2 G - - - - - -

3 L 0.665 yes 0.592 yes yes 0.642

4 L 0.761 yes 0.678 yes yes 0.733

5 S 0.742 yes 0.684 yes yes 0.732

6 R 0.721 yes 0.643 yes yes 0.691

7 L 0.760 yes 0.674 yes yes 0.725

8 R 0.775 yes 0.714 yes yes 0.762

9 K 0.694 yes 0.632 yes yes 0.681

10 R 0.658 yes 0.558 yes yes 0.596

11 E 0.713 yes 0.639 yes yes 0.665

12 P - - - - - -

13 I 0.632 yes 0.459 yes yes 0.539

14 S 0.703 yes 0.635 yes yes 0.683

15 I 0.888 yes 0.888 yes yes 0.890

16 Y 0.903 yes 0.909 yes yes 0.907

17 D 0.898 yes 0.900 yes yes 0.901

18 K 0.897 yes 0.896 yes yes 0.898

19 I 0.891 yes 0.890 yes yes 0.891

20 G 0.833 yes 0.833 yes yes 0.836

21 G 0.853 yes 0.845 yes yes 0.849

22 H 0.870 yes 0.853 yes yes 0.861

23 E 0.877 yes 0.865 yes yes 0.872

24 A 0.873 yes 0.865 yes yes 0.869

25 I 0.903 yes 0.896 yes yes 0.899

26 E 0.915 yes 0.914 yes yes 0.915

27 V 0.908 yes 0.905 yes yes 0.907

28 V 0.912 yes 0.912 yes yes 0.913

29 V 0.935 yes 0.935 yes yes 0.935

30 E 0.922 yes 0.922 yes yes 0.922

31 D 0.920 yes 0.920 yes yes 0.920

32 F 0.931 yes 0.931 yes yes 0.932

33 Y 0.924 yes 0.923 yes yes 0.924

34 V 0.889 yes 0.882 yes yes 0.885

35 R 0.917 yes 0.915 yes yes 0.916

36 V 0.927 yes 0.925 yes yes 0.925

37 L 0.919 yes 0.914 yes yes 0.916

38 A 0.853 yes 0.855 yes yes 0.854


274


Residue

M1a M2

b


S2 Conv.

c S

2 Conv. Conv.

d S

2

39 D 0.709 yes 0.723 yes yes 0.716

40 D 0.821 yes 0.831 yes yes 0.824

41 Q 0.808 yes 0.838 yes yes 0.822

42 L 0.863 yes 0.886 yes yes 0.872

43 S 0.900 yes 0.901 yes yes 0.901

44 A 0.892 yes 0.885 yes yes 0.890

45 F 0.882 yes 0.882 yes yes 0.883

46 F 0.882 yes 0.848 no yes 0.869

47 S 0.756 no 0.378 no no 0.586

48 G 0.628 no 0.063 no no 0.244

49 T 0.597 no 0.339 no no 0.418

50 N 0.802 yes 0.770 no yes 0.793

51 M 0.875 yes 0.868 yes yes 0.874

52 S 0.893 yes 0.895 yes yes 0.893

53 R 0.906 yes 0.908 yes yes 0.906

54 L 0.915 yes 0.916 yes yes 0.915

55 K 0.924 yes 0.929 yes yes 0.928

56 G 0.888 yes 0.894 yes yes 0.892

57 K 0.904 yes 0.911 yes yes 0.908

58 Q 0.925 yes 0.929 yes yes 0.928

59 V 0.927 yes 0.932 yes yes 0.930

60 E 0.923 yes 0.924 yes yes 0.923

61 F 0.931 yes 0.932 yes yes 0.932

62 F 0.933 yes 0.931 yes yes 0.933

63 A 0.924 yes 0.924 yes yes 0.925

64 A 0.917 yes 0.921 yes yes 0.919

65 A 0.918 yes 0.921 yes yes 0.920

66 L 0.907 yes 0.907 yes yes 0.908

67 G 0.881 yes 0.879 yes yes 0.881

68 G 0.861 yes 0.863 yes yes 0.863

69 P - - - - - -

70 E 0.635 no 0.731 yes yes 0.678

71 P - - - - - -

72 Y 0.733 no 0.771 no no 0.490

73 T 0.743 no 0.715 no no 0.721

74 G 0.768 no 0.739 no no 0.762

75 A 0.779 no 0.751 no no 0.779


275


Residue

M1a M2

b


S2 Conv.

