Thesis Reference

Thesis

Antioxidant profiles of new chemical entities as multifunctional hits for the treatment of neurodegenerative disorders

CRESSEND, Delphine

Abstract

Le stress oxydatif est impliqué dans de nombreuses maladies telles que les maladies neurodégénératives. Pour prévenir et retarder ces maladies, les composés antioxydants suscitent un grand intérêt. Pour accéder plus rapidement à ces molécules, des paramètres nécessaires pour la réalisation de criblages à haut débit ont été définis dans quatre tests à l'aide d'un ensemble de composés. Ces méthodes évaluent les propriétés antioxydantes de composés contre les radicaux peroxyls (test ALP et ORAC), DPPH• et ABTS•–. Ensuite, une analyse par classification ascendante hiérarchique des paramètres de puissance et de vitesse de réaction a conduit à des profils antioxydants. Ils regroupent des composés caractérisés par des propriétés antioxydantes similaires envers les trois radicaux et de leur activité en présence de la protéine. De plus, une procédure définissant un ordre précis pour la réalisation des tests a pu être mise en place. Pour mieux caractériser les antioxydants, une étude électrochimique a permis d'estimer le potentiel formel de l'acide caféique et de la mélatonine [...]

Reference

CRESSEND, Delphine. Antioxidant profiles of new chemical entities as multifunctional hits for the treatment of neurodegenerative disorders. Thèse de doctorat : Univ. Genève, 2010, no. Sc. 4271

URN : urn:nbn:ch:unige-132486 DOI : 10.13097/archive-ouverte/unige:13248

Available at: http://archive-ouverte.unige.ch/unige:13248

Disclaimer: layout of this document may differ from the published version.

1 / 1 UNIVERSITÉ DE GENÈVE FACULTÉ DES SCIENCES Section des sciences pharmaceutiques Professeur P.-A. Carrupt

Antioxidant Profiles Of New Chemical Entities As Multifunctional Hits For The Treatment Of Neurodegenerative Disorders

THÈSE

présentée à la Faculté des sciences de l’Université de Genève pour obtenir le grade de Docteur ès sciences, mention interdisciplinaires

par Delphine Cressend de Annecy (France)

Thèse No 4271

GENÈVE Atelier d’impression ReproMail 2010

UNIVERSITÉ DE GENÈVE FACULTÉ DES SCIENCES Section des sciences pharmaceutiques Professeur P.-A. Carrupt

Antioxidant Profiles Of New Chemical Entities As Multifunctional Hits For The Treatment Of Neurodegenerative Disorders

THÈSE

présentée à la Faculté des sciences de l’Université de Genève pour obtenir le grade de Docteur ès sciences, mention interdisciplinaires

par Delphine Cressend de Annecy (France)

Thèse No 4271

GENÈVE Atelier d’impression ReproMail 2010

Remerciements

Je tiens à exprimer ma plus sincère reconnaissance au Professeur Pierre- Alain Carrupt pour m’avoir permis d’effectuer ce travail de thèse au sein du groupe de pharmacochimie. Je le remercie de m’avoir fait confiance tout au long de ces années et d’avoir mis à ma disposition les moyens pour réaliser ce travail dans de très bonnes conditions. Merci aussi pour sa disponibilité et pour son soutien.

Je remercie profondément le Docteur Christophe Roussel qui m’a transmis les compétences nécessaires en électrochimie, domaine qui m’était inconnu et qui se trouve être intéressant et utile. Je le remercie également pour sa patience et sa disponibilité ainsi que pour sa bonne humeur et son enthousiasme.

Je remercie également le Docteure Marianne Reist-Oechslin pour son soutien, son encadrement et pour tous les bons conseils qu’elle m’a donné au cours de ces années.

Merci au Professeur Olivier Duval, au Docteure Marianne Reist-Oechslin, au Docteur Christophe Roussel et au Docteur Philippe Christen, membres du jury de thèse, pour leur lecture attentive du manuscrit, leurs remarques pertinentes et leurs précieuses recommandations.

Je remercie tous mes collègues de Pharmacochimie, en particulier le Docteure Liliana Sintra-Grilo et Céline Le Bourdonnec pour leur bonne humeur et les bons moments passés au 418, mais aussi Bénédicte Gross-Valoton, Fabrice Gillerat, Virginie Beyeler, Christophe Fancey, Laurent Starrenberger, Sandra Alvarez, Cédric Devard, Elisabeth Favre, les Drs. Alessandra Nurisso, Sophie Martel, Antoine Daina et Alessandra Zizzari, Philippe Eugster, Karine Vuignier, Nils Oberhauser, Christina Pomilio, Kevin Nadin, les Drs. Juan Bravo, Julien Boccard, Saviana Di Giovani, Francesca Bertolini, Vincent Gasparik et Yveline Henchoz pour avoir rendu ma thèse agréable. Je remercie également Sylvia Passaquay-Rion de m’avoir aidée dans les démarches administratives, pour sa bonne humeur et pour sa capacité à référencer un grand nombre de publications en un temps très court ! Un merci aussi aux Drs. Hanjörg Eder et Elisabeth Rivara-Minten ainsi qu’à Jessica Ortelli pour leur bonne humeur pendant les TPs et pour leur soutien. Merci à Julia Dreyer, Fabien Guarrasi et Jérémy De Mooij pour leur contribution à mon travail de thèse et pour leur sympathie.

J’adresse un merci particulier au Docteure Amandine Lebeau-Guillot qui m’a accompagnée tout au long de ma thèse, qui m’a soutenue et aidée grâce à son énergie. Merci aussi pour ces bons moments de rigolade !

Enfin, je remercie mes amis et particulièrement Yannick et Cyndy, ma belle-famille, ma famille et particulièrement mes parents, mon frère et mon mari Julien pour leur soutien et leur présence tout au long de ces années.

Table of content

Résumé de la thèse

Abbreviations

Chapter 1 Introduction ...... 1

1-1 Oxidative stress ...... 1

1-2 Antioxidant defenses ...... 4

1-2.1 Enzymatic antioxidant defenses ...... 4 1-2.2 Non-enzymatic antioxidant defenses ...... 6

1-2.2.1 Endogenous antioxidant compounds ...... 6

1-2.2.2 Exogenous antioxidant compounds ...... 8

1-2.2.2.1 Vitamins ...... 8

1-2.2.2.2 Polyphenols ...... 9

1-2.2.2.3 Other antioxidant compounds ...... 12

1-3 Oxidative stress-induced diseases ...... 15

1-3.1 Alzheimer’s disease ...... 15 1-3.2 Parkinson’s disease ...... 17 1-3.3 Antioxidant strategies in Alzheimer’s and Parkinson’s diseases ...... 18

1-4 Multifunctional drugs ...... 19

1-5 Scope and aims of the thesis ...... 23

Chapter 2 Determination of parameters for the characterization of new antioxidant compounds ...... 43

2-1 Introduction ...... 43

2-2 Materials and methods ...... 46

2-2.1 Materials ...... 46 2-2.2 Methods ...... 47

2-2.2.1 ALP assay ...... 47

2-2.2.2 ORAC assay ...... 49

2-2.2.3 DPPH• assay ...... 50

2-2.2.4 ABTS•– assay ...... 51

2-3 Results and discussion ...... 53

2-3.1 Screening parameters ...... 53

2-3.1.1 ALP assay ...... 53

2-3.1.2 ORAC assay ...... 58

2-3.1.3 DPPH• assay ...... 60

2-3.1.4 ABTS•– assay ...... 66

2-3.2 Ranking of the compounds into potency classes ...... 70 2-3.3 Similarity of ALP assay with ORAC assay ...... 74

2-4 Conclusion ...... 75

Chapter 3 Antioxidant profiles as a result of a clustering analyses ...... 79

3-1 Introduction – How to perform a cluster analysis ...... 79

3-2 Materials and methods ...... 83

3-2.1 Materials ...... 83 3-2.2 Methods ...... 84

3-2.2.1 Antioxidant potency toward the three radicals ...... 84

3-2.2.2 Cluster analysis ...... 84

3-3 Results and discussion – Formation of antioxidant profiles ...... 85

3-3.1 Kinetics cluster analysis ...... 85 3-3.2 Potency cluster analysis ...... 89 3-3.3 Antioxidant profiles ...... 93

3-4 Antioxidant assay procedure as a strategy to reach the antioxidant profiles ...... 96

3-4.1 Determination of the antioxidant properties of trolox®-derivatives as an application of the method ...... 101

3-5 Conclusion ...... 105

Chapter 4 Estimation of formal potential thanks to ferrocene- derivatives using cyclic voltammetry ...... 109

4-1 Introduction ...... 109

4-1.1 Cyclic voltammetry, a relevant method ...... 109 4-1.2 Formal standard potential ...... 110 4-1.3 Electrochemical reactions ...... 111

4-1.3.1 Mass transport ...... 112

4-1.3.2 Electron transfer ...... 113

4-1.3.3 Reversible electron transfer ...... 113

4-1.3.4 Quasi-reversible electron transfer ...... 116

4-1.3.5 Irreversible electron transfer ...... 117

4-1.4 Coupled homogeneous chemical reaction ...... 118 4-1.5 Materials ...... 120

4-1.5.1 Electrochemical cell ...... 120

4-1.5.2 Working electrode (WE) ...... 120

4-1.5.3 Reference electrode (RE) ...... 122

4-1.5.4 Counter electrode (CE) ...... 122

4-1.5.5 Solvents and electrolytes ...... 123

4-1.6 Double layer ...... 123 4-1.7 Chronoamperometry ...... 124 4-1.8 Aim of the study ...... 125

4-2 Materials and methods ...... 126

4-2.1 Materials ...... 126 4-2.2 Methods ...... 127

4-3 Results and discussion ...... 127

4-3.1 Validation of the system ...... 127 4-3.2 Electrochemical behavior of studied compounds alone ...... 131 4-3.2.1 Cyclic voltammetry study of caffeic acid ...... 131

4-3.2.2 Cyclic voltammetry study of melatonin ...... 133

4-3.2.3 Cyclic voltammetry study of para-hydoquinone ...... 134

4-3.3 Determination of the formal standard potential of caffeic acid in the ferrocene system ...... 135

4-3.3.1 Caffeic acid and aminoferrocene ...... 135

4-3.3.2 Caffeic acid and ferrocene ...... 137

4-3.3.3 Caffeic acid and ferrocenecarboxylic acid ...... 140

4-3.3.4 Electrochemical study of the caffeic acid-ferrocene solution by chronoamperometry ...... 142

4-3.4 Determination of the formal standard potential of melatonin in the ferrocene system ...... 143

4-3.4.1 Melatonin and aminoferrocene ...... 143

4-3.4.2 Melatonin and ferrocene ...... 145

4-3.4.3 Melatonin and ferrocenecarboxylic acid ...... 147

4-3.5 Para-hydroquinone as a proof of principle ...... 149

4-3.5.1 Para-hydroquinone and aminoferrocene acid ...... 149

4-3.5.2 Para-hydroquinone and ferrocene ...... 150

4-3.5.3 Para-hydroquinone and ferrocenecarboxylic acid ...... 153

4-4 Conclusions ...... 154

Chapter 5 Structure activity relationship ...... 159

5-1 Introduction ...... 159

5-2 Materials and methods ...... 161 5-2.1 Materials ...... 161 5-2.2 Methods ...... 162

5-2.2.1 DPPH• assay ...... 162

5-2.2.2 ABTS•− assay ...... 162

5-3 Results and discussion ...... 162

5-3.1 Reference compounds ...... 163

5-3.1.1 Caffeic acid (profile A2) ...... 163

5-3.1.2 Ascorbic acid (profile A8) ...... 165

5-3.1.3 Trolox® (profile A8) ...... 167

5-3.1.4 Chlorogenic acid (profile B1) ...... 169

5-3.1.5 Gallic acid (profile B1) ...... 170

5-3.1.6 Resveratrol (profile B1) ...... 172

5-3.1.7 Mangiferin (profile B2) ...... 173

5-3.1.8 Quercetin (profile B2)...... 175

5-3.1.9 Cysteine (profile B8) ...... 176

5-3.1.10 Melatonin (profile C5) ...... 178

5-3.1.11 Phenol (profile C5) ...... 179

5-3.1.12 Glutathione (profile C6) ...... 180

5-3.1.13 Uric acid (profile C7) ...... 181

5-3.1.14 Aniline and N,N-diethylaniline (profile E4) ...... 182

5-3.1.15 Mannitol ...... 183

5-3.1.16 α-tocopherol ...... 184

5-3.2 Aurone- and azaaurone-derivatives ...... 184

5-3.2.1 Aurone-derivatives ...... 185

5-3.2.2 Azaaurone-derivatives ...... 188

5-3.3 N-derivatives of (3S,4S)- and (3R,4R)-pyrrolidine-3,4-diols ...... 191 5-3.4 Racemic conduramine B-1 analogs ...... 193 5-3.5 Secondary amines ...... 195 5-3.6 Esters of (3S, 4S)-pyrrolidine-3,4-diols ...... 195 5-3.7 Racemic conduramine F-1 analogs ...... 195 5-3.8 Racemic conduramine F-1 epoxides ...... 197 5-3.9 Compounds extracted from Jacaranda caucana ...... 198 5-3.10 1-substituted and 2-substituted pyrazolopyrrolizine derivatives ...... 202 5-3.11 Xanthone-derivatives ...... 204 5-3.12 Quercetin-derivatives extracted from Flavo Psidium Cattleianium ... 211 5-3.13 Trolox®-derivatives ...... 216 5-3.14 SAR overview regarding the antioxidant profiles ...... 218

5-4 Conclusion ...... 219

Conclusions and perspectives ...... 229

Annexes

I Supporting data

II Structures and antioxidant properties of the compounds under study

III Publications

Frédéric Martin, Anne-Emmanuelle Hay, Delphine Cressend, Marianne Reist, Livia Vivas, Mahabir P. Gupta, Pierre-Alain Carrupt, Kurt Hostettmann. Antioxidant C- Glucosylxanthones from the Leaves of Arrabidaea patellifera, J. Nat. Prod. 2008, 71, 1887–1890

Frédéric Martin, Anne-Emmanuelle Hay, Valentin R. Quinteros Condoretty, Delphine Cressend, Marianne Reist, Mahabir P. Gupta, Pierre-Alain Carrupt, Kurt Hostettmann. Antioxidant Phenylethanoid Glycosides and a Neolignan from Jacaranda caucana, J. Nat. Prod. 2009, 72, 852–856

Raimana Ho, Aude Violette, Delphine Cressend, Phila Raharivelomanana, Pierre-Alain Carrupt, Kurt Hostettmann. Antioxidant potential and radical scavenging effects of flavonoids from the leaves of Psidium cattleianum grown in French Polynesia. Accepted in Natural Product Research.

Résumé

Les antioxydants sont devenus un argument de vente dans la société actuelle. Ils doivent cette promotion à leur effet bénéfique pour lutter contre le stress oxydatif. L’implication du stress oxydatif a été prouvé dans de nombreuses maladies comme le cancer, les maladies cardiovasculaires ou encore les maladies neurodégénératives. Pour prévenir, retarder voire même guérir ces maladies, les médicaments possédant des propriétés antioxydantes suscitent donc un grand intérêt. Cependant, le terme de stress oxydatif regroupe un grand nombre d’espèces réactives induisant des oxydations sur les biomolécules ce qui, souvent, ne permet pas une caractérisation complète et précise des antioxydants. En effet, un composé peut avoir des propriétés antioxydantes contre seulement certaines espèces réactives. De plus, un antioxydant peut avoir une efficacité différente selon les biomolécules impliquées, même si les dégâts oxydatifs sont provoqués par la même espèce réactive. Devant cette multitude de possibilités, il serait irréalisable de tester les propriétés antioxydantes de tous les composés connus à ce jour pour trouver des molécules médicamenteuses. Pour accéder plus rapidement à des molécules d’intérêt, des paramètres nécessaires pour la réalisation de criblages à haut débits ont été définis dans quatre tests à l’aide d’un ensemble de composés, dont des antioxydants de référence. Ces méthodes évaluent les propriétés antioxydantes de composés contre les radicaux peroxyls, DPPH• et ABTS•–. Les radicaux peroxyls sont utilisés dans le test ORAC et dans un test qui met en évidence l’oxydation d’une protéine, la phosphatase alcaline. Ensuite, les activités antioxydantes d’un plus grand panel de composés de structures variées ont été évaluées. Les composés actifs ont été décrits par des paramètres de puissance et de vitesse de réaction. Une analyse par classification ascendante hiérarchique de ces paramètres a conduit à des profils antioxydants. Ils regroupent des composés caractérisés par des propriétés antioxydantes similaires envers les trois radicaux et de leur activité en présence de la protéine. De plus, une procédure définissant un ordre précis pour la réalisation des tests a pu être mise en place. Seulement deux tests sur les quatre sont nécessaires pour atteindre certains profils et ainsi caractériser le composé grâce aux propriétés antioxydantes de ce profil. L’étude électrochimique des pouvoirs réducteurs et oxydants permet d’obtenir des informations utiles pour la caractérisation de composés. Le potentiel formel, obtenu par des méthodes électrochimiques, est un paramètre qui décrit la réactivité des composés. Pour le moment, aucune méthode ne permet d’évaluer efficacement ce paramètre pour des antioxydants. L’idée est donc d’estimer le potentiel formel de composés grâce à des intervalles de potentiel. Les bornes de ces intervalles sont fixées par des espèces électrochimiques bien connues telles que le ferrocene, l’aminoferrocene et l’acide carboxyferrocene. La capacité d’un composé à régénérer le ferrocene, ou l’un de ses dérivés, oxydé électrochimiquement permet de placer son potentiel formel sur l’échelle de potentiel. Ainsi, le potentiel formel de l’acide caféique est estimé entre 2 et 233 mV et celui de la mélatonine et de la para-hydroquinone est supérieur à 313 mV. Les caractéristiques antioxydantes obtenues et la diversité des structures des composés ont mis en évidence une relation de structure-activité dans les profils antioxydants. De plus, de nouvelles structures présentant des propriétés réductrices se sont révélées efficaces contre les radicaux peroxyl, DPPH• et ABTS•–. Enfin, certains composés ont montré une activité seulement dans la protection de la protéine ce qui suggère, pour ces composés, une capacité à se lier sur la protéine pour subir l’oxydation à la place de l’enzyme. Abbreviations

Δ Temperature Δ BDE BDE of a compound – BDE of phenol Δ IP IP of a compound – IP of phenol ε Molar extinction coefficient λ Wavelength

λEm Emission wavelength

λEx Excitation wavelength ν Scan rate [compound] Concentration of the compound 8-OH-G 8-Hydroxyguanosine 8-OH-dG 8-Hydroxy-2-deoxyguanosine Aβ Beta amyloid peptide AAPH 2,2’-Azobis(2-methylpropionamidine) dihydrochloride Abs Absorbance ABTS 2,2’-Azino-bis-(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt ABTS•– ABTS radical anion AChE Acetylcholinesterase AcOH Acetic acid AD Alzheimer’s disease AFMK N1-acetyl-N2-formyl-5-methoxykynuramine Ag Silver atom ALP Alkaline phosphatase AMK N1-acetyl-5-methoxykynuramine APP Amyloid β precursor protein APOE Apolipoprotein E AO Antioxidant AU Absorbance unity BDE Bond dissociation enthalpy BHA Butylated hydroxyanisole BHT Butylated hydroxytoluene Br Bromide atom c3OHM Cyclic 3-hydroxymelatonin C Homogeneous chemical transfer in electrochemical process Cl Chlorine atom COMT Catechol-O-methyl transferase Cr Chromium atom Cu Copper atom CuZn-SOD Superoxide dismutase containing copper and zinc atom as cofactors CV Cyclic voltammetry DMSO Dimethylsulfoxide DPPH• 2,2-diphenyl-1-picrylhydrazine E Heterogeneous electron transfer in electrochemical process

Epa Anodic peak potential

Epc Cathodic peak potential

ER50 Ratio of the compound concentration over the radical concentration necessary to reduce 50% of radical ET Electron transfer Et Ethyl residue EtOH Ethanol F Fluoride atom Fc Ferrocene FcCOOH Ferrocenecarboxylic acid Fe Iron atom Fl Fluorescein

Fluonon-ox Fluorescence value in presence of non-oxidized controls in ORAC assay

Fluoox Fluorescence value in presence of oxidized controls in ORAC assay

Fluosample Fluorescence value in presence of test compounds in ORAC assay

FcNH2 Aminoferrocene G6PD Glucose-6-phosphate dehydrogenase GC Glassy carbone GPx Glutathione peroxidase GRd Glutathione reductase GS• Glutathione radical GSH Glutathione GSSG Disulfide glutathione GST Glutathione S-transferase

H2O2 Hydrogen peroxide HAT Hydrogen atom transfer haox Hydrolytic activity in presence of oxidized controls in ALP assay hanon ox Hydrolytic activity in presence of non-oxidized controls in ALP assay hasample Hydrolytic activity in presence of test compounds in ALP assay HCA Hierarchical cluster analysis HNE 4-hydroxy-2-nonenal HOCl Hypochlorous acid HTS High-throughput screening IP Ionization potential ipa Anodic peak current ipc Cathodic peak current l Path length LDL Low-density lipoprotein L-Dopa Levodopa MAO-A Monoamine oxidase A MAO-B Monoamine oxidase B Me Methyl residue

MgCl2 Magnesium chloride Mn Manganese atom Mn-SOD Superoxide dismutase containing manganese atom as cofactor MU 4-Methylumbelliferone MUP 4-Methylumbelliferyl phosphate n Number of transferred electrons in an electrochemical process NaOH Sodium hydroxide NCE New chemical entity NO• Nitric oxide radical NOS Nitric oxide synthase NFT Neurofibrillary tangles

•– O2 Superoxide radical anion OH• Hydroxyl radical OMe Methoxy residue ONOO– Peroxynitrite ORAC Oxygen radical absorbance capacity pEC50 Negative logarithm of the half maximal activity Ph Phenyl residue p-HQ para-Hydroquinone PD Parkinson’s disease pKa Logarithm of the acid dissociation constant pKb Logarithm of the basic dissociation constant Pt Platinum RCS Reactive chlorine species RE Reference electrode RNS Reactive nitrogen species ROO• Peroxyl radical ROS Reactive oxygen species rs Spearman’s ranking coefficient SET Single electron transfer SMe Thiomethyl residue SOD Superoxide dismutase SPLET Sequential proton loss electron transfer tBu Tertiobutyl residue WE Working electrode Zn Zinc atom

ZnCl2 Zinc chloride

Chapter 1

Chapter 1 Introduction

1-1 Oxidative stress

The human body is a complex organism made up of about 100 trillion cells which carry out the vital functions. To exist and to accomplish their role, the cells require biomolecules, including nucleic acids, lipids, carbohydrates, vitamins, and radicals. The free radicals defined as chemical species containing an unpaired electron in their outer orbital possess a high reactivity inducing a redox homeostasis within the cell1,2. The free radicals are derived either from oxygen, nitrogen, or chlorine atoms, and are included in groups labeled reactive oxygen species (ROS), reactive nitrogen species (RNS), and reactive chlorine species (RCS), respectively (Table 1-1)3,4. Besides free radicals such as the superoxide

•– • anion (O2 ) and the nitric oxide radical (NO ), non-radical reactive species exist

– including hydrogen peroxide (H2O2) and peroxynitrite (ONOO ). The free radicals are mainly produced endogenously within the mitochondria during the electron transport chain. Among the large range of physiological roles, they are involved at low/moderate concentration in the defense against infectious agents5, in the regulation of cardiac, and vascular cell functioning6, and in the induction of mitogenic response7. Nitric oxide also has a key role in neurotransmission, smooth muscle relaxation, blood pressure regulation, and immune regulation7. The endogenous production of ROS and RNS is carried out not only by xanthine oxidase, superoxide dismutase (SOD), and nitric oxide synthase (NOS) among other enzymes but also by uncontrolled processes. During the mitochondrial

•– electron transport chain, a 1-3% electron leak occurs and generates O2 , which does not contribute to the reduction of oxygen to water.

Chapter 1

Table 1-1: Reactive species produced from endogenous and exogenous sources.

ROS RNS RCS

Free radicals:

Superoxide O2•– Nitric oxide NO• Chlorine atom Cl•

Hydroxyl OH• Nitrogen dioxide NO2•

Hydroperoxyl HO2•

Peroxyl RO2• Alkoxyl RO•

Carbonate CO3•–

Carbon dioxide CO2•–

Non-radicals:

Hydrogen H2O2 Nitrous acid HNO2 Hypochlorous HOCl peroxide Nitrosyl cation NO+ acid

– Singlet oxygen 1O2Δg Nitroxyl anion NO Nitryl NO2Cl

Organic ROOH Dinitrogen tetroxide N2O4 (nitronium) peroxides Dinitrogen trioxide N2O3 chloride

– Ozone O3 Peroxynitrite ONOO Chloramine

Peroxynitrite ONOO– Peroxynitrous acid ONOOH Chlorine gas Cl2

Peroxynitrous ONOOH Nitronium (nitryl) NO2+ acid cation Hypobromous HOBr Alkyl peroxynitrites ROONO acid Nitryl (nitronium) NO2Cl Hypochlorous HOCl chloride acid

Furthermore, exogenous causes such as UV radiation and tobacco smoke lead to an increase of the ROS/RNS/RCS concentrations8-10. These processes, by raising the amount of ROS/RNS/RCS, lead to the shift toward more oxidizing conditions of the redox homeostasis within cells. A cell differentiation is initiated by a slight shift toward oxidizing environment, while a larger shift leads to apoptosis and necrosis8,11,12. Indeed, a high concentration of free radicals induces oxidative damage on proteins, lipids, carbohydrates, and DNA3. Therefore, mutations and losses of enzymatic activity, among other oxidative damage (Table 2

Chapter 1

1-2), occur within cells and disturb the cell functioning. Irreversible damage may trigger cell death.

Table 1-2 : Most common biomarkers of oxidized biomolecules.

Biomolecules Biomarkers of oxidative damages

Protein Protein carbonyls Nitrotyrosine S-Glutathionylated proteins Advanced glycation end products (AGE)

Lipid Malonaldehyde (MDA) 4-Hydroxy-2-nonenal (HNE) Acrolein Isoprostanes Neuroprostanes

DNA 8-Hydroxy-2-deoxyguanosine (8-OH-dG) 8-Hydroxyguanosine (8-OH-G)

Carbohydrate Advanced glycation end products (AGE)

Defense systems have been developed to counteract the ROS/RNS overproduction through preventative, repair, and antioxidative mechanisms. The antioxidant (AO) system acts at the redox homeostasis by shifting it toward a more reducing environment. It prevents the radical overproduction13, the metal ion-promoted radical production14-16, the ROS/RNS propagation17, and repairs the oxidative damage18. In healthy organisms, the rate of ROS/RNS production and the rate of removal by antioxidants are balanced. However, an imbalance due to either an increase of ROS/RNS production or to a deficiency in the antioxidant activity may occur. This imbalance is referred to as oxidative stress.

Chapter 1

AO systems ROS RNS, RCS

Figure 1-1 : Oxidative stress, an imbalance between a high amount of reactive species and a deficiency in antioxidant system activity.

1-2 Antioxidant defenses

The defense systems countering oxidative stress are capable of interacting with reactive species and converting them into stable and non-toxic substances. Enzymatic and non-enzymatic mechanisms are involved in the defense systems. An antioxidant compound was firstly defined by “any substance that, when present at low concentration compared to that of an oxidizable substrate, significantly delays or prevents oxidation of that substrate”19. However, repair systems, inhibitors of reactive species generation or chaperons are not taken into account in this definition. Furthermore, the large excess of plasma albumin used to bind copper and protect extracellular targets does not match to the definition. An antioxidant compound is therefore defined by “any substance that delays, prevents or removes oxidative damage to a target molecule”11.

1-2.1 Enzymatic antioxidant defenses

Superoxide dismutase (SOD), catalase, and glutathione peroxidase (GPx) are the primary enzymes included in the enzymatic antioxidant defenses. They induce a direct detoxification of ROS/RNS. Superoxide dismutase catalyzes the dismutation of superoxide anion into oxygen and hydrogen peroxide (Equation (1.1)).

Chapter 1

i−+ 22OH22+→+O HO22 (1.1)

Copper (Cu), zinc (Zn), and manganese (Mn) are the cofactors of the three forms existing in human. CuZn-SOD is located in the cytoplasm (SOD1) and in the extracellular matrix (SOD3) while Mn-SOD is found within the mitochondria (SOD2). A promising study has reported the preservation of the SOD effectiveness after oral administration20.

Catalase carries out the decomposition of hydrogen peroxide to oxygen and water (Equation (1.2)). It is a tetramer containing four porphyrin heme groups. A deficiency in catalase activity promotes the development of type II diabetes21, 22.

22HO22→+ HO 2 O 2 (1.2)

Glutathione peroxidase (GPx) is a selenium-containing tetramer, present in several cellular locations. It converts hydrogen peroxide into water by using glutathione (GSH) as a reducing agent (Equation (1.3)). The resulting oxidized GSH is the disulfide glutathione (GSSG).

22GSH+ H22 O→+ GSSG H2 O (1.3)

The secondary enzymes, which are glutathione reductase (GRd), glucose-6- phosphate dehydrogenase (G6PD), and cytosolic glutathione S-transferase (GST), promote the elimination of reactive species. They decrease the peroxide levels and maintain a steady supply of metabolic intermediates, such as GSH, necessary for an optimum activity of the primary enzymes.

Chapter 1

1-2.2 Non-enzymatic antioxidant defenses

The endogenous antioxidants, including glutathione, uric acid, albumin, and bilirubin, and the exogenous antioxidant, such as ascorbic acid, α-tocopherol, and resveratrol, form the non-enzymatic antioxidant defense family. These compounds directly counteract ROS, RNS, and RCS through electron or hydrogen atom transfer, leading to non-reactive species, or act as repair agents on damaged sites. Some non-enzymatic antioxidant compounds, along with a glimpse of their properties, are further described hereinafter.

1-2.2.1 Endogenous antioxidant compounds

Glutathione (γ-L-glutamyl-L-cysteinglycine, GSH) is a tripeptide enzymatically formed by cysteine, glutamate, and glycine. This major endogenous antioxidant is produced within the cells and is located in the cytosol, in the mitochondra, in the peroxisomes, in the endoplasmic reticulum, and in the cellular matrix. The redox state of the GSSG/GSH is a representative indicator of the environmental redox homeostasis1,2. Besides, glutathione depletion is a feature of apoptotic cell death12,23,24. Thanks to the large localization sites and to its oxydoreduction properties, the tripeptide is involved in maintaining the redox homeostasis and in many physiological reactions, such as DNA synthesis and repair25. Glutathione has a key role in counteracting oxidative stress since it may act directly or indirectly25,26. In its reduced form (GSH), glutathione is a HOCl- removing agent27, a potent peroxyl radical scavenger, a weak peroxynitrite, and hydroxyl radical detoxifying agent28,29. It is necessary as a substrate of some antioxidant enzymes30,31, regenerates other non-enzymatic antioxidants and is involved in the elimination of xenobiotic compounds through conjugation32. After oxidation, glutathione readily interacts with another molecule of glutathione leading to a disulfide bridge (GSSG). However, pro-oxidant effects of the thiol have been reported in an iron ion-mediated hydroxyl radical-rich environment29.

Chapter 1

Uric acid is produced by xanthone oxidase as the final product of purine catabolism in humans. It is excreted during the urination process. An unusually high concentration of uric acid in blood or in urine is associated with medical disorders such as gout, cardiovascular diseases, cancer, diabetes33-36. On the other hand, a low concentration is associated with multiple sclerosis37. Other sources of uric acid are animal food products and vegetables such as cauliflower, asparagus, and spinach. The endogenous water-soluble compound interacts with peroxyl radicals before they attack polyunsaturated fatty acids, that is in the initiation step of lipid peroxidation38. It is also capable of detoxifying ONOO–,

• • 28,39,40 •– 41 18 OH , ROO , HOCl , and O2 , preventing DNA damage from oxidation , forming urate metal ion complexes by binding to the metals iron and copper, and has anticarcinogen properties40. However, the ability to bind copper may induce pro-oxidant effects42. Melatonin (N-acetyl-5-methoxytryptamine) is a natural hormone widespread in plants, yeasts, and in vertebrates. Referred to as the “hormone of darkness”, melatonin plays a key role among other functions in circadian rhythms in vertebrates43 as well as in the regulation of seasonal reproduction, and retinal functions44. It is produced by the pineal gland and is also found in high concentration within other organs and body fluids45. Thanks to its electron- rich indolamine, melatonin protects DNA, membrane lipids, and proteins from

•46-48 1 49,50 16,50,51 •52,53 – oxidative damages induced by OH , O2 , H2O2 , NO , ONOO 48,50,52, HOCl, NOS54,55 both in vitro and in vivo. Melatonin is also able to bind metals (aluminum, cadmium, copper, iron, lead, zinc, lithium, potassium, sodium, calcium)14-16. A main feature of melatonin is its amphiphilic property due to the O-methyl and N-acetyl residues which allows its propagation in many cellular compartments. The indolamine can also counteract the ROS- and RNS-induced damages by stimulating the gene expression for antioxidant enzymes including SOD56,57, catalase58, GPx57-59, GRd59, and G6PD60. The antioxidant properties of oxidation products from melatonin lead the indolamine to sometimes counteract up to 10 molecules of ROS and RNS61-63. The large spectra of antioxidant properties makes melatonin a compound of great interest, which has been widely reported44,64.

Chapter 1

1-2.2.2 Exogenous antioxidant compounds

1-2.2.2.1 Vitamins

L-ascorbic acid, referred to as vitamin C, is present in plant forms and in the liver and kidney of most animals. However, due to a genetic mutation, some mammals, among them humans, are unable to synthesize it. They need to consume foods rich in the nutrient in order to prevent a vitamin C deficiency, which may lead to diseases such as scurvy65. The richest natural sources of L- ascorbic acid are plants, fruits, and vegetables. Acerola, kiwifruit, and broccoli are among the fruit and vegetables containing the highest concentration of vitamin C. This water-soluble compound is involved in several physiological functions and exhibits antioxidant and pro-oxidant properties. Indeed, its ability to reduce some metal agents such as Fe(III), Cr(II), and Cr(VI)66-68 is one of ascorbic acid’s pro-oxidant effects. However, it is capable of scavenging ROS and

67 69 •69 •28 –28 RNS such as H2O2 , OH , ROO , and ONOO . Thanks to its antioxidant capacities, L-ascorbic acid is used as a food additive (European number E300). For instance, it may prevent the flavor loss from the polyphenol oxidation in wine by undergoing the oxidation before the other components and by regenerating the oxidized polyphenols70,71. It is also capable of stimulating SOD, catalase, and GST72, and of regenerating α-tocopherol71,73. The generic term of vitamin E stands for a family of four forms, namely α-, β-, γ-, and δ-tocopherol. They differ in the number of methyl groups on the chromanol moiety. α-Tocopherol is the major lipid-soluble antioxidant in cells, thanks to the phytyl tail and has the highest bioavailability. The recommended daily intake is 15 mg, which can be found in nuts, sunflower, and safflower oils, or in green leafy vegetables, among other sources. The lipophilic feature and the chromanol moiety play a key role in preventing lipid peroxidation; α-tocoperol is among the most powerful chain-breaking antioxidants. It is therefore a potent ROO• scavenger17,74,75. Furthermore, it stimulates SOD, catalase, and GST72, and

• •– scavenges OH and O2 but with a weak potency. However, it fails to detoxify RNS69,76. In presence of specific metals, vitamin E exhibits pro-oxidant properties

Chapter 1 inducing toxic effects77-79. In addition to ascorbic acid, vitamin E may also be regenerated by glutathione.