c S

2 Conv. Conv.

d S

2

79 Q 0.872 yes 0.864 yes yes 0.867

80 V 0.885 yes 0.889 yes yes 0.887

81 H 0.897 yes 0.829 no yes 0.874

82 Q 0.861 yes 0.882 yes yes 0.869

83 G 0.821 no 0.826 no no 0.807

84 R 0.748 no 0.801 no no 0.754

85 G 0.802 no 0.694 no no 0.751

86 I 0.863 yes 0.850 yes yes 0.862

87 T 0.812 yes 0.785 yes yes 0.793

88 M 0.907 yes 0.909 yes yes 0.907

89 H 0.912 yes 0.909 yes yes 0.911

90 H 0.912 yes 0.915 yes yes 0.913

91 F 0.916 yes 0.918 yes yes 0.917

92 S 0.927 yes 0.926 yes yes 0.927

93 L 0.926 yes 0.926 yes yes 0.925

94 V 0.930 yes 0.932 yes yes 0.931

95 A 0.937 yes 0.938 yes yes 0.938

96 G 0.918 yes 0.917 yes yes 0.918

97 H 0.917 yes 0.919 yes yes 0.918

98 L 0.929 yes 0.929 yes yes 0.929

99 A 0.935 yes 0.933 yes yes 0.934

100 D 0.933 yes 0.931 yes yes 0.932

101 A 0.931 yes 0.928 yes yes 0.930

102 L 0.935 yes 0.932 yes yes 0.934

103 T 0.925 yes 0.922 yes yes 0.923

104 A 0.927 yes 0.923 yes yes 0.925

105 A 0.868 yes 0.866 yes yes 0.867

106 G 0.806 yes 0.801 yes yes 0.804

107 V 0.876 yes 0.874 yes yes 0.875

108 P - - - - - -

109 S 0.892 yes 0.883 yes yes 0.888

110 E 0.903 yes 0.894 yes yes 0.899

111 T 0.892 yes 0.885 yes yes 0.889

112 I 0.921 yes 0.915 yes yes 0.919

113 T 0.921 yes 0.913 yes yes 0.918

114 E 0.918 yes 0.912 yes yes 0.915

115 I 0.926 yes 0.913 yes yes 0.919


276


Residue

M1a M2

b


S2 Conv. c S2 Conv. Conv.d S2

119 I 0.872 no 0.842 no no 0.861

120 A 0.869 no 0.851 no no 0.873

121 P - - - - - -

122 L 0.848 yes 0.869 yes yes 0.857

123 A 0.888 yes 0.887 yes yes 0.889

124 V 0.884 yes 0.872 yes yes 0.875

125 D 0.858 yes 0.861 yes yes 0.851

126 V 0.916 yes 0.916 yes yes 0.916

127 T 0.902 yes 0.874 no yes 0.891

128 S 0.858 yes 0.604 no yes 0.193

129 G 0.353 no 0.197 no no 0.109

130 E 0.067 no 0.450 no no 0.151

131 S 0.190 no 0.121 no no 0.235

132 T 0.217 no 0.007 no no 0.146

133 T 0.157 no 0.128 no no 0.143

134 A 0.205 no 0.074 no no 0.177

135 P - - - - - -

136 V 0.105 no 0.083 no no 0.173

119 I 0.872 no 0.842 no no 0.861

120 A 0.869 no 0.851 no no 0.873

121 P - - - - - -

122 L 0.848 yes 0.869 yes yes 0.857

123 A 0.888 yes 0.887 yes yes 0.889

124 V 0.884 yes 0.872 yes yes 0.875

125 D 0.858 yes 0.861 yes yes 0.851

126 V 0.916 yes 0.916 yes yes 0.916

127 T 0.902 yes 0.874 no yes 0.891

128 S 0.858 yes 0.604 no yes 0.193

129 G 0.353 no 0.197 no no 0.109

130 E 0.067 no 0.450 no no 0.151

131 S 0.190 no 0.121 no no 0.235

132 T 0.217 no 0.007 no no 0.146

133 T 0.157 no 0.128 no no 0.143

134 A 0.205 no 0.074 no no 0.177

135 P - - - - - -

136 V 0.105 no 0.083 no no 0.173

a. Internal autocorrelation in invidual simulations

b. Internal autocorrelation by randomization of all simulations

c. Converged during the simulation

d. Converged in at least one simulation

277

References

1. Couture, M., S. R. Yeh, B. A. Wittenberg, J. B. Wittenberg, Y. Ouellet, D. L. Rousseau,

and M. Guertin. 1999. A cooperative oxygen-binding hemoglobin from Mycobacterium

tuberculosis. Proc Natl Acad Sci U S A 96:11223-11228.

2. Scott, N. L., and J. T. Lecomte. 2000. Cloning, expression, purification, and

preliminary characterization of a putative hemoglobin from the cyanobacterium

Synechocystis sp. PCC 6803. Protein Sci 9:587-597.