1-2.2.2.2 Polyphenols

Polyphenols are naturally occurring compounds in plants, fruits, cereals, vegetables, etc. They are of great interest not only for the treatment of diseases but are also used as additive compounds in food, cosmetic, and pharmaceutical industries. Caffeic acid (3,4-dihydroxycinnamic acid) is a natural product, present in almost every plants due to its involvement in the biosynthesis of lignin. It is also widely present in food such as fruits, vegetables, wine, olive oil, tea, and coffee bean80. Its therapeutic property extent covers a broad spectrum since dihydroxycinnamic acid acts as an inflammatory81 and mutagenic82-84 inhibitor. It prevents and inhibits the lipid peroxidation85-87, acts as an inhibitor of the nickel- induced oxidative damages88, and inhibits DNA methylation by stimulating the formation of S-adenosyl-L-homocysteine82. Furthermore, it is capable of regenerating the antioxidant defenses in rat liver88. Caffeic acid exhibits such a high number of therapeutic properties because it is a potent metal ion

86,88,89 •90 •87,91 •–86 92 •93,94 chelator and a potent scavenger of OH , ROO , O2 , H2O2 , NO , and ONOO–95. Caffeic acid is present in blood circulation of rats after ingestion and is absorbed by humans in the stomach and in the small intestine96. This bioavailability has been reported in in vivo antioxidant assays94,97-99. Chlorogenic acid is the major phenolic compound in coffee and is also present in apple, beans, sunflower, almond as well as many other plants, fruits, and vegetables80. It is an ester of caffeic acid with a quinic acid residue. The extent of chlorogenic acid properties is as wide as that of caffeic acid both in vitro and in vivo. It inhibits DNA damage82,100, the peroxyl-, nitric oxide-, and superoxide anion-induced lipid peroxidation87,90,101,102. It is an anticarcinogenic agent99, modifies the concentration of cholesterol, and triacylglycerol in plasma and liver103. Caffeic acid and chlorogenic acid act as inhibitors of the lipid peroxidation chain initiation and as chain-breaking antioxidants87,90.

Chapter 1

Gallic acid (3,4,5-trihydroxybenzoic acid) is an organic compound widely present in plants, tea leaves, fruits (e.g. blackberry, mango, raspberry), oak bark. It occurs both free and as part of tannins. It is a powerful antioxidant, due to its

•– • • –95,104-106 detoxifying activities of O2 , OH , ROO , and ONOO , and has neuroprotective effect due to both its antioxidant abilities and hydrophobicity107. It acts as a chain-breaking agent in lipid peroxidation propagation106,108 and counteracts the aging-related activity decrease of catalase and GPx108. Furthermore, the pro-oxidant effects of gallic acid have benefic outcomes toward some diseases109 Gallic acid ester derivatives are used in pharmaceutical field to increase the bioavailability of orally administered compounds110 and exhibit a high number of properties107,111-113. Propylgallate, octylgallate, and laurylgallate are currently used as antioxidant food additives (European Community codes E- 310, E-311, and E-312, respectively). Mangiferin (2-β-D-glucopyranosyl-1,3,6,7-tetrahyxdroxy-9H-xanthen-9-one) is a xanthone-derivative mainly found in mango and in plants. It is produced in industrial scale in Cuba for use as a nutritional supplement, in cosmetic products and phytomedicine. Some of the biological activities are gastroprotective114, hypoglycemic115, analgesic116, and anti-inflammatory117 effects. It is a powerful

•–118,119 •119,120 •119,120 scavenger of O2 , OH , ROO . Mangiferin inhibits the lipid peroxidation at the initial phase due to its iron-chelating properties121. In highly ROS-containing mitochondria, oxidized mangiferin accumulates and induces mitochondrial dysfunctions122. This feature is expected to lead to promising outcomes in the apoptosis trigger of damaged cells. Quercetin (3,3’,4’,5,7-pentahydroxyflavone) belongs to the flavonoid family and is naturally widely present in plants, fruit, vegetables, and flowers including onions, apples, berries, red grapes, tea. An intake of more than 33 mg a day has been related to a decreased risk of cardiovascular diseases123-125 and it has been reported that quercetin remains present in plasma during several hours (11- 28h)126,127. However, quercetin is not present in plasma as aglycone but it is converted to glucuronide and sulfate conjugates within several organs such as small intestine, liver, and kidney128,129. Among the wide biological effect range130,131, quercetin is a potential drug for the hypertension132 and anti-

Chapter 1 inflammatory treatment133. Its neuroprotective effect is controversial since the in vitro and in vivo studies provide contradictory results due to the aglycone form used in in vitro assays and the conjugated quercetin form present in vivo134. The powerful antioxidant properties of quercetin are due to its potent ability to

•–135 •136 –136,137 scavenge, among other ROS/RNS, O2 , NO , ONOO . Besides, the flavonoid is a inhibitor of lipid peroxidation, LDL oxidation, DNA damage138,139, and a metal chelator138. Quercetin is not a carcinogenic agent131 but can be a pro- oxidant140,141. Furthermore, oxidized quercetin preferentially reacts with thiols rather than with ascorbate, which may alter the enzymatic activities142. Therefore, some precautions to be taken in are necessary using quercetin as a nutritional supplement, such as a glutathione supply. Its antioxidative and anti- inflammatory properties have been assessed in vivo143,144. Resveratrol (3,4’,5-trihyxdroxystilbene) is a phytoalexin produced by some plants in response to injury, pathogen attack or UV exposure145,146. It is also present in red grape, blackberry, peanuts, and dark chocolate as trans- and cis- isomers. Both forms act by the same way but cis-resveratrol is less effective147. The polyphenol is one of the factors involved in the so-called “French paradox” since the resveratrol concentration in red wine seems to be sufficient to induce benefic effects on health148. It has anti-inflammatory, anti-aging effect149, and exhibits benefic effects in preventing and treating both obesity and diabetes60.

•–150 Regarding its antioxidant properties, it is capable of detoxifying O2 ,

•39,69,151,152 •138 150,153 149 OH , ROO , and H2O2 , and inhibiting nitric oxide synthase . Furthermore, resveratrol is able to bind copper, but not iron, and it is thereby capable of inhibiting Cu-induced low-density lipoprotein (LDL) oxidation38. It also exhibits a potent activity in inhibiting DNA damage39,139,152 and lipid peroxidation150,154. Besides its antioxidant properties, the pro-oxidant effects of the stilbene in presence of copper may be useful in some clinical therapies155. Due to its large property range, resveratrol is of great interest in replacing butylated hydroxyanisole (BHA) and butylated hydroxytoluene (BHT)150. BHA, and BHT are used to preserve fats and oils in food, pharmaceuticals, and cosmetics. However, doubts over their toxic and carcinogenic effects exist156,157.

Chapter 1

1-2.2.3 Other antioxidant compounds

Trolox® (6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid) is a water-soluble derivative of α-tocopherol. Trolox® equivalency is often used to assess the antioxidant properties of entities158. It is capable of interacting with ONOO–28, and ROO•28,159,160 but poorly with OH•28. It also inhibits DNA fragmentation161, lipid peroxidation159,162, low-density lipoprotein (LDL) oxidation163, and protein oxidation164,165. It is able to repair damaged DNA166 but

•– 167,168 169 fails to inhibit H2O2-, O2 -induced DNA damage , and recent study has reported the benefic effect of trolox® in enhancing the arsenic trioxide-induced anti-leukemic effects. However, a pro-oxidant effect was detected under certain conditions; trolox® is thereby involved in the copper-induced LDL oxidation163 and in the chrome-induced DNA damage68.

Mannitol (1,2,3,4,5,6-hexanol) is not only used as a sweetener in food but also as an osmotic diuretic and as an adjunct during a cardiopulmonary bypass to preserve renal function170, among other applications. It is also used to deliver drug within the brain thanks to its ability to induce a blood-brain barrier disruption171,172. In addition, mannitol is a well-known hydroxyl radical scavenger173,174 and may enhance the antioxidant enzyme activities175.

The antioxidant properties of the described compounds are summarized in Table 1-3, along with the diseases in which they may have potential benefic effects.

Chapter 1

Table 1-3: Antioxidant properties of some well-known compounds and some diseases for which they may induce benefic effects.

Antioxidant Scavenged Oxidative stress-counteracting Potential benefic compounds reactive species properties effects in

Ascorbic acid H2O269, OH•69, Activator of SOD, catalase, Cancer176 ROO•28, and GST72 ONOO–28 Cardiovascular α-tocopherol-71,73 and disease177,178 polyphenol-70, 71 regenerating agent

α-tocopherol ROO•74,75, OH•, Lipid peroxidation chain- Cancer179 , O2•–69 177 breaking agent17 Cardiovascular disease177,180

Diabetes181,182

Caffeic acid OH•90, Inhibitor of inflammation, Cancer183-185 ROO•87,90,91,138, mutagenesis , Alzheimer’s O2•–86 90, H2O292 186 NO•93,94, Lipid peroxidation (initiation- disease ONOO–90,95 inhibiting and chain-breaking agent)

Metal ion chelator

Iron reducing power138

, Chlorogenic O2•–90, OH•90, Inhibitor of DNA damage82 100 Cancer187 acid ROO•87,90,102, 87,90,101,102 ONOO–90 NO•183 Lipid peroxidation Alzheimer’s disease 188

, , Gallic acid O2•–104, OH•69, Inhibitor of CuSO4– Cancer111 192 193 ONOO–95 ROO•-induced LDL Parkinson’s oxidation104 disease112

Iron-reducing power104,189

Iron chelator190

Chain-breaking agent of lipid peroxidation106,108,191

Chapter 1

Glutathione OH•28, Activating agent of enzymatic Cardiovascular ROO•28,29, antioxidant 30,31 diseases194 NO•28, HOCl27 Regenerator of ascorbic acid Parkinson’s disease and α-tocopherol 25,26

, , Mangiferin O2•–118 195, Iron chelator118 Cancer122 195 OH•119,120, ROO•119,120 Inhibitor of lipid Cardiovascular peroxidation121 diseases195

2-deoxiribose oxidation88 Alzheimer’s disease196

Inflammation195

Mannitol OH•173,174 Preventive agent of lipid peroxidation174

, Melatonin OH•46-48, 1O249 Metal chelator14-16 Alzheimer’s , , 50, H2O216 50 51, disease198-200 56,57 NO•52,53, ONOO– Activator of SOD , 58 57-59 59 48,52, HOCl, catalase , GPx , GRd , and Parkinson’s 197 201 NOS54,55 G6PD disease

Quercetin O2•–135, NO•136, Inhibitor of LDL oxidation, Cancer 202 ONOO–136,137, lipid peroxidation138, DNA ROO•138 damage139 Alzheimer’s disease203,204 Iron reducing power138 Parkinson’s disease205,206

Hypertension207

Inflammation144

Resveratrol O2•–150, Inhibitor of lipid Cancer210-212 , , , , , OH•39 39 69 151 152, peroxidation138 208, of CuSO4-, ROO•138, ROO•-induced LDL Cardiovascular , diseases213 H2O2150 153 oxidation38, and DNA 39 oxidation Alzheimer’s 196,214,215 Metal chelator38,150 disease

Initiation-inhibiting agent209 Parkinson’s disease215

Obesity, diabetes60

Chapter 1

Trolox® ONOO–28, Inhibitor of lipid Cancer161,169 ROO•28,159,160 peroxidation138 and DNA fragmentation161

Uric acid ONOO–, OH•, Iron and copper chelator40 Parkinson’s ROO•, disease216,217 HOCl28,39,40, Huntington’s O2•– 41 disease218

1-3 Oxidative stress-related diseases

It is now well-established that oxidative stress is involved in many diseases. It is obvious that free radical-induced oxidation plays a key role in the propagation of cell injury leading to neuropathology219 and that the mechanism of carcinogenesis is initiated by oxidative DNA lesions220, however it is unclear if oxidative stress is the initiating event responsible for the disorders. Oxidative lesions on biomolecules such as DNA, proteins, lipids, and carbohydrates are involved in cancer220,221, diabetes22,182,222,223, cardiovascular diseases177,224, Down’s syndrome225,226, and rheumatoid arthritis227 among other diseases, and in ageing228-230. Each disease is associated with specific biomarkers4,231. In addition to the low level of antioxidant defenses, the high content of polyunsaturated fatty acids, the presence of metal ions (Fe, Cu, Zn…), and the high oxygen consumption make the brain particularly susceptible to oxidation. The oxidative brain damage are relevant in neurodegenerative disorders including Alzheimer’s and Parkinson’s diseases, and amyotrophic lateral sclerosis10,215,232-234. The two most common neurodegenerative disorders are further described hereinafter.

1-3.1 Alzheimer’s disease

Alzheimer’s disease (AD) is the most common form of dementia, it affects 35.6 million of people today235. This progressive and fatal brain disorder induces memory loss and affects the ability to learn, communicate, reason, and carry out

Chapter 1 daily activities. The advanced age, hypercholesterolemia, and smoking are among the risk factors for AD10. The three main hallmarks of the complex disease- related cascade are senile plaques, neurofibrillary tangles (NFT), and loss of connection among neurons. Inflammation, severe tissue shrinkage, and neuronal death are secondary hallmarks. Oxidative biomarkers, such as acrolein, 4- hydroxynonenal (HNE), protein carbonyls, and 8-hydroxyguanosine (8-OH-G) have been found in brain of AD’s patients that emphasizes oxidative stress as a promoting agent236-238. Two hypotheses have been put forward as prime trigger of Alzheimer’s disease; they are the amyloid cascade and the mitochondrial cascade since mitochondrial dysfunctions have been reported as a key role in each stage of the disorders. Oxidative stress is involved in both of them239,240. The senile plaques are made up of the aggregation of amyloid β-peptides (Aβ), which is produced from the cleavage of amyloid β precursor protein (APP).

The most common Aβ isoforms are Aβ1−40 and Aβ1-42 depending on their length. β- and γ-secretase are responsible of the cleavage using two proteins, which are presenilin 1 and presenilin 2; they are a part of the γ-secretase protease complex. The senile plaques are formed in response to oxidative stress since the Aβ peptide has antioxidant properties241,242. However, at high concentration, the peptide has

2+ + pro-oxidant abilities; it reduces Cu to Cu that leads to the H2O2 production followed by hydroxyl radical formation243,244. In the early-onset AD, induced by a genetic predisposition (5-10% of AD’s sufferers), mutations of the gene encoding

APP and both presenilin lead to the overproduction of Aβ1-42, which is the major constituent of the senile plaques. The possession of the apolipoprotein E (APOE) e4 allele is a further genetic risk factor for the familial AD since it is the least effective at exhibiting antioxidant properties while the other APOE alleles showed potent antioxidant activity245. The formation of neurofibrillary tangles (NFT) within neurons is caused by the hyperphosphorylation of the microtubule-associated protein tau. The protein is involved in microtubule assembly and stabilization. The increased activities of protein kinases and the decreased activities of protein phosphatases lead to the tau hyperphosphorylation246,247. Oxidative damage precede NFT formation as it is the case with senile plaques226,238.

Chapter 1

Cholinergic neurotransmission is the third main physiological function affected in Alzheimer’s disease. The cognitive impairments are apparently the consequence of a deficit in the neurotransmitter acetylcholine (AChE). In health body, acetylcholinesterase catalyzes the hydrolysis of acetylcholine. The therapeutic strategy based on the inhibition of AChE is the most useful and developed path to delay the neurodegenerative disorder. The enzyme also promotes the Aβ aggregation248,249. Four AChE inhibitors are commonly prescribed so far; they are Donepezil250 (1996), Rivastigmine251 (2000), Galantamine252 (2001), and memantine253 (2003). The first approved AChE inhibitor was Tacrine in 1993 but it is currently rarely prescribed due to its side effects; it was even removed from the French market in 2004. Some promising drugs are resveratrol, which is currently in phase III clinic trial200,254, melatonin200, retinoids255, heparine256, and lipid-lowering drugs257.

1-3.2 Parkinson’s disease

Parkinson’s disease (PD) is the second most common neurodegenerative disorder, which affects about 6 million worldwide. The main symptoms of the aging-related disease are uncontrolled tremors, the rigidity of muscles, and slowness of movements, called bradykinesia. Memory problems, hallucinations, and anxiety are among the secondary symptoms of PD. The symptoms are caused by the selective degeneration of dopaminergic neurons in the substantia niagra pars compacta, which are necessary for motor functions, and by the formation of protein aggregates called Lewy bodies in surviving neurons made up of α- synuclein and ubiquitin258. The familial PD (5%) has been linked with mutations of gene encoding α-synuclein, parkin, and ubiquitin carboxyterminal hydrolase L1259,260. The role of oxidative stress as cause or consequence of PD is still unclear but its involvement is obvious since biomarkers of oxidative damage have been revealed including HNE, 8-OH-dG, and carbonylated proteins. As it has been reported for AD, mitochondrial dysfunctions have a prominent role in PD, including gene alterations261,262. Dopamine can auto-oxidize and be metabolized

Chapter 1 by monoamine oxidase A (MAO-A) and B (MAO-B) which lead to the formation of

•– O2 and H2O2. A deficiency in sequestering dopamine within synaptic vesicles would be a key event of the dopaminergic neuron degeneration263. Other sources of ROS include activation of phospholipases and induction of NADPH oxidase- inactivated microglia264. The depletion of GSH is a further feature of PD. In addition to the glutathione decay due to its ROS-reducing activity, the ROS overproduction induce mitochondria dysfunctions which lead to the inhibition of GRd activity25. Some current therapeutic strategies are treatments with dopamine agonists (bromocritpine, pergolide, etc), and MAO-B inhibitors, such as selegiline, and rasagiline. Several drugs are currently prescribed depending on the symptoms. The current most effective drug is levodopa (L-Dopa, 3,4- dihydroxyphenylalanine), which is a natural amino acid converted into dopamine in the brain. However, it requires to be combined with dopamine agonists, catechol-O-methyl transferase (COMT), and MAO inhibitors to prolong its activity. Furthermore, other drugs are necessary to counteract the high number of L-Dopa-induced side effects265. Other strategies are currently developed. Among them, the inhibition of α-synuclein aggregation appears promising266,267.

1-3.3 Antioxidant strategies in Alzheimer’s and Parkinson’s diseases

It is now well-established that antioxidant compounds may act as free radical scavengers in preventing and/or inhibiting oxidative damage on biomolecules and ROS-activated proteins such as kinases and secretases. The mitochondria are the main source of ROS. Due to the prominent role of mitochondrial dysfunction in AD and PD, more and more studies focus on mitochondria-targeted antioxidants to prevent and detoxify ROS overproduction262,268,269. However, besides radical scavenging activity, other properties are required including a selective deliverance into the oxidative damage-affected mitochondria. A promising antioxidant drug, named MitoQ is

Chapter 1 currently in phase II clinical trials for PD270. In addition, metal ions including iron, copper, zinc, and aluminum are involved in some pathological features of both diseases232,271. In Alzheimer’s disease, the metals accelerate the Aβ aggregation and promote the tau-dependant pathology272. A restoration of metal ion homeostasis may not only reverse the amyloid aggregation but also dissolve the amyloid plaques involving a delay in cognitive impairments273. The iron level in PD, significantly higher in substantia niagra relative to other brain regions,

• may readily be involved in the Fenton reaction which converts H2O2 into OH ; hence a trigger of lipid peroxidation. It has been proven that iron-related oxidative stress also promotes the α-synuclein aggregation274. The metal- targeting strategy consists in chelating the metal overload which may prevent or inhibit the protein aggregation and other metal-mediated processes. Some metal chelators are currently used such as desferrioxamine (DFO), and D- penicillamine. Other chelators combining radical scavenging activity have been designed275,276

1-4 Multifunctional drugs

The development of multifunctional drugs is of great interest in the treatment of complexe diseases. They allow the interaction with several disease- related targets by reducing the drug-drug interaction and by ameliorating the treatment regimen through reduced drug-related side effects277,278. A multifunctional drug may be a design of hybrid molecules or an entity exhibiting several therapeutic abilities. Hybrid molecules are synthesized by coupling two or more pharmacophores, each one with a specific biological activity. Recently, a multifunctional drug for the therapy of Alzheimer’s disease has been designed made up of three pharmacophores, the properties of which are the AChE inhibition, antioxidant properties, and the penetration into the central nervous system198. Due to the involvement of oxidative stress in a wide disease spectrum, a hybrid molecule bearing an antioxidant pharmacophore appears a promising drug to delay or hopefully cure the disorder. Many multifunctional drugs with an

Chapter 1 antioxidant moiety have already been developed and exhibited an activity in several diseases like cancer279,280, neurodegenerative199,281-286, and cardiovascular diseases277,278,287,288. However, hybrid molecules require sometimes a long and difficult synthesis and a specific knowledge of the behavior in vivo of these new structures. An entity accepted as a multifunctional drug has to reveal effectiveness against several targets. When considering Alzheimer’s and Parkinson’s diseases some targets are common to both diseases, i.e. acetylcholinesterase, monoamine oxidase B, and oxidative stress. Indeed, these neurodegenerative disorders are due to abnormalities in cholinergic functions. However, AChE inhibitors (iAChE) are necessary in AD due to the loss of cholinergic neurons while they have to be avoided in PD since an increased activity of cholinergic neurons is one of the pathological features. Besides its harmful contribution in PD, MAO-B is a target to inhibit in both diseases. The MAO-B-catalyzed oxidation of dopamine yields the formation of H2O2 which could be then involved in other deleterious ROS- mediated processes. Furthermore, the MAO-B inhibition may induce antidepressant activity due to the low level of dopamine, serotonin, and noradrenalin in the central nervous system. The large contribution of oxidative stress at each stage of AD and PD makes the antioxidants have a decisive role in the therapy. Inhibitors of other targets, specific to each disease are also relevant for the therapy. These targets are the Aβ aggregation and the tau protein hyperphosphorylation for AD and the α-synuclein aggregation and the COMT for PD (Figure 1-2).

Chapter 1

Figure 1-2: Experimental assay procedure leading to multifunctional compounds for AD and PD therapy. i stands for inhibitor, iTau hyperP is inhibitor of tau hyperphosphorylation.

Table 1-4 summarizes some multifunctional compounds combining antioxidant and other properties necessary to treat Alzheimer’s or Parkinson’s diseases. Resveratrol, quercetin, and caffeic acid, previously described, are also considered as multifunctional compounds.

Chapter 1

Table 1-4: Some multifunctional drugs for AD and PD therapy.

Targeted- Compound Structure Properties diseases

Antioxidant: radical scavenger, Curcumin metal chelator AD289-291 Aβ aggregation

Ladostigil AChE and MAO-B inhibitor AD, PD292 (TV3326)

Antioxidant: radical scavenger, M30 iron chelator PD293

MAO inhibitor

Antioxidant: radical scavenger

Memoquin Aβ aggregation inhibitor AD294

Antioxidant: peroxynitrite PF9601N scavenger PD295 MAO-B inhibitor

Chapter 1

1-5 Scope and aims of the thesis

Oxidative stress is involved in a wide human-related disease spectrum due to the high number of reactive oxygen and nitrogen species which induce damage on several biomolecules and cellular constituents. All these oxidative events make the characterization of an antioxidant complex, and antioxidants often lack of specificity toward some ROS/RNS and biomolecules. By drawing up profiles, which describe the antioxidant properties of a compound toward radicals and in presence of biomolecules, the multifunctional drug research will be enhanced. Furthermore, new chemical entities (NCE) will readily be characterized. A specific profile will provide the antioxidant properties of NCE against a high number of radicals and the behavior toward the biomolecules. To perform high throughput screening and readily detect the potential antioxidants from a large set of compounds, parameters are necessary. They are determined thanks to a small set of compounds with four methods which evaluate the antioxidant properties toward peroxyl, DPPH•, and ABTS•– radicals. Furthermore, parameters of potency and kinetics are determined for these compounds; they allow the formation of classes for a ranking. In order to provide well-known antioxidant-related profiles, reference compounds are included in the set of compounds. These reference compounds are ascorbic acid, α-tocopherol, caffeic acid, chlorogenic acid, gallic acid, glutathione, mangiferin, mannitol, melatonin, quercetin, resveratrol, trolox®, and uric acid. After the screening and ranking parameters are defined, a larger set of compounds is screened. The properties of the emerged potential antioxidants, which are the potency and the kinetics to reduce the radicals, is further characterized and ranked into the defined classes. A clustering analysis of the antioxidant activities is performed yielding antioxidant profiles. Trolox®- derivatives are used for the application of the characterization by the antioxidant profiles. To further characterize the antioxidant power, the standard formal potential of caffeic acid, melatonin, and para-hydroquinone are assessed by cyclic voltammetry. Since a study of these compounds alone does not provide any

Chapter 1 accessible standard formal potential, the latter is estimated by the ability of the compounds to regenerate electrochemically oxidized ferrocene-derivatives. Finally, as a good knowledge of the chemical structure is useful in drug design, a study of structure-activity relationship is considered.

Chapter 1

Bibliographic references

1. Droge, W. Free radicals in the physiological control of cell function. Physiol. Rev. 2002, 82, 47-95. 2. Noctor, G. Metabolic signalling in defence and stress: the central roles of soluble redox couples. Plant Cell Environ. 2006, 29, 409-425. 3. Halliwell, B. Free radical reactions in human disease. Oxidative Stress Disease 2001, 7, 1-16. 4. Halliwell, B.; Whiteman, M. Measuring reactive species and oxidative damage in vivo and in cell culture: how should you do it and what do the results mean? Brit. J. Pharmacol. 2004, 142, 231-255. 5. Keisari, Y.; Braun, L.; Flescher, E. The oxidative burst and related phenomena in mouse macrophages elicited by different sterile inflammatory stimuli. Immunobiol. 1983, 165, 78-89. 6. Griendling, K. K.; Sorescu, D.; Lassegue, B.; Ushio-Fukai, M. Modulation of protein kinase activity and gene expression by reactive oxygen species and their role in vascular physiology and pathophysiology. Arterioscler. Thromb. Vasc. Biol. 2000, 20, 2175-2183. 7. Bergendi, L.; Benes, L.; Durackova, Z.; Ferencik, M. Chemistry, physiology and pathology of free radicals. Life Sci. 1999, 65, 1865-1874. 8. Franco, R.; Panayiotidis, M. I. Environmental toxicity, oxidative stress, human disease and the "black box" of their synergism: How much have we revealed? Mutation Res. 2009, 674, 1-2. 9. Limon-Pacheco, J.; Gonsebatt, M. E. The role of antioxidants and antioxidant- related enzymes in protective responses to environmentally induced oxidative stress. Mutation Res. 2009, 674, 137-147. 10. Migliore, L.; Coppede, F. Environmental-induced oxidative stress in neurodegenerative disorders and aging. Mutation Res. 2009, 674, 73-84. 11. Halliwell, B. Biochemistry of oxidative stress. Biochem. Soc. Trans. 2007, 35, 1147-1150. 12. Franco, R.; Cidlowski, J. A. Apoptosis and glutathione: beyond an antioxidant. Cell Death. Differ. 2009, 16, 1303-1314. 13. Tan, D. X.; Manchester, L. C.; Sainz, R.; Mayo, J. C.; Alvares, F. L.; Reiter, R. J. Antioxidant strategies in protection against neurodegenerative disorders. Exp. Opin. Ther. Patents 2003, 13, 1513-1543. 14. Lack, B.; Daya, S.; Nyokong, T. Interaction of serotonin and melatonin with sodium, potassium, calcium, lithium and aluminium. J. Pineal Res. 2001, 31, 102-108. 15. Limson, J.; Nyokong, T.; Daya, S. The interaction of melatonin and its precursors with aluminium, cadmium, copper, iron, lead, and zinc: An adsorptive voltammetric study. J. Pineal Res. 1998, 24, 15-21. 16. Gulcin, I.; Buyukokuroglu, M. E.; Kufrevioglu, O. I. Metal chelating and hydrogen peroxide scavenging effects of melatonin. J. Pineal Res. 2003, 34, 278-281. 17. Traber, M. G.; Atkinson, J. Vitamin E, antioxidant and nothing more. Free Rad. Biol. Med. 2007, 43, 4-15. 18. Simic, M. G.; Jovanovic, S. V. Antioxidation mechanisms of uric-scid. J. Am. Chem. Soc. 1989, 111, 5778-5782. 19. Halliwell, B.; Aeschbach, R.; Löliger, J.; Aruoma, O. I. The characteization of antioxidants. Fd Chem. Toxic. 1995, 33, 601-617. 20. Menvielle-Bourg, F. J. La superoxyde dismutase, puissant antioxydant naturel, d+®sormais disponible par voie orale. Phytothérapie 2005, 3, 118-121.

Chapter 1

21. Goth, L.; Lenkey, A.; Bigler, W. N. Blood catalase deficiency and diabetes in Hungary. Diabetes Care 2001, 24, 1839-1840. 22. Goth, L. Catalase deficiency and type 2 diabetes. Diabetes Care 2008, 31, e93. 23. Franco, R.; Schoneveld, O. J.; Pappa, A.; Panayiotidis, M. I. The central role of glutathione in the pathophysiology of human diseases. Arch. Physiol. Biochem. 2007, 113, 234-258. 24. Forman, H. J.; Zhang, H.; Rinna, A. Glutathione: overview of its protective roles, measurement, and biosynthesis. Mol. Aspects of Med. 2002, 30, 1-12. 25. Zeevalk, G. D.; Razmpour, R.; Bernard, L. P. Glutathione and Parkinson's disease: Is this the elephant in the room? Biomed. Pharmacother. 2004, 62, 236-249. 26. Martin, H. L.; Teismann, P. Glutathione--a review on its role and significance in Parkinson's disease. FASEB J. 2009, 23, 3263-3272. 27. Winterbourn, C. C.; Brennan, S. O. Characterization of the oxidation products of the reaction between reduced glutathione and hypochlorous acid. Biochem. J. 1997, 326, 87-92. 28. Regoli, F.; Winston, G. W. Quantification of total oxidant scavenging capacity of antioxidants for peroxynitrite, peroxyl radicals, and hydroxyl radicals. Toxicol. Appl. Pharmacol. 1999, 156, 96-105. 29. Sagrista, M. L.; Garcia, A. F.; De Madariaga, M. A.; Mora, M. Antioxidant and pro-oxidant effect of the thiolic compounds N-acetyl-L-cysteine and glutathione against free radical-induced lipid peroxidation. Free Rad. Res. 2002, 36, 329-340. 30. Brigelius-Flohe, R. Tissue-specific functions of individual glutathione peroxidases. Free Rad. Biol. Med. 1999, 27, 951-965. 31. Ralat, L. A.; Manevich, Y.; Fisher, A. B.; Colman, R. F. Direct evidence for the formation of a complex between 1-cysteine peroxiredoxin and glutathione S-transferase π with activity changes in both enzymes. Biochemistry 2005, 45, 360-372. 32. Alary, J.; Guθraud; Cravedi, J. P. Fate of 4-hydroxynonenal in vivo: disposition and metabolic pathways. Mol. Aspects of Med. 2008, 24, 177-187. 33. Dawson, J.; Walters, M. Uric acid and xanthine oxidase: future therapeutic targets in the prevention of cardiovascular disease? Brit. J. Clin. Pharmacol. 2006, 62, 633-644. 34. Feig, D. I.; Kang, D. H.; Johnson, R. J. Uric acid and cardiovascular risk. New Engl. J. Medicine 2008, 359, 1811-1821. 35. Dehghan, A.; van Hoek, M.; Sijbrands, E. J. G.; Hofman, A.; Witteman, J. C. M. High serum uric acid as a novel risk factor for type 2 diabetes. Diabetes Care 2008, 31, 361-362. 36. Strasak, A. M.; Rapp, K.; Hilbe, W.; Oberaigner, W.; Ruttmann, E.; Concin, H.; Diem, G.; Pfeiffer, K. P.; Ulmer, H. The role of serum uric acid as an antioxidant protecting against cancer: prospective study in more than 28 000 older Austrian women. Ann. Oncol. 2007, 18, 1893-1897. 37. Sotgiu, S.; Pugliatti, M.; Sanna, A.; Sotgiu, A.; Fois, M. L.; Arru, G.; Rosati, G. Serum uric acid and multiple sclerosis. Neurol. Sci. 2002, 23, 183-188. 38. Belguendouz, L.; Fremont, L.; Linard, A. Resveratrol inhibits metal ion- dependent and independent peroxidation of porcine low-density lipoproteins. Biochem. Pharmacol. 1997, 53, 1347-1355. 39. Burkhardt, S.; Reiter, R. J.; Tan, D. X.; Hardeland, R.; Cabrera, J.; Karbownik, M. DNA oxidatively damaged by chromium(III) and H2O2 is protected by the antioxidants melatonin, N-1-acetyl-N-2-formyl-5-methoxykynuramine, resveratrol and uric acid. Int. J. Biochem. Cell. Biol. 2001, 33, 775-783.

Chapter 1

40. Allen, R. N.; Shukla, M. K.; Leszczynski, J. A theoretical study of the structure and properties of uric acid: A potent antioxidant. Int. J. Quant. Chem. 2004, 100, 801-809. 41. Shukla, M. K.; Mishra, P. C. Electronic structures and spectra of two antioxidants, uric acid and ascorbic acid. J. Mol. Struct. 1996, 377, 247- 259. 42. Bagnati, M.; Perugini, C.; Cau, C.; Bordone, R.; Albano, E.; Bellomo, G. When and why a water-soluble antioxidant becomes pro-oxidant during copper- induced low-density lipoprotein oxidation: a study using uric acid. Biochem. J. 1999, 340, 143-152. 43. Arendt, J. Importance and relevance of melatonin to human biological rhythms. J. Neuroendocrinol. 2003, 15, 427-431. 44. Reiter, R. J.; Paredes, S. D.; Manchester, L. C.; Tan, D. X. Reducing oxidative/nitrosative stress: a newly-discovered genre for melatonin. Crit. Rev. Biochem. Mol. Biol. 2009, 44, 175-200. 45. Reiter, R. J.; Tan, D. X.; Manchester, L. C.; Qi, W. B. Biochemical reactivity of melatonin with reactive oxygen and nitrogen species - A review of the evidence. Cell Biochem. Biophys. 2001, 34, 237-256. 46. Tan, D. X.; Chen, L. D.; Poeggeler, B.; Manchester, L. C.; Reiter, R. J. Melatonin: a potent, endogenous hydroxyl radical scavenger. Endocrine Journal 1993, 1, 57-60. 47. Poeggeler, B.; Saarela, S.; Reiter, R. J.; Tan, D. X.; Chen, L. D.; Manchester, L. C.; Barlow-Walden, L. R. Melatonin, a highly potent endogenous radical scavenger and electron donor: new aspects of the oxidation chemistry of this indole accessed in vitro. Ann. N. Y. Acad. Sci. 1994, 738, 419-420. 48. Teixeira, A.; Morfim, M. P.; de Cordova, C. A. S.; Charao, C. C. T.; de Lima, V. R.; Creczynski-Pasa, T. B. Melatonin protects against pro-oxidant enzymes and reduces lipid peroxidation in distinct membranes induced by the hydroxyl and ascorbyl radicals and by peroxynitrite. J. Pineal Res. 2003, 35, 262-268. 49. Poeggeler, B.; Reiter, R. J.; Hardeland, R.; Tan, D. X.; Barlow-Walden, L. R. Melatonin and structurally-related, endogenous indoles act as potent electron donors and radical scavengers in vitro. Redox Report 1996, 2, 179- 184. 50. Zang, L. Y.; Cosma, G.; Gardner, H.; Vallyathan, V. Scavenging of reactive oxygen species by melatonin. Biochim. Biophys. Acta 1998, 1425, 469-477. 51. Tan, D. X.; Manchester, L. C.; Reiter, R. J.; Plummer, B. F.; Limson, J.; Weintraub, S. T.; Qi, W. B. Melatonin directly scavenges hydrogen peroxide: A potentially new metabolic pathway of melatonin biotransformation. Free Rad. Biol. Med. 2000, 29, 1177-1185. 52. Cuzzocrea, S.; Costantino, G.; Caputi, A. P. Protective effect of melatonin on cellular energy depletion mediated by peroxynitrite and poly (ADP-ribose) synthetase activation in a non-septic shock model induced by zymosan in the rat. J. Pineal Res. 1998, 25, 78-85. 53. Blanchard, B.; Pompon, D.; Ducrocq, C. Nitrosation of melatonin by nitric oxide and peroxynitrite. J. Pineal Res. 2000, 29, 184-192. 54. Pozo, D.; Reiter, R. J.; Calvo, J. R.; Guerrero, J. M. Inhibition of cerebellar nitric oxide synthase and cyclic GMP production by melatonin via complex formation with calmodulin. J. Cell. Biochem. 1997, 65, 430-442. 55. Crespo, E.; Macias, M.; Pozo, D.; Escames, G.; Martin, M.; Vives, F.; Guerrero, J. M.; cuna-Castroviejo, D. Melatonin inhibits expression of the inducible NO synthase II in liver and lung and prevents endotoxemia in