3. Berry, E. A., and B. L. Trumpower. 1987. Simultaneous determination of hemes a, b,

and c from pyridine hemochrome spectra. Anal Biochem 161:1-15.

4. Tjandra, N., P. Wingfield, S. Stahl, and A. Bax. 1996. Anisotropic rotational diffusion

of perdeuterated HIV protease from 15

N NMR relaxation measurements at two

magnetic fields. J Biomol NMR 8:273-284.

5. Gong, Q., and R. Ishima. 2007. 15

N-{1H} NOE experiment at high magnetic field

strengths. J Biomol NMR 37:147-157.

6. d'Auvergne, E. J., and P. R. Gooley. 2008. Optimisation of NMR dynamic models II. A

new methodology for the dual optimisation of the model-free parameters and the

Brownian rotational diffusion tensor. J Biomol NMR 40:121-133.

7. d'Auvergne, E. J., and P. R. Gooley. 2008. Optimisation of NMR dynamic models I.

Minimisation algorithms and their performance within the model-free and Brownian

rotational diffusion spaces. J Biomol NMR 40:107-119.

8. Morin, S., and S. M. Gagne. 2009. NMR dynamics of PSE-4 beta-lactamase: an

interplay of ps-ns order and µs-ms motions in the active site. Biophys J 96:4681-4691.

9. Lipari, G., and A. Szabo. 1982. Model-Free Approach to the Interpretation of Nuclear

Magnetic Resonance Relaxation in Macromolecules - Theory and Range of Validity. J

Am Chem Soc 104:4546-4559.

10. Lipari, G., and A. Szabo. 1982. Model-Free Approach to the Interpretation of Nuclear

Magnetic Resonance Relaxation in Macromolecules - Analysis of Experimental

Results. J Am Chem Soc 104:4559-4570.

11. Morin, S., and M. G. S. 2009. Simple tests for the validation of multiple field spin

relaxation data. J Biomol NMR 45:361-372.

12. Pawley, N. H., C. Wang, S. Koide, and L. K. Nicholson. 2001. An improved method for

distinguishing between anisotropic tumbling and chemical exchange in analysis of 15

N

relaxation parameters. J Biomol NMR 20:149-165.

278

13. Bai, Y., J. S. Milne, L. Mayne, and S. W. Englander. 1993. Primary structure effects on

peptide group hydrogen exchange. Proteins 17:75-86.

14. Connelly, G. P., Y. Bai, M. F. Jeng, and S. W. Englander. 1993. Isotope effects in

peptide group hydrogen exchange. Proteins 17:87-92.

15. MacKerell, A. D., D. Bashford, M. Bellott, R. L. Dunbrack, J. D. Evanseck, M. J. Field,

S. Fischer, J. Gao, H. Guo, S. Ha, D. Joseph-McCarthy, L. Kuchnir, K. Kuczera, F. T.

K. Lau, C. Mattos, S. Michnick, T. Ngo, D. T. Nguyen, B. Prodhom, W. E. Reiher, B.

Roux, M. Schlenkrich, J. C. Smith, R. Stote, J. Straub, M. Watanabe, J. Wiorkiewicz-

Kuczera, D. Yin, and M. Karplus. 1998. All-Atom Empirical Potential for Molecular

Modeling and Dynamics Studies of Proteins. Journal of Physical Chemistry B

102:3586-3616

16. Frisch, M. J., G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,

J. R. Cheeseman, J. A. Montgomery, T. Vreven, K. N. Kudin, J. C. Burant,

J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi,

G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota,

R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai,

M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo,

J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi,

Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador,

J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. arkas,

D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A.