Chapter 1

lipopolysaccharide-induced multiple organ dysfunction syndrome in rats. FASEB J. 1999, 13, 1537-1546. 56. Antolin, I.; Rodriguez, C.; Sainz, R. M.; Mayo, J. C.; Uria, H.; Kotler, M. I.; RodriguezColunga, M. J.; Tolivia, D.; MenendezPelaez, A. Neurohormone melatonin prevents cell damage: Effect on gene expression for antioxidant enzymes. FASEB J. 1996, 10, 882-890. 57. Kotler, M.; Rodriguez, C.; Sainz, R. M.; Antolin, I.; Menendez-Pelaez, A. Melatonin increases gene expression for antioxidant enzymes in rat brain cortex. J. Pineal Res. 1998, 24, 83-89. 58. Reiter, R. J.; Tan, D. X.; Osuna, C.; Gitto, E. Actions of melatonin in the reduction of oxidative stress - A review. J. Biomed. Sci. 2000, 7, 444-458. 59. Pablos, M. I.; Reiter, R. J.; Ortiz, G. G.; Guerrero, J. M.; Agapito, M. T.; Chuang, J. I.; Sewerynek, E. Rhythms of glutathione peroxidase and glutathione reductase in brain of chick and their inhibition by light. Neurochem. Int. 1998, 32, 69-75. 60. Szkudelska, K.; Szkudelski, T. Resveratrol, obesity and diabetes. Eur. J. Pharmacol. 2010, 635, 1-8. 61. Tan, D. X.; Manchester, L. C.; Burkhardt, S.; Sainz, R. M.; Mayo, J. C.; Kohen, R.; Shohami, E.; Huo, Y. S.; Hardeland, R.; Reiter, R. J. N-1-acetyl-N-2- formyl-5-methoxykynuramine, a biogenic amine and melatonin metabolite, functions as a potent antioxidant. FASEB J. 2001, 15, 2294-+. 62. Tan, D. X.; Hardeland, R.; Manchester, L. C.; Poeggeler, B.; Lopez-Burillo, S.; Mayo, J. C.; Sainz, R. M.; Reiter, R. J. Mechanistic and comparative studies of melatonin and classic antioxidants in terms of their interactions with the ABTS cation radical. J. Pineal Res. 2003, 34, 249-259. 63. Tan, D. X.; Manchester, L. C.; Terron, M. P.; Flores, L. J.; Reiter, R. J. One molecule, many derivatives: a never-ending interaction of melatonin with reactive oxygen and nitrogen species? J. Pineal Res. 2007, 41, 28-42. 64. Maharaj, D. S.; Glass, B. D.; Daya, S. Melatonin: New places in therapy. Biosci. Rep. 2007, 27, 299-320. 65. Davey, M. W.; Van Montagu, M.; Inze, D.; Sanmartin, M.; Kanellis, A.; Smirnoff, N.; Benzie, I. J. J.; Strain, J. J.; Favell, D.; Fletcher, J. Plant L-ascorbic acid: chemistry, function, metabolism, bioavailability and effects of processing. J. Sci. Food Agric. 2000, 80, 825-860. 66. Buettner, G. R.; Jurkiewicz, B. A. Catalytic metals, ascorbate and free radicals: Combinations to avoid. Radiat. Res. 1996, 145, 532-541. 67. Carr, A.; Frei, B. Does vitamin C act as a pro-oxidant under physiological conditions? FASEB J. 1999, 13, 1007-1024. 68. Poljsak, B.; Raspor, P. The antioxidant and pro-oxidant activity of vitamin C and trolox in vitro: a comparative study. J. Appl. Toxicol. 2008, 28, 183-188. 69. Soares, D. G.; Andreazza, A. C.; Salvador, M. Sequestering ability of butylated hydroxytoluene, propyl gallate, resveratrol, and vitamins C and E against ABTS, DPPH, and hydroxyl free radicals in chemical and biological systems. J. Agric. Food Chem. 2003, 51, 1077-1080. 70. Makhotkina, O.; Kilmartin, P. A. Uncovering the influence of antioxidants on polyphenol oxidation in wines using an electrochemical method: Cyclic voltammetry. J. Electroanal. Chem. 2009, 633, 165-174. 71. Sawai, Y.; Moon, J. H. NMR snalytical spproach to vlarify the molecular mechanisms of the sntioxidative and tadical-dcavenging sctivities of sntioxidants in tea using 1,1-Diphenyl-2-picrylhydrazyl. J. Agric. Food Chem. 2000, 48, 6247-6253.

Chapter 1

72. Zaidi, S. M. K.; Banu, N. Antioxidant potential of vitamins A, E and C in modulating oxidative stress in rat brain. Clin. Chim. Acta 2004, 340, 229- 233. 73. Born, M. Redox studies of some natural and synthetic antioxidants. PhD Thesis; University of Lausanne: Lausanne, 1997. 74. Dangles, O.; Fargeix, G.; Dufour, C. One-electron oxidation of quercetin and quercetin derivatives in protic and non protic media. J. Chem. Soc. , Perkin Trans. 2 1999, 1387-1395. 75. Bowry, V. W.; Stocker, R. Tocopherol-mediated peroxidation. The prooxidant effect of vitamin E on the radical-initiated oxidation of human low-protein lipoprotein. J. Am. Chem. Soc. 1993, 115, 6029-6044. 76. Carr, A. C.; Mccall, M. R.; Frei, B. Oxidation of LDL by myeloperoxidase and reactive nitrogen species - Reaction pathways and antioxidant protection. Arterioscler. Thromb. Vasc. Biol. 2000, 20, 1716-1723. 77. Burkitt, M. J.; Milne, L. Hydroxyl radical formation from Cu(II)-Trolox mixtures: Insights into the pro-oxidant properties of alpha-tocopherol. FEBS Lett. 1996, 379, 51-54. 78. Miller, E. R.; Pastor-Barriuso, R.; Dalal, D.; Riemersma, R. A.; Appel, L. J.; Guallar, E. Meta-analysis: high-dosage vitamin E supplementation may increase all-cause mortality. Ann. Intern. Med. 2005, 142, 37-46. 79. Winterbone, M.; Sampson, M.; Saha, S.; Hughes, J.; Hughes, D. Pro-oxidant effect of alpha-tocopherol in patients with Type 2 Diabetes after an oral glucose tolerance test - a randomised controlled trial. Cardiovascul. Diabetol. 2007, 6, 8. 80. Shahidi, F.; Chandrasekara, A. Hydroxycinnamates and their in vitro and in vivo antioxidant activities. Phytochem. Rev. 2010, 9, 147-170. 81. Orban, Z.; Mitsiades, N.; Burke, T. R.; Tsokos, M.; Chrousos, G. P. Caffeic acid phenethyl ester induces leukocyte apoptosis, modulates nuclear factor- kappa B and suppresses acute inflammation. Neuroimmunomodul. 2000, 7, 99-105. 82. Lee, W. J.; Zhu, B. T. Inhibition of DNA methylation by caffeic acid and chlorogenic acid, two common catechol-containing coffee polyphenols. Carcinogenesis 2006, 27, 269-277. 83. Nardini, M.; Leonardi, F.; Scaccini, C.; Virgili, F. Modulation of ceramide-induced NF-kappa B binding activity and apoptotic response by caffeic acid in U937 cells: comparison with other antioxidants. Free Rad. Biol. Med. 2001, 30, 722-733. 84. Tapiero, H.; Tew, K. D.; Ba, G. N.; Mathe, G. Polyphenols: do they play a role in the prevention of human pathologies? Biomed. Pharmacother. 2002, 56, 200-207. 85. Andreasen, M. F.; Landbo, A. K.; Christensen, L. P.; Hansen, A.; Meyer, A. S. Antioxidant effects of phenolic rye (Secale cereale L.) extracts, monomeric hydroxycinnamates, and ferulic acid dehydrodimers on human low-density lipoproteins. J. Agric. Food Chem. 2001, 49, 4090-4096. 86. Gülcin, I. Antioxidant activity of caffeic acid (3,4-dihydroxycinnamic acid). Toxicology 2006, 217, 213-220. 87. Cheng, J. C.; Dai, F.; Zhou, B.; Yang, L.; Liu, Z. L. Antioxidant activity of hydroxycinnamic acid derivatives in human low density lipoprotein: Mechanism and structure-activity relationship. Food Chem. 2007, 104, 132-139. 88. Pari, L.; Prasath, A. Efficacy of caffeic acid in preventing nickel induced oxidative damage in liver of rats. Chem. Biol. Interactions 2008, 173, 77-83.

Chapter 1

89. Nardini, M.; D'Aquino, M.; Tomassi, G.; Gentili, V.; Di Felice, M.; Scaccini, C. Inhibition of human low-density lipoprotein oxidation by caffeic acid and other hydroxycinnamic acid derivatives. Free Rad. Biol. Med. 1995, 19, 541-552. 90. Kono, Y.; Kobayashi, K.; Tagawa, S.; Adachi, K.; Ueda, A.; Sawa, Y.; Shibata, H. Antioxidant activity of polyphenolics in diets - Rate constants of reactions of chlorogenic acid and caffeic acid with reactive species of oxygen and nitrogen. Biochim. Biophys. Acta 1997, 1335, 335-342. 91. Kweon, M. H.; Hwang, H. J.; Sung, H. C. Identification and antioxidant activity of novel chlorogenic acid derivatives from bamboo (Phyllostachys edulis). J. Agric. Food Chem. 2001, 49, 4646-4655. 92. Mansouri, A.; Makris, D. P.; Kefalas, P. Determination of hydrogen peroxide scavenging activity of cinnamic and benzoic acids employing a highly sensitive peroxyoxalate chemiluminescence-based assay: Structure- activity relationships. J. Pharm. Biomed. Anal. 2005, 39, 22-26. 93. Saija, A. Ferulic and caffeic acids as potential protective agents against photooxidative skin damage (vol 79, pg 476, 1999). J. Sci. Food Agric. 1999, 79, 1319. 94. Saija, A.; Tomaino, A.; Trombetta, D.; De Pasquale, A.; Uccella, N.; Barbuzzi, T.; Paolino, D.; Bonina, F. In vitro and in vivo evaluation of caffeic and ferulic acids as topical photoprotective agents. Int. J. Pharm. 2000, 199, 39-47. 95. Ketsawatsakul, U.; Whiteman, M.; Halliwell, B. A reevaluation of the peroxynitrite scavenging activity of some dietary phenolics. Biochem. Biophys. Res. Commun. 2000, 279, 692-699. 96. Olthof, M. R.; Hollman, P. C. H.; Katan, M. B. Chlorogenic acid and caffeic acid are absorbed in humans. J. Nutr. 2001, 131, 66-71. 97. Kuenzig, W.; Chau, J.; Norkus, E.; Holowaschenko, H.; Newmark, H.; Mergens, W.; Conney, A. H. Caffeic and ferulic acid as blockers of nitrosamine formation. Carcinogenesis 1984, 5, 309-313. 98. Wattenberg, L. W.; Coccia, J. B.; Lam, L. K. T. Inhibitory effects of phenolic- compounds on benzo(A)pyrene-induced neoplasia. Cancer Res. 1980, 40, 2820-2823. 99. Tanaka, T.; Kojima, T.; Kawamori, T.; Wang, A.; Suzui, M.; Okamoto, K.; Mori, H. Inhibition of 4-nitroquinoline-1-oxide-induced rat tongue carcinogenesis by the naturally occurring plant phenolics caffeic, ellagic, chlorogenic and ferulic acids. Carcinogenesis 1993, 14, 1321-1325. 100. Grace, S. C.; Salgo, M. G.; Pryor, W. A. Scavenging of peroxynitrite by a phenolic/peroxidase system prevents oxidative damage to DNA. FEBS Letters 1998, 426, 24-28. 101. Heinonen, I. M.; Meyer, A. S.; Frankel, E. N. Antioxidant activity of berry phenolics on human low-density lipoprotein and liposome oxidation. J. Agric. Food Chem. 1998, 46, 4107-4112. 102. Marinova, E. M.; Toneva, A.; Yanishlieva, N. Comparison of the antioxidative properties of caffeic and chlorogenic acids. Food Chem. 2009, 114, 1498- 1502. 103. de Sotillo, D. V. R.; Hadley, M. Chlorogenic acid modifies plasma and liver concentrations of: cholesterol, triacylglycerol, and minerals in (fa/fa) Zucker rats. J. Nutr. Biochem. 2002, 13, 717-726. 104. Hseu, Y. C.; Chang, W. H.; Chen, C. S.; Liao, J. W.; Huang, C. J.; Lu, F. J.; Chia, Y. C.; Hsu, H. K.; Wu, J. J.; Yang, H. L. Antioxidant activities of Toona Sinensis leaves extracts using different antioxidant models. Food Chem. Toxicol. 2008, 46, 105-114.

Chapter 1

105. Adom, K. K.; Liu, R. H. Rapid peroxyl radical scavenging capacity (PSC) assay for assessing both hydrophilic and lipophilic antioxidants. J. Agric. Food Chem. 2005, 53, 6572-6580. 106. Roginsky, V. Chain-breaking antioxidant activity of natural polyphenols as determined during the chain oxidation of methyl linoleate in Triton X-100 micelles. Arch. Biochem. Biophys. 2003, 414, 261-270. 107. Lu, Z.; Nie, G.; Belton, P. S.; Tang, H.; Zhao, B. Structure-activity relationship analysis of antioxidant ability and neuroprotective effect of gallic acid derivatives. Neurochem. Int. 2006, 48, 263-274. 108. Li, L.; Ng, T. B.; Gao, W.; Li, W.; Fu, M.; Niu, S. M.; Zhao, L.; Chen, R. R.; Liu, F. Antioxidant activity of gallic acid from rose flowers in senescence accelerated mice. Life Sci. 2005, 77, 230-240. 109. Ximenes, V. F.; Lopes, M. G.; Petronio, M. S.; Regasini, L. O.; Siqueira Silva, D. H.; da Fonseca, L. M. Inhibitory effect of gallic acid and iIts esters on 2,2'- Azobis(2-amidinopropane)hydrochloride (AAPH)-induced hemolysis and depletion of intracellular glutathione in erythrocytes. J. Agric. Food Chem. 2010, 58, 5355-5362. 110. Wacher, V. J.; Benet, L. Z. Use of gallic acid esters to increase bioavailability of orally administered pharmaceutical compounds; 2001. 111. Serrano, A.; Palacios, C.; Roy, G.; Cespon, C.; Villar, M. L.; Nocito, M.; Gonzalez- Porque, P. Derivatives of gallic acid induce apoptosis in tumoral cell lines and inhibit lymphocyte proliferation. Arch. Biochem. Biophys. 1998, 350, 49-54. 112. Kasture, V. S.; Katti, S. A.; Mahajan, D.; Wagh, R.; Mohan, M.; Kasture, S. B. Antioxidant and antiparkinson activity of gallic acid derivatives. Pharmacologyonline 2009, 1, 385-395. 113. Rangel, L. P.; Fritzen, M.; Yunes, R. A.; Leal, P. C.; Creczynski-Pasa, T. B.; Ferreira-Pereira, A. Inhibitory effects of gallic acid ester derivatives on Saccharomyces cerevisiae multidrug resistance protein Pdr5p. Fems Yeast Res. 2010, 10, 244-251. 114. Carvalho, A. C. S.; Guedes, M. M.; de Souza, A. L.; Trevisan, M. T. S.; Lima, A. F.; Santos, F. A.; Rao, V. S. N. Gastroprotective effect of mangiferin, a xanthonoid from Mangifera indica, against gastric injury induced by ethanol and indomethacin in rodents. Planta Med. 2007, 73, 1372-1376. 115. Miura, T.; Ichiki, H.; Hashimoto, I.; Iwamoto, N.; Kato, M.; Kubo, M.; Ishihara, E.; Komatsu, Y.; Okada, M.; Ishida, T.; Tanigawa, K. Antidiabetic activity of a xanthone compound, mangiferin. Phytomedicine 2001, 8, 85-87. 116. Dar, A.; Faizi, S.; Naqvi, S.; Roome, T.; Zikr-ur-Rehman, S.; Ali, M.; Firdous, S.; Moin, S. T. Analgesic and antioxidant activity of mangiferin and its derivatives: the structure activity relationship. Biol. Pharm. Bull. 2005, 28, 596-600. 117. Garrido, G.; Gonzalez, D.; Lemus, Y.; Garcia, D.; Lodeiro, L.; Quintero, G.; Delporte, C.; Nunez-Selles, A. J.; Delgado, R. In vivo and in vitro anti- inflammatory activity of Mangifera indica L. extract (VIMANG«). Pharmacol. Res. 2004, 50, 143-149. 118. Pardo-Andreu, G. L.; Sanchez-Baldoquin, C.; vila-Gonzalez, R.; Delgado, R.; Naal, Z.; Curti, C. Fe(III) improves antioxidant and cytoprotecting activities of mangiferin. Eur. J. Pharmacol. 2006, 547, 31-36. 119. Stoilova, I.; Jirovetz, L.; Stoyanova, A. Antioxidant activity of the polyphenol mangiferin. Electro. J. Environ. Agric. Food Chem. 2008, 13, 2706-2716.

Chapter 1

120. Mishra, B.; Priyadarsini, K. I.; Sudheerkumar, M.; Unnikrishhnan, M. K.; Mohan, H. Pulse radiolysis studies of mangiferin: s C-glycosyl xanthone isolated from Mangifera indica. Radiat. Phys. Chem. 2006, 75, 70-77. 121. Andreu, G. P.; Delgado, R.; Velho, J. A.; Curti, C.; Vercesi, A. E. Iron complexing activity of mangiferin, a naturally occurring glucosylxanthone, inhibits mitochondrial lipid peroxidation induced by Fe2+-citrate. Eur. J. Pharmacol. 2005, 513, 47-55. 122. Andreu, G. L. P.; Delgado, R.; Velho, J. A.; Curti, C.; Vercesi, A. E. Mangiferin, a natural occurring glucosyl xanthone, increases susceptibility of rat liver mitochondria to calcium-induced permeability transition. Arch. Biochem. Biophys. 2005, 439, 184-193. 123. Geleijnse, J. M.; Launer, L. J.; van der Kuip, D. A. M.; Hofman, A.; Witteman, J. C. M. Inverse association of tea and flavonoid intakes with incident myocardial infarction: the Rotterdam Study. Am. J. Clin. Nutr. 2002, 75, 880-886. 124. Hertog, M. G. L.; Feskens, E. J. M.; Hollman, P. C. H.; Katan, M. B.; Kromhout, D. Dietary Antioxidant Flavonoids and Risk of Coronary Heart-Disease - the Zutphen Elderly Study. Lancet 1993, 342, 1007-1011. 125. Knekt, P.; Kumpulainen, J.; Jarvinen, R.; Rissanen, H.; Heliovaara, M.; Reunanen, A.; Hakulinen, T.; Aromaa, A. Flavonoid intake and risk of chronic diseases. Am. J. Clin. Nutr. 2002, 76, 560-568. 126. Manach, C.; Williamson, G.; Morand, C.; Scalbert, A.; Remesy, C. Bioavailability and bioefficacy of polyphenols in humans. I. Review of 97 bioavailability studies. Am. J. Clin. Nutr. 2005, 81, 230S-242S. 127. Egert, S.; Wolffram, S.; Bosy-Westphal, A.; Boesch-Saadatmandi, C.; Wagner, A. E.; Frank, J.; Rimbach, G.; Mueller, M. J. Daily quercetin supplementation dose-dependently increases plasma quercetin concentrations in healthy humans. J. Nutr. 2008, 138, 1615-1621. 128. Erlund, I.; Kosonen, T.; Alfthan, G.; Maenpaa, J.; Perttunen, K.; Kenraali, J.; Parantainen, J.; Aro, A. Pharmacokinetics of quercetin from quercetin aglycone and rutin in healthy volunteers. Eur. J. Clin. Pharmacol. 2000, 56, 545-553. 129. Day, A. J.; Mellon, F.; Barron, D.; Sarrazin, G.; Morgan, M. R. A.; Williamson, G. Human metabolism of dietary flavonoids: Identification of plasma metabolites of quercetin. Free Rad. Res. 2001, 35, 941-952. 130. Boots, A. W.; Haenen, G. R. M. M.; Bast, A. Health effects of quercetin: From antioxidant to nutraceutical. Eur. J. Pharmacol. 2008, 585, 325-337. 131. Bischoff, S. C. Quercetin: potentials in the prevention and therapy of disease. Curr. Opin. Clin. Nutr. Metabol. Care 2008, 11, 733-740. 132. Larson, A. J.; Symons, J. D.; Jalili, T. Quercetin: a treatment for hypertension? A review of efficacy and mechanisms. Pharmaceuticals 2010, 3, 237-250. 133. Orsolic, N.; Knezevic, A. H.; Sver, L.; Terzic, S.; Basic, I. Immunomodulatory and antimetastatic action of propolis and related polyphenolic compounds. J. Ethnopharmacol. 2004, 94, 307-315. 134. Ossola, B.; Kaariainen, T. M.; Mannisto, P. T. The multiple faces of quercetin in neuroprotection. Exp. Opin. Drug Saf. 2009, 8, 397-409. 135. Cushnie, T. P. T.; Lamb, A. J. Antimicrobial activity of flavonoids. Int. J. Antimicrob. Agents 2005, 26, 343-356. 136. Paquay, J. B. G.; Haenen, G. R. M. M.; Korthouwer, R. E. M.; Bast, A. Peroxynitrite scavenging by wines. J. Agric. Food Chem. 1997, 45, 3357- 3358.

Chapter 1

137. Heijnen, C. G. M.; Haenen, G. R. M. M.; van Acker, F. A. A.; van der Vijgh, W. J. F.; Bast, A. Flavonoids as peroxynitrite scavengers: the role of the hydroxyl groups. Toxicol. in Vitro 2001, 15, 3-6. 138. Capitani, C. D.; Carvalho, A. C. L.; Rivelli, D. P.; Barros, S. B. M.; Castro, I. A. Evaluation of natural and synthetic compounds according to their antioxidant activity using a multivariate approach. Eur. J. Lipid. Sci. Technol. 2009, 111, 1090-1099. 139. Celik, H.; Arinc, E. Evaluation of the protective effects of quercetin, rutin, resveratrol, naringenin and trolox against idarubicin-induced DNA damage. J. Pharm. Pharmaceut. Sci. 2010, 13, 231-241. 140. Metodiewa, D.; Jaiswal, A. K.; Cenas, N.; Dickancaite, E.; Segura-Aguilar, J. Quercetin may act as a cytotoxic prooxidant after its metabolic activation to semiquinone and quinoidal product. Free Rad. Biol. Med. 1999, 26, 107- 116. 141. Martins, R.; Braga, H.; da Silva, A. P.; Dalmarco, J. B.; de Bem, A. F.; dos Santos, A. R.; Dafre, A. L.; Pizzolatti, M. G.; Latini, A.; Aschner, M.; Farina, M. Synergistic neurotoxicity induced by methylmercury and quercetin in mice. Food Chem. Toxicol. 2009, 47, 645-649. 142. Boots, A. W.; Kubben, N.; Haenen, G. R. M. M.; Bast, A. Oxidized quercetin reacts with thiols rather than with ascorbate: implication for quercetin supplementation. Biochem. Biophys. Res. Commun. 2003, 308, 560-565. 143. Edwards, R. L.; Lyon, T.; Litwin, S. E.; Rabovsky, A.; Symons, J. D.; Jalili, T. Quercetin reduces blood pressure in hypertensive subjects. J. Nutr. 2007, 137, 2405-2411. 144. Boots, A. W.; Drent, M.; de Boer, V. C. S.; Bast, A.; Haenen, G. R. M. M. Quercetin reduces markers of oxidative stress and inflammation in sarcoidosis. 2009; pp 161-176. 145. Langcake, P.; Pryce, R. J. Production of resveratrol and viniferins by grapevines in response to Uv irradiation. Phytochemistry 1977, 16, 1193-1196. 146. Soleas, G. J.; Diamandis, E. P.; Goldberg, D. M. Resveratrol: A molecule whose time has come? And gone? Clin. Biochem. 1997, 30, 91-113. 147. Basly, J. P.; Marre-Fournier, F.; Le Bail, J. C.; Habrioux, G.; Chulia, A. J. Estrogenic/antiestrogenic and scavenging properties of (E)- and (Z)- resveratrol. Life Sci. 2000, 66, 769-777. 148. Bertelli, A.; Bertelli, A. A. E.; Gozzini, A.; Giovannini, L. Plasma and tissue resveratrol concentrations and pharmacological activity. Drugs Exp. Clin. Res. 1998, 24, 133-138. 149. de la Lastra, C. A.; Villegas, I. Resveratrol as an anti-inflammatory and anti- aging agent: Mechanisms and clinical implications. Mol. Nutr. Food Res. 2005, 49, 405-430. 150. Gülcin, I. Antioxidant properties of resveratrol: a structure-activity insight. Innov. Food Sci. Emerg. Technol. 2010, 11, 210-218. 151. Lu, M.; Fang, J. G.; Liu, Z. L.; Wu, L. M. Effects of resveratrol and its analogs on scavenging hydroxyl radicals: evaluation by EPR spin trapping method. Appl. Magn. Reson. 2002, 22, 475-481. 152. Burkitt, M. J.; Duncan, J. Effects of trans-resveratrol on copper-dependent hydroxyl-radical formation and DNA damage: evidence for hydroxyl- radical scavenging and a novel, glutathione-sparing mechanism of action. Arch. Biochem. Biophys. 2000, 381, 253-263. 153. Vieira de Almeida, L. M.; Celintano Pineiro, C.; Leite, M.; Brolese, G.; Leal, R.; Gottfried, C.; Gonçalves, C. A. Protective effects of resveratrol on hydrogen

Chapter 1

peroxide induced toxicity in primary cortical astrocyte cultures. Neurochem. Res. 2008, 33, 8-15. 154. Fang, J. G.; Lu, M.; Chen, Z. H.; Zhu, H. H.; Li, Y.; Yang, L.; Wu, L. M.; Liu, Z. L. Antioxidant effects of resveratrol and its analogues against the free- radical-induced peroxidation of linoleic acid in micelles. Chem. Eur. J. 2002, 8, 4191-4198. 155. de la Lastral, C. A.; Villegas, I. Resveratrol as an antioxidant and pro-oxidant agent: mechanisms and clinical implications. Biochem. Soc. Trans. 2007, 35, 1156-1160. 156. 2,6-di-tert-butyl-p-cresol(BHT); 2002. 157. Chemicals known to the state to cause cancer or reproductive toxicity. 2010. 158. Re, R.; Pellegrini, N.; Proteggente, A.; Pannala, A.; Yang, M.; Rice-Evans, C. Antioxidant activity applying and improved abts radical cation decolorization assay. Free Rad. Biol. Med. 1999, 26, 1231-1237. 159. Castro, I. A.; Rogero, M. M.; Junqueira, R. M.; Carrapeiro, M. M. Free radical scavenger and antioxidant capacity correlation of +¦-tocopherol and Trolox measured by three in vitro methodologies. Int. J. Food Sci. Nutr. 2006, 57, 75-82. 160. Huang, S. W.; Frankel, E. N.; Schwarz, K.; German, J. B. Effect of pH on antioxidant activity of alpha-tocopherol and Trolox in oil-in-water emulsions. J. Agric. Food Chem. 1996, 44, 2496-2502. 161. Chung, M. J.; Walker, P. A.; Hogstrand, C. Dietary phenolic antioxidants, caffeic acid and Trolox, protect rainbow trout gill cells from nitric oxide-induced apoptosis. Aquatic Toxicol. 2006, 80, 321-328. 162. Lucio, M.; Nunes, C.; Gaspar, D.; Ferreira, H.; Lima, J. L. F. C.; Reis, S. Antioxidant activity of vitamin E and trolox: understanding of the factors that govern lipid peroxidation studies in vitro. Food Biophys. 2009, 4, 312- 320. 163. Albertini, R.; Abuja, P. M. Prooxidant and antioxidant properties of Trolox C, analogue of vitamin E, in oxidation of low-density lipoprotein. Free Rad. Res. 1999, 30, 181-188. 164. Salvi, A.; Brühlmann, C.; Migliavacca, E.; Carrupt, P. A.; Hostettmann, K.; Testa, B. Protein protection by antioxidants: development of a convenient assay and structure-activity relationships of natural polyphenols. Helv. Chim. Acta 2002, 85, 867-881. 165. Bertolini, F.; Novaroli, L.; Carrupt, P. A.; Reist, M. Novel screening assay for antioxidant protection against peroxyl radical-induced loss of protein function. J. Pharm. Sci. 2007, 96, 2931-2944. 166. Misiaszek, R.; Crean, C.; Joffe, A.; Geacintov, N. E.; Shafirovich, V. Oxidative DNA damage derived from a combination of superoxide with guanine and 8-oxo-7,8-dihydroguanine radicals. Chem. Res. Toxicol. 2004, 17, 40. 167. Melidou, M.; Riganakos, K.; Galaris, D. Protection against nuclear DNA damage offered by flavonoids in cells exposed to hydrogen peroxide: The role of iron chelation. Free Rad. Biol. Med. 2005, 39, 1591-1600. 168. Anderson, D.; Yu, T. W.; Phillips, B. J.; Schmezer, P. The effect of various antioxidants and other modifying agents on oxygen-radical-generated Dna-damage in human-lymphocytes in the comet assay. Mutation Res. 1994, 307, 261-271. 169. Diaz, Z.; Laurenzana, A.; Mann, K. K.; Bismar, T. A.; Schipper, H. M.; Miller, W. Trolox enhances the anti-lymphoma effects of arsenic trioxide, while protecting against liver toxicity. Leukemia 2007, 21, 2117-2127.

Chapter 1

170. Larsen, M.; Webb, G.; Kennington, S.; Kelleher, N.; Sheppard, J.; Kuo, J.; Unsworth-White, J. Mannitol in cardioplegia as an oxygen free radical scavenger measured by malondialdehyde. Perfusion 2002, 17, 51-55. 171. Brown, R. C.; Egleton, R. D.; Davis, T. P. Mannitol opening of the blood-brain barrier: regional variation in the permeability of sucrose, but not Rb-86(+) or albumin. Brain Res. 2004, 1014, 221-227. 172. van Vliet, E. A.; Gorter, J. A. Reply: Complexities in the association of human blood−brain barrier disruption with seizures: importance of patient population and method of disruption. Brain 2007, 130, e78. 173. Shen, B.; Jensen, R. G.; Bohnert, H. J. Mannitol protects against oxidation by hydroxyl radicals. Plant Physiol. 1997, 115, 527-532. 174. Kostopanagiotou, G.; Pandazi, A. K.; Andreadou, I.; Markantonis, S. L.; Niokou, D.; Teloudis, A.; Costopanagiotou, C.; Arkadopoulos, N.; Smyrniotis, V. Effects of mannitol in the prevention of lipid peroxidation during liver resection with hepatic vascular exclusion. J. Clin. Anesthesia 2006, 18, 570-574. 175. Seckin, B.; Sekmen, A. H.; Turkan, I. An enhancing effect of exogenous mannitol on the antioxidant enzyme activities in roots of wheat under salt stress. J. Plant Growth Regul. 2009, 28, 12-20. 176. Ashino, H.; Shimamura, M.; Nkajima, H.; Dombou, M.; Kawanaka, S.; Oikawa, T.; Iwaguchi, T.; Kawashima, S. Novel function of ascorbic acid as an angiostatic factor. Angiogenesis 2003, 6, 259-269. 177. Abudu, N.; Miller, J. J.; Attaelmannan, M.; Levinson, S. S. Vitamins in human arteriosclerosis with emphasis on vitamin C and vitamin E. Clin. Chim. Acta 2004, 339, 11-25. 178. Xu, A. M.; Vita, J. A.; Keaney, J. F. Ascorbic acid and glutathione modulate the biological activity of S-nitrosoglutathione. Hypertension 2000, 36, 291-295. 179. Bairati, I.; Meyer, F.; Gélinas, M.; Fortin, A.; Nabid, A.; Brochet, F.; Mercier, J. P.; Têtu, B.; Harel, F.; Mâsse, B.; Vigneault, E.; Vass, S.; Del Vecchio, P.; Roy, J. A randomized trial of antioxidant vitamins to prevent second primary cancers in head and neck cancer patients. J. Nat. Cancer Institute 2005, 97, 481-488. 180. Munteanu, A.; Zingg, J. M.; Azzi, A. Anti-atherosclerotic effects of vitamin E - myth or reality? J. Cell. Mol. Med. 2004, 8, 59-76. 181. Knekt, P.; Reunanen, A.; Marniemi, J.; Leino, A.; Aromaa, A. Low vitamin E status is a potential risk factor for insulin-dependent diabetes mellitus. J. Intern. Med. 1999, 245, 99-102. 182. Pazdro, R.; Burgess, J. R. The role of vitamin E and oxidative stress in diabetes complications. Mech. Ageing Dev. 2010, 131, 276-286. 183. Kono, Y.; Shibata, H.; Kodama, Y.; Sawa, Y. The suppression of the N-nitrosating reaction by chlorogenic acid. Biochem. J. 1995, 312, 947-953. 184. Hirose, M.; Takesada, Y.; Tanaka, H.; Tamano, S.; Kato, T.; Shirai, T. Carcinogenicity of antioxidants BHA, caffeic acid, sesamol, 4- methoxyphenol and catechol at low doses, either alone or in combination, and modulation of their effects in a rat medium-term multi-organ carcinogenesis model. Carcinogenesis 1997, 19, 207-212. 185. Chung, T. W.; Moon, S. K.; Chang, Y. C.; Ko, J. H.; Lee, Y. C.; CHO, G. U. N.; Kim, S. H.; Kim, J. G.; Kim, C. H. Novel and therapeutic effect of caffeic acid and caffeic acid phenyl ester on hepatocarcinoma cells: complete regression of hepatoma growth and metastasis by dual mechanism. FASEB J. 2004, 18, 1670-1681.