G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz,

I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, A. Laham, C. Y. Peng, A. Nanayakkara,

M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez, and

J. A. Pople. 2003. Gaussian 03, Revision C.02.

17. Autenrieth, F., E. Tajkhorshid, J. Baudry, and Z. Luthey-Schulten. 2004. Classical force

field parameters for the heme prosthetic group of cytochrome c. J Comput Chem

25:1613-1622.

18. Daigle, R., M. Guertin, and P. Lague. 2009. Structural characterization of the tunnels of


simulations. Proteins : Struc. Func. Bioinf. 75:735-747.

19. Milani, M., Y. Ouellet, H. Ouellet, M. Guertin, A. Boffi, G. Antonini, A. Bocedi,

M. Mattu, M. Bolognesi, and P. Ascenzi. 2004. Cyanide binding to truncated

hemoglobins: a crystallographic and kinetic study. Biochemistry 43:5213-5221.

20. Brooks, B. R., R. E. Bruccoleri, B. D. Olafson, D. J. States, S. Swaminathan, and

M. Karplus. 1983. CHARMM: A program for macromolecular energy, minimization,

and dynamics calculations. Journal of Computational Chemistry 4:187-217.

279

21. Phillips, J. C., R. Braun, W. Wang, J. Gumbart, E. Tajkhorshid, E. Villa, C. Chipot, R.

D. Skeel, L. Kalé, and K. Schulten. 2005. Scalable molecular dynamics with NAMD. J

Comput Chem 26:1781-1802.



mechanics in reproducing protein conformational distributions in molecular dynamics

simulations. J Comput Chem 25:1400-1415.



24. Darden, T., D. York, and L. Pedersen. 1993. Particle mesh Ewald: An N·log(N) method

for Ewald sums in large systems. The Journal of Chemical Physics 98:10089-10092.

25. Miyamoto, S., and P. Kollman. 1992. Settle: An analytical version of the SHAKE and

RATTLE algorithm for rigid water models. Journal of Computational Chemistry

13:952-962.

26. Ryckaert, J.-P., G. Ciccotti, and H. J. C. Berendsen. 1977. Numerical integration of the

cartesian equations of motion of a system with constraints: molecular dynamics of

nalkanes. Journal of Computational Physics 23:327-341.

280

Annexe 4


Complete experimental procedures

Mutagenesis, expression and purification – Amino acid substitutions were carried out

using the QuickChange Site-Directed Mutagenesis kit (Stratagene) following the

recommended protocol. The cloned M. tuberculosis glbN gene was used as a template with

the complementary oligonucleotide primers, Ala24(B1)Leu: 5’GATCGGCGGGCATG

AGCTGATCGAAGTCGTCGTCG3’ with 5’CGACGACGACTTCGATCAGCTCATGC

CCGCCGATC3’; Ala65(E18)Ile: 5’GAGTTTTTCGCGGCCATACTTGGCGGGCCCG

AG3’ with 5’CTCGGGCCCGCCAAGTATGGCCGCGAAAAACTC3’; Ala95(G9)Ile:

5’CCACTTCAGCCTGGTCATCGGACACTTGGCCGACG3’ with 5’CGTCGGCCAAGT

GTCCGATGACCAGGCTGAAGTGG3’. The expression and purification of the

recombinant proteins were performed in accordance with the previously published

method (1).

NOD reaction – NOD reaction we measured by stopped-flow spectrophotometry under

single turnover conditions as previously described (2). Reaction rates were measured at 5°C

using an Applied Photophysics SX.18MV-R stopped-flow spectrophotometer (Leatherhead,

U.K.) equipped with a photodiode array detector. The integration time was 2.5 ms. 1600

Spectra were collected on time scales ranging from 1.3 to 4100 ms. Singular value

decomposition and global analysis were performed using the Specfit/32 (3.0.37) program.

Kinetics constants obtained from fitting had uncertainties of 5%. The results shown in

Figures 2, 3, S1 and S2 are representative of at least two experiments.

Flash-photolysis experiments – Laser flash-photolysis studies of the ferric •NO complexes

of the different proteins were carried out as previously described (3) using the LKS.60

281

Spectrometer from Applied Photolysis (Leatherhead, U.K.) at 23 °C. Photolysis was

initiated by a 5 ns pulse of light at 532 nm provided by a Brillant B Nd:YAG laser

(QUANTEL S.A., Fr.). Absorbance changes were measured at 392 nm using the

monochromator-filtered light from a 150 W xenon arc lamp. Passing through the sample,

the probe light beam was refocused on the slits (slits widths at 1 mm) of a second

monochromator. Changes in transmitted probe light intensity were detected by a 1P28 PMT

coupled with a HP 54830B DSO digital oscilloscope (Agilent Technologies Inc., USA) and

transferred on a RISC platform PC (Acorn, U.K.) for processing. An average of at least ten

kinetic traces from at least two separate experiments were averaged and analyzed with the

instrument manufacturer software (Applied Photolysis, U.K.) to obtain the rate constants.

The fraction of geminate rebinding was calculated as described in (4). Plots showing

absorbance changes following •NO photolysis were obtained using the KaleidaGraph

software (Synergy Software, USA).

MD simulations - Simulations were performed using CHARMM (5) and the CHARMM22

all-atom potential energy parameter set (6) with φ, ψ cross term map correction

(CMAP) (7) and modified TIP3P waters (8). Electrostatic interactions were calculated via

the Particle Mesh Ewald method (9) using a sixth-order spline interpolation for

complementary function, with κ = 0.34 Å-1

, and a fast-Fourier grid density of ≈ 1 Å.

Cutoffs for the real space portion of the Particle Mesh Ewald calculation and the truncation

of the Lennard-Jones interactions were 10 Å, with the latter smoothed via a shifting

function over the range of 8 Å to 10 Å. The SHAKE algorithm (10) was used to constrain

all covalent bonds involving hydrogen atoms. All simulations employed the leapfrog

algorithm and an integration step of 1 femtosecond (fs). Coordinates were saved every

picosecond (ps) for analysis. Nonbond and image lists were updated heuristically. All

simulations were performed at constant pressure and temperature (NPT ensemble) using

Hoover algorithm for temperature control (11). The mass of the thermal piston was 20 000

kcal •mol-1

•ps2 and the mass of the pressure piston equaled 1000 amu. All simulations were

carried out at 298 K and 1 atm. The net translation and rotation of the systems were

removed every 10 000 steps.

282

System setup - The structure and dynamics of the mutant, under the FeIIO2 form, was

studied by performing a 30 ns MD simulation. The coordinates of the mutant protein were

built from an equilibrated MD frame of wild-type oxygenated TrHbN (wt-TrHbN). The

simulation methodology employed was the same as previously used for wt-TrHbN (12).

The first 5 ns were considered as the equilibration phase giving 25 ns in production mode.

The analysis of this simulation was performed in comparisons with two 30 ns MD

simulations of wt-TrHbN presented earlier (13).

Evaluation of tunnel entrance openings – The opening of the tunnel entrances was

evaluated as follows. First, each tunnel path is mapped on a 3D cylindrical grid of 4 Å

radius and 8.2 Å length. Grids were centered on mutated residues, and tunnel paths were

taken from a previous study (39). Points on the grids were separated by 0.25A (i.e. voxels

of (0.25 Å3). Second, voxels occupied by protein atoms were found; the atomic radii used

were 1.20 Å for H, 1.70 Å for C, 1.52 Å O, 1.55 Å for N, 1.80 Å for S and Fe. Next, empty

voxels accessible by a probe of 1.4 Å radius were grouped into cavities. If a cavity

extended from the solvent to the protein core by at least 3 Å each side of the mutated

residue (center of the grid) the tunnel was considered opened. All of the 25 000 MD frames

of wt-TrHbN (from previous work) and the mutant were analyzed. Similarly, all MD

frames from LES simulations were analyzed to measure impacts of •NO molecules.

Interactions formed by the mutated side chain and other tunnel residues -

LT: Leu24(B1) makes contacts with side chains from internal residues Ile19(A15),

Ile25(B2), Val28(B5), Val29(B6), Leu102(G16) and Val107(GH5). Among these residues,

interactions with residues on positions B5, G16 and GH5 are new while contacts with

residue B6 are significantly increased.