Chapter 1

186. Sul, D.; Kim, H. S.; Lee, D.; Joo, S. S.; Hwang, K. W.; Park, S. Y. Protective effect of caffeic acid against beta-amyloid-induced neurotoxicity by the inhibition of calcium influx and tau phosphorylation. Life Sci. 2009, 84, 257-262. 187. Belkaid, A.; Currier, J. C.; Desganes, J.; Annabi, B. The chemopreventive properties of chlorogenic acid reveal a potential new role for the microsomal glucose-6-phosphate translocase in brain tumor progression. Cancer Cell Int. 2006, 6, 1-12. 188. Kwon, S. H.; Lee, H. K.; Kim, J. A.; Hong, S. I.; Kim, H. C.; Jo, T. H.; Park, Y. I.; Lee, C. K.; Kim, Y. B.; Lee, S. Y.; Jang, C. G. Neuroprotective effects of chlorogenic acid on scopolamine-induced amnesia via anti- acetylcholinesterase and anti-oxidative activities in mice. Eur. J. Pharmacol. 2010, In Press, Corrected Proof. 189. Strlic, M.; Radovic, T.; Kolar, J.; Pihlar, B. s. Anti- and prooxidative properties of gallic acid in fenton-type systems. J. Agric. Food Chem. 2002, 50, 6313- 6317. 190. Fazary, A. E.; Taha, M.; Ju, Y. H. Iron complexation studies of gallic acid. J. Chem. Eng. Data 2009, 54, 35-42. 191. Pereira, R.; Fachinetto, R.; de Souza Prestes, A.; Puntel, R.; Santos da Silva, G.; Heinzmann, B.; Boschetti, T.; Athayde, M.; Bürger, M.; Morel, A.; Morsch, V.; Rocha, J. Antioxidant effects of different extracts from Melissa officinalis, Matricaria recutita and Cymbopogon citratus. Neurochem. Res. 2009, 34, 973-983. 192. Saeki, K.; Yuo, A.; Isemura, M.; Abe, I.; Seki, T.; Noguchi, H. Apoptosis-inducing activity of lipid derivatives of gallic acid. Biol. Pharm. Bull. 2000, 1391- 1394. 193. Sakagami, H.; Satoh, K.; Hatano, T.; Yoshida, T.; Okuda, T. Possible role of radical intensity and oxidation potential for gallic acid induced apoptosis. Anticancer Res. 1997, 17, 377-380. 194. Harbrecht, B. G.; DiSilvio, M.; Chough, V.; Kim, Y. M.; Simmons, R. L.; Billiar, T. R. Glutathione regulates nitric oxide synthase in cultured hepatocytes. Annals of Surgery 1997, 225, 76-87. 195. Leiro, J. M.; Alvarez, E.; Arranz, J. A.; Siso, I. G.; Orallo, F. In vitro effects of mangiferin on superoxide concentrations and expression of the inducible nitric oxide synthase, tumour necrosis factor-[alpha] and transforming growth factor-[beta] genes. Biochem. Pharmacol. 2003, 65, 1361-1371. 196. Darvesh, A. S.; Carroll, R. T.; Bishayee, A.; Geldenhuys, W. J.; Van der Schyf, C. J. Oxidative stress and Alzheimer's disease: dietary polyphenols as potential therapeutic agents. Exp. Rev. Neurother. 2010, 10, 729-745. 197. Reiter, R. J.; Tan, D. X.; Mayo, J. C.; Sainz, R. M.; Lopez-Burillo, S. Melatonin, longevity and health in the aged: An assessment. Free Rad. Res. 2002, 36, 1323-1329. 198. Arce, M. P.; Rodriguez-Franco, M. I.; Gonzalez-Munoz, G. C.; Perez, C.; Lopez, B.; Villarroya, M.; Lopez, M. G.; Garcia, A. G.; Conde, S. Neuroprotective and cholinergic properties of multifunctional hlutamic acid derivatives for the treatment of Alzheimer's disease. J. Med. Chem. 2009, 52, 7249-7257. 199. Fernandez-Bachiller, M. I.; Perez, C.; Campillo, N. E.; Paez, J. A.; Gonzalez- Munoz, G. C.; Usan, P.; Garcia-Palomero, E.; Lopez, M. G.; Villarroya, M.; Garcia, A. G.; Martinez, A.; Rodriguez-Franco, M. I. Tacrine-melatonin hybrids as multifunctional agents for Alzheimer's disease, with cholinergic, antioxidant, and neuroprotective properties. Chem. Med. Chem. 2009, 4, 828-841.

Chapter 1

200. Camins, A.; Sureda, F. X.; Junyent, F.; Verdaguer, E.; Folch, J.; Beas-Zarate, C.; Pallas, M. An overview of investigational antiapoptotic drugs with potential application for the treatment of neurodegenerative disorders. Exp. Opin. Invest. Drugs 2010, 19, 587-604. 201. Mayo, J. C.; Sainz, R. M.; Tan, D. X.; Antolin, I.; Rodriguez, C.; Reiter, R. J. Melatonin and Parkinson's disease. Endocrine 2005, 27, 169-178. 202. Murakami, A.; Ashida, H.; Terao, J. Multitargeted cancer prevention by quercetin. Cancer Let. 2008, 269, 315-325. 203. Kim, H.; Park, B. S.; Lee, K. G.; Choi, C. Y.; Jang, S. S.; Kim, Y. H.; Lee, S. E. Effects of naturally occurring compounds on fibril formation and oxidative stress of β-amyloid. J. Agric. Food Chem. 2005, 53, 8537-8541. 204. Tong-un, T.; Wannanon, P.; Wattanathorn, J.; Phachonpai, W. Cognitive- enhancing and antioxidant activities of Quercetin liposomes in animal model of Alzheimer's disease. J. Biol. Sci. 2010, 2, 84-91. 205. Bournival, J.; Quessy, P.; Martinoli, M. G. Protective effects of resveratrol and quercetin against MPP+-Induced oxidative stress act by modulating markers of apoptotic death in dopaminergic neurons. Cell. Mol. Neurobiol. 2009, 29, 1169-1180. 206. Singh, A.; Naidu, P. S.; Kulkarni, S. K. Quercetin potentiates L-dopa reversal of drug-induced catalepsy in rats: possible COMT/MAO inhibition. Pharmacology 2003, 68, 81-88. 207. Bagli, J.; Bogri, T.; Voith, K.; Lee, D. Troponoids. 7. Chemistry and dopamine agonist activity of ciladopa and related aralkyltroponylpiperazines. J. Med. Chem. 1986, 29, 186-193. 208. Wang, M.; Jin, Y.; Ho, C. T. Evaluation of resveratrol derivatives as potential antioxidants and identification of a reaction product of resveratrol and 2,2- Diphenyl-1-picryhydrazyl radical. J. Agric. Food Chem. 1999, 47, 3974- 3977. 209. Belguendouz, L.; Frémont, L.; Gozzelino, M. T. Interaction of transresveratrol with plasma lipoproteins. Biochem. Pharmacol. 1998, 55, 811-816. 210. Jang, M. S.; Cai, E. N.; Udeani, G. O.; Slowing, K. V.; Thomas, C. F.; Beecher, C. W. W.; Fong, H. H. S.; Farnsworth, N. R.; Kinghorn, A. D.; Mehta, R. G.; Moon, R. C.; Pezzuto, J. M. Cancer chemopreventive activity of resveratrol, a natural product derived from grapes. Science 1997, 275, 218-220. 211. Bhat, K. P. L.; Pezzuto, J. M. Cancer chemopreventive activity of resveratrol. Ann. N. Y. Acad. Sci. 2002, 957, 210-229. 212. Aziz, M. H.; Kumar, R.; Ahmad, N. Cancer chemoprevention by resveratrol: In vitro and in vivo studies and the underlying mechanisms (Review). Int. J. Oncol. 2003, 23, 17-28. 213. Goldberg, D. M. Does wine work? Clin. Chem. 1995, 41, 14-16. 214. Queen, B. L.; Tollefsbol, T. O. Polyphenols and aging. Curr. Aging Sci. 2010, 3, 34-42. 215. Sun, A.; Wang, Q.; Simonyi, A.; Sun, G. Resveratrol as a therapeutic agent for neurodegenerative diseases. Mol. Neurobiol. 2010, 41, 375-383. 216. Annanmaki, T.; Muuronen, A.; Murros, K. Low plasma uric acid level in Parkinson's disease. Movement Disorders 2007, 22, 1133-1137. 217. Winquist, A.; Steenland, K.; Shankar, A. Higher serum uric acid associated with decreased Parkinson's disease prevalence in a large community-based survey. Movement Disorders 2010, 25, 932-936. 218. Auinger, P.; Kieburtz, K.; McDermott, M. P. The relationship between uric acid levels and Huntington's disease progression. Movement Disorders 2010, 25, 224-228.

Chapter 1

219. Andersen, J. K. Oxidative stress in neurodegeneration: cause or consequence ? Neurodegeneration 2004, S18-S25. 220. Valko, M.; Rhodes, C. J.; Moncol, J.; Izakovic, M.; Mazur, M. Free radicals, metals and antioxidants in oxidative stress-induced cancer. Chem. Biol. Interactions 2006, 160, 1-40. 221. Halliwell, B. Oxidative stress and cancer: have we moved forward? Biochem. J. 2007, 401, 1-11. 222. Khullar, M.; Binepal, G.; Al-Shudiefat, A. A. R. S.; Ludke, A.; Singal, P. K. Oxidative stress: a key contributor to diabetic cardiomyopathy. Can. J. Physiol. Pharmacol. 2010, 88, 233-240. 223. Shen, G. X. Oxidative stress and diabetic cardiovascular disorders: roles of mitochondria and NADPH oxidase. Can. J. Physiol. Pharmacol. 2010, 88, 241-248. 224. de Bernonville, T. D.; Guyot, S.; Paulin, J. P.; Gaucher, M.; Loufrani, L.; Henrion, D.; Derbre, S.; Guilet, D.; Richomme, P.; Dat, J. F.; Brisset, M. N. Dihydrochalcones: Implication in resistance to oxidative stress and bioactivities against advanced glycation end-products and vasoconstriction. Phytochemistry 2010, 71, 443-452. 225. Zana, M.; Janka, Z.; Kalman, J. Oxidative stress: A bridge between Down's syndrome and Alzheimer's disease. Neurobiology of Aging 2007, 28, 648- 676. 226. Nunomura, A.; Perry, G.; Pappolla, M. A.; Friedland, R. P.; Hirai, K.; Chiba, S.; Smith, M. A. Neuronal oxidative stress precedes amyloid-beta deposition in Down syndrome. J. Neuropathol. Exp. Neurol. 2000, 59, 1011-1017. 227. Altindag, O.; Karakoc, M.; Kocyigit, A.; Celik, H.; Soran, N. Increased DNA damage and oxidative stress in patients with rheumatoid arthritis. Clin. Biochem. 2007, 40, 167-171. 228. Finkel, T.; Holbrook, N. J. Oxidants, oxidative stress and the biology of ageing. Nature 2000, 408, 239-247. 229. Lithgow, G. J.; Andersen, J. K. Models of oxidative stress in the biology of aging. Drug Discovery Today 2005, 2, 273-277. 230. Stadtman, E. R. Role of oxidant species in aging. Curr. Med. Chem. 2004, 11, 1105-1112. 231. Valko, M.; Leibfritz, D.; Moncol, J.; Cronin, M. T. D.; Mazur, M.; Telser, J. Free radicals and antioxidants in normal physiological functions and human disease. Int. J. Biochem. Cell. Biol. 2007, 39, 44-84. 232. Crichton, R. R.; Dexter, D. T.; Ward, R. J. Metal based neurodegenerative diseases--From molecular mechanisms to therapeutic strategies. Coord. Chem. Rev. 2008, 252, 1189-1199. 233. Uttara, B.; Singh, A. V.; Zamboni, P.; Mahajan, R. T. Oxidative stress and neurodegenerative diseases: a review of upstream and downstream antioxidant therapeutic options. Curr. Neuropharmacol. 2009, 7, 65-74. 234. Adibhatla, R. M.; Hatcher, J. F. Lipid oxidation and peroxidation in CNS health and disease: from molecular mechanisms to therapeutic opportunities. Antioxid. Redox Signal. 2010, 12, 125-169. 235. Wimo, A.; Prince, M. Alzheimer's Disease International - World Alzheimer Report 2010 - The Global Economic Impact of Dementia; 2010. 236. Butterfield, D. A.; Lauderback, M. Lipid peroxidation and protein oxidation in Alzheimer's disease brain: potential causes and consequences involving amyloid β-peptide-associated free radical oxidative stress. Free Rad. Biol. Med. 2002, 32, 1050-1060.

Chapter 1

237. Pratico, D. Oxidative stress hypothesis in Alzheimer's disease: a reappraisal. Trends Pharmacol. Sci. 2008, 29, 609-615. 238. Nunomura, A.; Castellani, R. J.; Zhu, X.; Moreira, P. I.; Perry, G.; Smith, M. A. Involvement of oxidative stress in Alzheimer disease. J. Neuropathol. Exp. Neurol. 2006, 65, 631-641. 239. Hardy, J. A.; Higgins, G. A. Alzheimers-disease - the amyloid cascade hypothesis. Science 1992, 256, 184-185. 240. Swerdlow, R. H.; Burns, J. M.; Khan, S. M. The Alzheimer's disease mitochondrial cascade hypothesis. J. Alzheimers Dis. 2010, 20, S265-S279. 241. Kontush, A.; Berndt, C.; Weber, W.; Akopyan, V.; Arlt, S.; Schippling, S.; Beisiegel, U. Amyloid-beta is an antioxidant for lipoproteins in cerebrospinal fluid and plasma. Free Rad. Biol. Med. 2001, 30, 119-128. 242. Atwood, C. S.; Obrenovich, M. E.; Liu, T. B.; Chan, H.; Perry, G.; Smith, M. A.; Martins, R. N. Amyloid-beta: a chameleon walking in two worlds: a review of the trophic and toxic properties of amyloid-beta. Brain Res. Rev. 2003, 43, 1-16. 243. Dikalov, S. I.; Vitek, M. P.; Mason, R. P. Cupric-amyloid beta peptide complex stimulates oxidation of ascorbate and generation of hydroxyl radical. Free Rad. Biol. Med. 2004, 36, 340-347. 244. Kontush, A. Amyloid-beta: an antioxidant that becomes a pro-oxidant and critically contributes to Alzheimer's disease. Free Rad. Biol. Med. 2001, 31, 1120-1131. 245. Miyata, M.; Smith, J. D. Apolipoprotein E allele-specific antioxidant activity and effects on cytotoxicity by oxidative insults and beta-amyloid peptides. Nature Genetics 1996, 14, 55-61. 246. Lee, H. G.; Perry, G.; Moreira, P. I.; Garrett, M. R.; Liu, Q.; Zhu, X. W.; Takeda, A.; Nunomura, A.; Smith, M. A. Tau phosphorylation in Alzheimer's disease: pathogen or protector? Trends Mol. Med. 2005, 11, 164-169. 247. Liu, Q.; Smith, M. A.; Avila, J.; DeBernardis, J.; Kansal, M.; Takeda, A.; Zhu, X.; Nunomura, A.; Honda, K.; Moreira, P. I.; Oliveira, C. R.; Santos, M. S.; Shimohama, S.; Aliev, G.; de la Torre, J.; Ghanbari, H. A.; Siedlak, S. L.; Harris, P. L. R.; Sayre, L. M.; Perry, G. Alzheimer-specific epitopes of tau represent lipid peroxidation-induced conformations. Free Rad. Biol. Med. 2005, 38, 746-754. 248. Alvarez, A.; Opazo, C.; Alarcon, R.; Garrido, J.; Inestrosa, N. C. Acetylcholinesterase promotes the aggregation of amyloid-beta-peptide fragments by forming a complex with the growing fibrils. J. Mol. Biol. 1997, 272, 348-361. 249. Bartolini, M.; Bertucci, C.; Cavrini, V.; Andrisano, V. β-Amyloid aggregation induced by human acetylcholinesterase: inhibition studies. Biochem. Pharmacol. 2003, 65, 407-416. 250. Bryson, H. M.; Benfield, P. Donepezil. Drugs Aging 1997, 10, 234-239. 251. Gabelli, C. Rivastigmine: an update on therapeutic efficacy in Alzheimer's disease and other conditions. Curr. Med. Res. Opin. 2003, 19, 69-82. 252. SrameK, J. J.; Frackiewicz, E. J.; Cutler, N. R. Review of the acetylcholinesterase inhibitor galanthamine. Exp. Opin. Invest. Drugs 2000, 9, 2393-2402. 253. Raina, P.; Santaguida, P.; Ismaila, A.; Patterson, C.; Cowan, D.; Levine, M.; Booker, L.; Oremus, M. Effectiveness of cholinesterase inhibitors and memantine for treating dementia: Evidence review for a clinical practice guideline. Ann. Int. Med. 2008, 148, 379-W85.

Chapter 1

254. Marambaud, P.; Zhao, H. T.; Davies, P. Resveratrol promotes clearance of Alzheimer's disease amyloid-beta peptides. J. Biol. Chem. 2005, 280, 37377-37382. 255. Fahrenholz, F.; Tippmann, F.; Endres, K. Retinoids as a perspective in treatment of Alzheimer's disease. Neurodegenerative Dis. 2010, 7, 190-192. 256. Klaver, D.; Hung, A. C.; Gasperini, R.; Foa, L.; Aguilar, M. I.; Small, D. H. Effect of Heparin on APP metabolism and Aβ production in cortical neurons. Neurodegenerative Dis. 2010, 7, 187-189. 257. Solomon, A.; Sippola, R.; Soininen, H.; Wolozin, B.; Tuomilehto, J.; Laatikainen, T.; Kivipelto, M. Lipid-lowering treatment is related to decreased risk of dementia: a population-based study (FINRISK). Neurodegenerative Dis. 2010, 7, 180-182. 258. Spillantini, M. G.; Schmidt, M. L.; Lee, V. M. Y.; Trojanowski, J. Q.; Jakes, R.; Goedert, M. alpha-synuclein in Lewy bodies. Nature 1997, 388, 839-840. 259. Moore, D. J.; West, A. B.; Dawson, V. L.; Dawson, T. M. Molecular pathophysiology of Parkinson's disease. Ann. Rev. Neurosci. 2005, 28, 57- 87. 260. Lee, F. J. S.; Liu, F. Genetic factors involved in the pathogenesis of Parkinson's disease. Brain Res. Rev. 2008, 58, 354-364. 261. Schapira, A. H. V.; Cooper, J. M.; Dexter, D.; Clark, J. B.; Jenner, P.; Marsden, C. D. Mitochondrial complex I deficiency in Parkinsons-disease. J. Neurochem. 1990, 54, 823-827. 262. Henchcliffe, C.; Beal, M. F. Mitochondrial biology and oxidative stress in Parkinson disease pathogenesis. Nat. Clin. Pract. Neurol. 2008, 4, 600- 609. 263. Lotharius, J.; Brundin, P. Pathogenesis of parkinson's disease: dopamine, vesicles and [alpha]-synuclein. Nat. Rev. Neurosci. 2002, 3, 932-942. 264. Miller, R.; James-Kracke, M.; Sun, G.; Sun, A. Oxidative and inflammatory pathways in Parkinson's disease. Neurochem. Res. 2009, 34, 55-65. 265. Foster, H. D.; Hoffer, A. The two faces of L-DOPA: benefits and adverse side effects in the treatment of Encephalitis lethargica, Parkinson's disease, multiple sclerosis and amyotrophic lateral sclerosis. Medical Hypotheses 2004, 62, 177-181. 266. Paleologou, K. E.; Irvine, G. B.; El-Agnaf, O. M. A. α-synuclein aggregation in neurodegenerative diseases and its inhibition as a potential therapeutic strategy. Biochem. Soc. Trans. 2005, 33, 1103-1110. 267. Gerard, M.; Deleersnijder, A.; Daniels, V.; Schreurs, S.; Munck, S.; Reumers, V.; Pottel, H.; Engelborghs, Y.; Van den Haute, C.; Taymans, J. M.; Debyser, Z.; Baekelandt, V. Inhibition of FK506 binding proteins reduces {alpha}- synuclein aggregation and Parkinson's disease-like pathology. J. Neurosci. 2010, 30, 2454-2463. 268. Smith, R. A. J.; Adlam, V. J.; Blaikie, F. H.; Manas, A. R. B.; Porteous, C. M.; James, A. M.; Ross, M. F.; Logan, A.; Cocheme, H. M.; Trnka, J.; Prime, T. A.; Abakumova, I.; Jones, B. A.; Filipovska, A.; Murphye, M. P. Mitochondria-targeted antioxidants in the treatment of cisease. Mitochondria and Oxidative Stress in Neurodegenerative Disorders 2008, 1147, 105-111. 269. Bonda, D. J.; Wang, X.; Gustaw-Rothenberg, K. A.; Perry, G.; Smith, M. A.; Zhu, X. Mitochondrial drugs for Alzheimer disease. Pharmaceuticals 2009, 2, 287-298. 270. Tauskela, J. S. MitoQ-a mitochondria targeted antioxidant. J. Drugs 2007, 10, 399-412.