ST: Ile95(G9) is making contacts with surface residues Phe91(G5), Leu116(H8),

Ala120(H12) and the internal residues Val94(G8), Leu98(G12) and Ile119(H11). Among

these residues, novel interactions concerns residues on positions G8, G12 and H12 while

interactions with residues G5, H8 and H11 are increased.

283

EHT: Ile65(E18) interacts with side chains from surface residues Phe61(E14),

Val118(H10), Leu122(H14) and internal residues Leu66(E19) and Ile119(H11). Among

these interactions, contacts with Ile119(H11) were not observed in wt-TrHbN while the

others are all increased.

284

Table S1. Rotameric species observed for mutated side chains

Mutation Tunnels Rotamers* Occupancy Blocking

†

Ala24(B1)Leu Long tp

tt

mm

mt

outliers

0.843

0.079

0.001

0.041

0.035

+

+

-

-

Ala65(E18)Ile EH tp

tt

mm

outliers

0.422

0.425

0.037

0.116

+

+

+

Ala95(G9)Ile Short pp

pt

tp

tt

mt

outliers

0.031

0.846

0.006

0.074

0.014

0.028

+

+

+

+

+

* Based on rotameric nomenclature of Lovell and al. (14)

† Indicate if the rotamer is blocking (+) or not (-) the tunnel entrance

285

Table S2. Side-chain conformations and dynamics observed in the mutant protein relative to wt-TrHbN

Residue Tunnel s-chain

position Contacts with other tunnel residues

*

wt-TrHbN versus triple mutant†

Dynamics Conformations Ile15(A11) LT Internal A15 - E19 - H7 Yes No

Ile19(A15) LT Internal A11 - B1 - B2 – B5 - E15 - E19 - G16 - GH5 Yes No

Leu/Ala(B1) LT Surface A15 - B2 – B5 – B6 - G16 - GH5 - -

Ile25(B2) LT Internal A15 – B1 - B5 – B6 - E15 - E19 Yes No

Val28(B5) LT Internal A15 - B1 - B2 – B6 - B9 - E15 - G12 - G16 No No

Val29(B6) LT Internal B2 – B5 - B9 - B10 - E11 - E15 No No

Phe32(B9) DHP-ST Internal B5 – B6 – B10 – E11 – E15 – G8 – G12 No No

Tyr33(B10) DHP Internal B6 – B9 – CD1 – E7 – E11 No No

Phe46(CD1) DHP Internal B10 – CD1 – E4 – E7 – G8 No No

Leu54(E7) DHP Internal B1- CD1 – E11 No No

Gln58(E11) DHP Internal B9 – B10 – E7 No No

Phe61(E14) EHT Surface E11 – E15 – E18 – H11 – H14 No No

Phe62(E15) LT-EHT Internal A15 – B2 – B5 – B6 – B9 – E11 – E14 – E18 – E19 – G12 – G16 – H7 – H11 Yes No

Ala/Ile65(E18) EHT Surface E14 – E15 – E19 – H10 – H11 - -

Leu66(E19) LT-EHT Internal A11 – A15 – B2 – E15 – E18 No No

Phe91(G5) ST Surface G9 – H11 – H12 – H14 No No

Val94(G8) DHP Internal B9 – CD1 – G8 – G9 No No

Ala/Ile95(G9) ST Surface G5 – G8 – G12 –H8 – H11 – H12 - -

Leu98(G12) EHT-LS-ST Internal B5 – B9 – E15 – G9 – G16 – H7- H8 – H11 No No

Leu102(G16) LT Internal A15 – B5 – E15 – G12 – GH5 – H7 – H8 No No

Ile115(H7) LT Internal A11 – A15 – E15 – E19 – G12 –G16 –GH5 – H10 No No

Leu116(H8) ST Surface G9 – G12 – G16 – H7 - H11 - H12 No No

Val118(H10) EHT Surface A11 – E18 – E19 – H7 – H11 Yes Yes

Ile119(H11) EHT-ST Internal E14 – E15 – E18 – E19 – G5 – G9 – G12 – H7 – H8 Yes Yes

Ala120(H12) ST Surface G5 – G9 – H8 No No

Leu122(H14) EHT Surface E14 –E18 – G5 – H10 – H11 No No

* Residues in bold are for side-chain contacts observed in the triple mutant trajectory that are not present in wt-TrHbN. Underlined

residues are those where contacts are significantly augmented. Side chain contacts were analyzed using a cutoff arbitrarily set to 3.5 Å.

286

Figure S1. Reaction of horse heart MbFeII

(O2) (5 M) with one equivalent of •NO at

5 ºC, pH 9.5. (a) Evolution of the optical spectra acquired during the first 500 ms and

collected on time scales ranging from 1.3 ms (red line) to 500 ms (blue line) with an

integration time of 2.5 ms. Abs, absorbance units. (b) First spectrum (417, 544 and 580 nm)

recorded after mixing (1.3 ms). (c) The reaction of oxidation of MbFeII(O2) by •NO was

well described using a double exponential function (ABC). The kinetics at 580 nm

(red) and the fit (black) are shown. (d) optical spectra of the species obtained by singular

value decomposition and global analysis of the rapid scan data from (a): Species A (red),

species B (black) and species C (blue). Abs: absorbance.

287

Figure S2. Reaction of the FeII

(O2) forms (5 M) of TrHbN and the LT/ST/EHT

mutant (5 M) with one equivalent of •NO at 5 ºC, pH 9.5. Optical pectrum recorded at

1.3 ms for (a) wt-TrHbN and (b) LT/ST/EHT mutant. The mutant shows a significant LS

character (409, 544 and 582 nm) relative to that of wt-TrHbN. Abs, absorbance units.

288

Figure S3. Conformational flexibility of side chains blocking (a) LT (A24L), (b) EHT

(A65I) and (c) ST (A95I). Plot showing side-chain dihedral over time (top) and (bottom)

the corresponding rotamer populations.

289

Figure S3. (continued, EHT (Ile65(E18)) Plot showing side-chain dihedral over time (top)

and (bottom) the corresponding rotamer populations.

290

Figure S3. (continued, EHT (Ile95(G9)) Plot showing side-chain dihedral over time (top)

and (bottom) the corresponding rotamer populations.

291

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

Figure S4. Side-chain flexibility of all tunnel-residues of the mutant (red) and wt-

TrHbN (blue). (left) Plot showing side-chain dihedral over time and (right) the

corresponding rotamer populations. Graphes a (top), b (middle) and c (bottom) are shown.

(Continued on next page)

292

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

Figure S4. (Continued) Graphes d (top), e (middle) and f (bottom) are shown.

293

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

Figure S4. (Continued) Graphes g (top), h (middle) and i (bottom) are shown.

294

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

Figure S4. (Continued) Graphes j (top), k (middle) and l (bottom) are shown.

295

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

Figure S4. (Continued) Graphes m (top), n (middle) and o (bottom) are shown.

296

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

Figure S4. (Continued) Graphes p (top), q (middle) and r (bottom) are shown.

297

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

-----------------------------------------------------------------------------------------------------

Figure S4. (Continued) Graphes s (top), t (middle) and u (bottom) are shown.

298

Figure S5. PMF profiles for •NO diffusion in the different tunnels for the wt-TrHbN

(blue) and mutant (red and green). Shaded gray zones correspond to tunnel region filled

by the mutations. For LT (a), profile in red was calculated for the mutant without

Ala24(B1) mm and mt rotamers. For tunnel passing by GHc (d), profile in red was

calculated mutant without Ile119(H11) rotamers tt and tp. To highlight Xe1, Xe2, Xe5

cavities along LT, PMF was calculated depending on Phe62(E15) conformations (rotamers

t80 or m30/m-85) and only rotamers m30/m-85 is shown here for picture clarity. PMF

profiles calculated with t80 rotamers is show in Figure 5 in the main text. ILS calculation

using all 25 000 MD frames from the mutant is shown by the green line. Errorbars,

depending on calculated PMF levels, are not shown for picture clarity (see method).

299

Figure S6. Exits and entry events as function of time observed in simulations of wt-

TrHbN and the mutant. wt-TrHbN : Total exit events observed (red filled circles), entry

events (blue filled squares). The exit of every single •NO (without considering •NO that

reentered), is plotted in green triangles. Mutant: Total exits events observed (open black

squares).

300

Figure S7. •NO diffusion pathways in wt-TrHbN (top) and the mutant, (bottom) from

locally enhanced sampling MD simulations. Diffusion occurring in different tunnels are

colored in purple for the LT, red for the ST, yellow for the EHT, blue for the GH and green

for the EH2. •NO molecule colored in gray are located where tunnel are merging.

301

References

1. Mukai, M., P. Y. Savard, H. Ouellet, M. Guertin, and S. R. Yeh. 2002. Unique

ligand-protein interactions in a new truncated hemoglobin from Mycobacterium

tuberculosis. Biochemistry 41:3897-3905.

2. Ouellet, H., Y. Ouellet, C. Richard, M. Labarre, B. Wittenberg, J. Wittenberg, and

M. Guertin. 2002. Truncated hemoglobin HbN protects Mycobacterium bovis from

nitric oxide. Proc Natl Acad Sci U S A 99:5902-5907.

3. Ouellet, Y. H., R. Daigle, P. Lague, D. Dantsker, M. Milani, M. Bolognesi, J. M.

Friedman, and M. Guertin. 2008. Ligand binding to truncated hemoglobin N from

Mycobacterium tuberculosis is strongly modulated by the interplay between the

distal heme pocket residues and internal water. J Biol Chem 283:27270-27278.

4. Salter, M. D., K. Nienhaus, G. U. Nienhaus, S. Dewilde, L. Moens, A. Pesce, M.

Nardini, M. Bolognesi, and J. S. Olson. 2008. The apolar channel in cerebratulus

lacteus hemoglobin is the route for O2 entry and exit. J Biol Chem. 283:35689-

35702.

5. Brooks, B. R., R. E. Bruccoleri, B. D. Olafson, D. J. States, S. Swaminathan, and

M. Karplus. 1983. CHARMM: A program for macromolecular energy,

minimization, and dynamics calculations. Journal of Computational Chemistry

4:187-217.

6. MacKerell, A. D., D. Bashford, M. Bellott, R. L. Dunbrack, J. D. Evanseck, M. J.

Field, S. Fischer, J. Gao, H. Guo, S. Ha, D. Joseph-McCarthy, L. Kuchnir, K.

Kuczera, F. T. K. Lau, C. Mattos, S. Michnick, T. Ngo, D. T. Nguyen, B. Prodhom,

W. E. Reiher, B. Roux, M. Schlenkrich, J. C. Smith, R. Stote, J. Straub, M.

Watanabe, J. Wiorkiewicz-Kuczera, D. Yin, and M. Karplus. 1998. All-Atom

Empirical Potential for Molecular Modeling and Dynamics Studies of Proteins. J.

Phys. Chem. B 102:3586-3616.



mechanics in reproducing protein conformational distributions in molecular

dynamics simulations. J Comput Chem 25:1400-1415.



9. Darden, T., D. York, and L. Pedersen. 1993. Particle mesh Ewald: An N•log(N)

method for Ewald sums in large systems. The Journal of Chemical Physics

98:10089-10092.

302

10. Ryckaert, J.-P., G. Ciccotti, and H. J. C. Berendsen. 1977. Numerical integration of

the cartesian equations of motion of a system with constraints: molecular dynamics

of n-alkanes. Journal of Computational Physics 23:327-341.

11. Feller, S. E., Y. Zhang, R. W. Pastor, and B. R. Brooks. 1995. Constant pressure

molecular dynamics simulation: The Langevin piston method. The Journal of


12. Daigle, R., M. Guertin, and P. Lague. 2009. Structural characterization of the

tunnels of Mycobacterium tuberculosis truncated hemoglobin N from molecular

dynamics simulations. Proteins-Structure Function and Bioinformatics 75:735-747.

13. Daigle, R., J. A. Rousseau, M. Guertin, and P. Lague. 2009. Theoretical

Investigations of Nitric Oxide Channeling in Mycobacterium tuberculosis

Truncated Hemoglobin N. Biophysical Journal 97:2967-2977.

14. Lovell, S. C., J. M. Word, J. S. Richardson, and D. C. Richardson. 2000. The

penultimate rotamer library. Proteins 40:389-408.

Étude thÉorique de la structure et de la dynamique

Documents