Chapter 1

271. Brown, D. R. Metalloproteins and neuronal death. Metallomics 2010, 2, 186-194. 272. Ricchelli, F.; Drago, D.; Filippi, B.; Tognon, G.; Zatta, P. Aluminum-triggered structural modifications and aggregation of β-amyloids. Cell. Mol. Life Sci. 2005, 62, 1724-1733. 273. Adlard, P. A.; Cherny, R. A.; Finkelstein, D. I.; Gautier, E.; Robb, E.; Cortes, M.; Volitakis, I.; Liu, X.; Smith, J. P.; Perez, K.; Laughton, K.; Li, Q. X.; Charman, S. A.; Nicolazzo, J. A.; Wilkins, S.; Deleva, K.; Lynch, T.; Kok, G.; Ritchie, C. W.; Tanzi, R. E.; Cappai, R.; Masters, C. L.; Barnham, K. J.; Bush, A. I. Rapid restoration of cognition in Alzheimer's transgenic mice with 8-Hydroxy quinoline analogs is associated with decreased interstitial A[beta]. Neuron 2008, 59, 43-55. 274. Uversky, V. N.; Li, J.; Fink, A. L. Metal-triggered structural transformations, aggregation, and fibrillation of human alpha-synuclein - A possible molecular link between Parkinson's disease and heavy metal exposure. J. Biol. Chem. 2001, 276, 44284-44296. 275. Storr, T.; Merkel, M.; Song-Zhao, G. X.; Scott, L. E.; Green, D. E.; Bowen, M. L.; Thompson, K. H.; Patrick, B. O.; Schugar, H. J.; Orvig, C. Synthesis, characterization, and metal coordinating ability of multifunctional carbohydrate-containing compounds for Alzheimer's therapy. J. Am. Chem. Soc. 2007, 129, 7453-7463. 276. Bolognin, S.; Drago, D.; Messori, L.; Zatta, P. Chelation therapy for neurodegenerative diseases. Med. Res. Rev. 2009, 29, 547-570. 277. Boschi, D.; Tron, G. C.; Lazzarato, L.; Chegaev, K.; Cena, C.; Di Stilo, A.; Giorgis, M.; Bertinaria, M.; Fruttero, R.; Gasco, A. NO-donor phenols: a new class of products endowed with antioxidant and vasodilator properties. J. Med. Chem. 2006, 49, 2886-2897. 278. Gasco, A.; Fruttero, R.; Rolando, B. Focus on recent approaches for the development of new NO-donors. Mini Rev. Med. Chem. 2005, 5, 217-229. 279. Horvathova, K.; Chalupa, I.; Sebova, L.; Tothova, D.; Vachalkova, A. Protective effect of quercetin and luteolin in human melanoma HMB-2 cells. Mutation Res. 2005, 565, 102-112. 280. Colic, M.; Pavelic, K. Molecular mechanisms of anticancer activity of natural dietetic products. J. Mol. Med. 2000, 78, 333-336. 281. Di Stefano, A.; Sozio, P.; Cocco, A.; Iannitelli, A.; Santucci, E.; Costa, M.; Pecci, L.; Nasuti, C.; Cantalamessa, F.; Pinnen, F. L-Dopa- and Dopamine-(R)-α- lipoic acid conjugates as multifunctional codrugs with antioxidant properties. J. Med. Chem. 2006, 49, 1486-1493. 282. Zheng, H.; Gal, S.; Weiner, L. M.; Bar-Am, O.; Warshawsky, A.; Fridkin, M.; Youdim, M. B. H. Novel multifunctional neuroprotective iron chelator- monoamine oxidase inhibitor drugs for neurodegenerative diseases: in vitro studies on antioxidant activity, prevention of lipid peroxide formation and monoamine oxidase inhibition. J. Neurochem. 2005, 95, 68-78. 283. Martinez, A.; Fernandez, E.; Castro, A.; Conde, S.; Rodriguez-Franco, I.; Banos, J. E.; Badia, A. N-benzylpiperidine derivatives of 1,2,4-thiadiazolidinone as new acetylcholinesterase inhibitors. Eur. J. Med. Chem. 2000, 35, 913-922. 284. Youdim, M. B. H.; Buccafusco, J. J. CNS targets for multi-functional drugs in the treatment of Alzheimer's and Parkinson's diseases. J. Neural Transm. 2005, 112, 519-537. 285. Zecca, L.; Youdim, M. B. H.; Riederer, P.; Connor, J. R.; Crichton, R. R. Iron, brain ageing and neurodegenerative disorders. Nature Rev. 2004, 5, 863- 873.

Chapter 1

286. Rosini, M.; Andrisano, V.; Bartolini, M.; Bolognesi, M. L.; Hrelia, P.; Minarini, A.; Tarozzi, A.; Melchiorre, C. Rational approach to discover multipotent anti- Alzheimer drugs. J. Med. Chem. 2005, 48, 360-363. 287. Rodriguez-Franco, M.; Fernandez-Bachiller, M. I.; Perez, C.; Hernandez-Ledesma, B.; Bartolomé, B. Novel tacrine-melatonin hybrids as dual-acting drugs for Alzheimer disease, with improved acetylcholinesterase inhibitory and antioxidant properties. J. Med. Chem. 2006, 49, 459-462. 288. Christiaans, J. A. M.; Timmerman, H. Cardiovascular hybrid drugs: combination of more than one pharmacological property in one single molecule. Eur. J. Pharm. Sci. 1996, 4, 1-22. 289. Litwinienko, G.; Ingold, K. U. Abnormal solvent effects on hydrogen atom abstraction. 2. Resolution of the curcumin antioxidant controversy. The role of sequential proton loss electron transfer. J. Org. Chem. 2004, 69, 5888-5896. 290. Ak, T.; Gulcin, I. Antioxidant and radical scavenging properties of curcumin. Chem. Biol. Interactions 2008, 174, 27-37. 291. Yang, F.; Lim, G. P.; Begum, A. N.; Ubeda, O. J.; Simmons, M. R.; Ambegaokar, S. S.; Chen, P.; Kayed, R.; Glabe, C. G.; Frautschy, S. A.; Cole, G. M. Curcumin inhibits formation of amlyoid β oligomers and fibrils, binds plaques, and reduces amyloid in vivo. J. Biol. Chem. 2005, 280, 5892-5901. 292. Weinstock, M.; Luques, L.; Bejar, C.; Shoham, S. Ladostigil, a novel multifunctional drug for the treatment of dementia co-morbid with depression. J. Neural Transm. Suppl. 2006, 443-446. 293. Gal, S.; Fridkin, M.; Amit, T.; Zheng, H.; Youdim, M. B. H. M30, a novel multifunctional neuroprotective drug with potent iron chelating and brain selective monoamine oxidase-ab inhibitory activity for Parkinson's disease. J. Neural Transm. Suppl. 2006, 447-456. 294. Cavalli, A.; Bolognesi, M. L.; Capsoni, S.; Andrisano, V.; Bartolini, M.; Margotti, E.; Cattaneo, A.; Recanatini, M.; Melchiorre, C. A small molecule targeting the multifactorial nature of Alzheimer's disease. Angew. Chem. Int. Ed. 2007, 46, 3689-3692. 295. Bellik, L.; Dragoni, S.; Pessina, F.; Sanz, E.; Unzeta, M.; Valoti, M. Antioxidant properties of PF9601N, a novel MAO-B inhibitor: assessment of its ability to interact with reactive nitrogen species. Acta Biochim. Pol. 2010, 57, 235-239.

Chapter 2

Chapter 2 Determination of parameters for the characterization of new antioxidant compounds

2-1 Introduction

Drug development is a long and expensive process which provides only a few efficient drugs among the huge number of drug candidates1,2. It includes drug discovery, pre-clinical research, and clinical trials (Figure 2-1). In a selected disease and after the identification and the choice of drug target, the capacity of a large set of compounds for acting on the target is assessed. The active compounds, which emerge from the library at this discriminatory step, are named hits. Afterward, the properties of the hits including cell membrane permeability, solubility, and selectivity against other related targets are checked. The hits exhibiting the most appropriate properties are selected and named lead compounds. This is the hit-to-lead phase. Pre-clinical research and clinical trials follow to assess the lead compound activity in vivo.

Figure 2-1 : Drug development process.

Chapter 2

To save time and money, high-throughput screening (HTS) is required to release hits from the first discriminatory step. This process may use robotics, sensitive detectors, microplates with 96, 384, 1536, and 3456 wells. A new approach based on drop-based microfluidics allows a faster screening at a more cheaply cost for directed evolution3. It could enhance the utility and the applicability of HTS in some biology and chemistry areas. In this study, four methods have been selected to assess the antioxidant properties against radicals of different stability. All of them may be used for high- throughput screening. The first method, which is the ALP assay4, is based on the oxidation of the protein alkaline phosphatase (ALP). Generally, the assays focused on protein oxidation are used to investigate the structural damages, such as the detection of carbonyl groups in the protein structure5, protein fragmentation, and aggregation6,7, as well as amino acid side chain modification8. This one is based on the protein activity loss induced by peroxyl radicals formed from the thermal decomposition of 2,2’-Azobis(2-methylpropionamidine) dihydrochloride (AAPH) (Figure 2-2). The antioxidant property is the capacity to protect the catalytic activity to hydrolyze 4-methylumbelliferyl phosphate (MUP) to the fluorescent product 4-methylumbelliferone (MU).

NH NH Δ H2N NN 2 + N NH2 H2N 2 2 O2 OO HN Figure 2-2: Generation of peroxyl radical from AAPH decomposition.

The oxygen radical absorbance capacity (ORAC)9 is also carried out with peroxyl radicals generated by AAPH, but the oxidative damage is related to the oxidation of fluorescein that is a fluorescent compound. Thus, the degree of oxidation is monitored by the fluorescence decrease. The antioxidant activity related to the capacity of a compound to avoid the oxidation of fluorescein is measured by the amount of fluorescence in presence of the antioxidant. In these

Chapter 2 two assays, the peroxyl radicals are produced continually by AAPH, contrary to the two other methods wherein the antioxidant property is the ability to scavenge an initial amount of radicals. These radicals are 2,2-diphenyl-1-picrylhydrazine (DPPH•)10 (Figure 2-3) and 2,2’-azino-bis-(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS) radical anion (ABTS•–)11 (Figure 2-4).

O2N O2N

N N NO N NO2 + AH N 2 + A

O2N O2N

Purple Yellow Figure 2-3 : Scavenging of DPPH• by an antioxidant AH.

N N O3S S N O3S S N S + AH N S + AH N SO3 SO3 N N Colorless Green Figure 2-4 : Reduction of ABTS•– by an antioxidant AH.

It is well-accepted that an antioxidant compound scavenges the DPPH• radical through hydrogen atom transfer. However, the reduction mechanism of ABTS•– is not that clear since both hydrogen atom transfer (HAT) and electron transfer (ET) mechanisms are considered10,12-14. It has to be mentioned that the ABTS• radical is most often written ABTS•+. However, the two sulfonic acid groups are likely deprotonated which induces the ABTS to exist in its ionic dianion form. The notation ABTS•– seems therefore more appropriate. In order to define the screening parameters in the four assays, different conditions were studied by using compounds known to have antioxidant abilities, namely ascorbic acid15, caffeic acid4, chlorogenic acid4, gallic acid16, glutathione17, mangiferin4, mannitol18,19, melatonin4, quercetin4, resveratrol4, α-tocopherol10, trolox®11, uric acid20. A large set of chemically diverse compounds has been screened to optimize the screening parameters. To more easily compare the

Chapter 2 compound activities, classes have been defined depending on the potency toward the radicals in each assay and on the free radical-scavenging kinetics in DPPH• and ABTS•– assays.

2-2 Materials and methods

2-2.1 Materials

Alkaline phosphatase (ALP) from calf intestinal mucosa (EC 3.1.3.1), [9001- 78-9], 4-methylumbelliferyl phosphate disodium salt (MUP), 2,2’-Azobis(2- methylpropionamidine) dihydrochloride (AAPH), glycine, magnesium chloride hexahydrate, sodium hydroxide, sodium chloride, ascorbic acid, melatonin, uric acid, and dimethylsulfoxide (DMSO) of microbiological quality were purchased from Fluka (Buchs, Switzerland). 2,2-diphenylpicrylhydrazine (DPPH•), 2,2’- azinobis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS), potassium persulfate, caffeic acid, chlorogenic acid, gallic acid, L-glutathione, mangiferin from Mangifer indica Burk, D-mannitol, quercetin, resveratrol, α- tocopherol, and trolox® were purchased from Sigma-Aldrich (Buchs, Switzerland). Fluorescein (FL) was purchased from Siegfried Handel (Zofingen, Switzerland). A 146-compound set has been used to validate the screening parameters. It includes aurone- and azaaurone-derivatives (annexes II-B and II-C, respectively) prepared in the group of Prof. Ahcene Boumendjel (Département de Pharmacologie Moléculaire, Grenoble, France), pyrrolidine-3,4-diol and conduramine analogs (annexes II-D to II-J) synthesized in the group of Prof. Pierre Vogel by Dr Robert Lysek (Ecole Polytechnique Fédérale de Lausanne, Switzerland), pyrazolo[4,3-b]pyrrolizine-derivatives (annexes II-L and II-M) synthesized in the group of Prof. Antonello Mai (Dipartimento di Studi Farmaceutici. Università degli Studi di Roma, Italy), and xanthone-derivatives either synthesized in the group of Prof. Madalena Pinto (Laboratorio de Quimica Organica. Faculdade de Farmacia. Porto, Portugal) or extracted from Chironia

Chapter 2

Kresbii in the group of Prof. Kurt Hostettmann (Laboratoire de pharmacognosie et phytochimie. Université de Genève-Lausanne. Geneva, Switzerland) (annex II- O). However, only the compounds yielding the most relevant data for the validation are described in this chapter. All values of the potency and the kinetic parameters are reported in annex II. Water solutions were prepared in demineralised and purified water obtained with the Elix 3 Millipore water purifying system. The fluorescence was monitored using a Bio Tek FLX 800 microplate fluorescence reader and the absorbance was measured using a Bio-Tek Power wave X microplate absorbance reader. KC4 v3.3 software (Bio Tec Instruments Inc., Winooski, USA) was used with both readers. The dose-response curves were plotted using GraphPad prism version 5.00 software (San Diego California, USA).

2-2.2 Methods

2-2.2.1 ALP assay

The antioxidant activity of a compound was determined by its ability to preserve the catalytic effectiveness of the enzyme alkaline phosphatase (ALP) despite the presence of peroxyl radicals generated by 2,2’-azobis(2- methylpropionamidine) dihydrochloride (AAPH). The reaction followed to assess the catalytic activity of ALP was the enzymatic dephosphorylation of 4- methylumbelliferyl phosphate (MUP) to fluorescent 4-methylumbelliferone (4- MU) (Figure 2-5). Enzymatic hydrolysis rates of MUP were determined by a continuous spectrofluorimetric assay4.

Chapter 2

OH ALP HO + H2O + H3PO4 P O O O O O O OH MUP MU Figure 2-5: Dephosphorylation of 4-methylumbelliferyl phosphate (MUP) to 4- methylumbelliferone (MU) catalyzed by alkaline phosphatase (ALP).

ALP, AAPH, and MUP solutions were prepared in glycine buffer (0.1 M glycine/NaOH buffer pH8.6, containing 1 mM of MgCl2 and 1 mM of ZnCl2), while the compounds were dissolved in DMSO. After the addition of MgCl2 and ZnCl2, the pH of the glycine buffer was 8.3. Black polypropylene 96-well plates obtained from Milian (Geneva, Switzerland) were filled with 166 μL of ALP solution (2 mU/mL in buffer). 4 μL of the DMSO solution of compounds under study or 4 μL of DMSO for the controls were added. After a pre-incubation for 15 minutes at 40 ± 0.1°C with a continuous shaking at 150 rpm in a Heidolph Titramax 1000 incubator (Geneva, Switzerland), 10 μL of AAPH solution (final concentration of 5 mM) were added to wells containing test samples and oxidized controls. The oxidation, i.e. the incubation with AAPH, was followed for 90 min at 40 ± 0.1°C (Heidolph Titramax 1000 incubator, shaking at 150 rpm). During the incubations, the 96-well plates were sealed with elastomer 96-well cap mats to avoid sample evaporation. After the oxidation, the cap mats were removed and the samples were cooled down to room temperature (5 min). For the non-oxidized controls, i.e. to obtain the ALP activity in the absence of oxidation, 10 μL of AAPH solution (final concentration 5 mM) were added just before the addition in every well of 20 μL of substrate solution (MUP) diluted in buffer at a final concentration of 5 μM. The plates were shaken for 45 seconds and the fluorescence (λEx 360 ± 20 nm, λEm 460 ± 20 nm) was monitored at room temperature for 15 min to obtain the hydrolytic activity of oxidized samples (hasample) and oxidized controls (haox) as well as of non-oxidized controls (hanon ox).

Chapter 2

The percentage of ALP protection by tested compounds was calculated from the intensity of the fluorescence’s signal according to Equation (2.1):

ha− ha ( sample ox ) % ALP protective capacity = 100 x (2.1) ()hanon ox− ha ox

The reference compounds were tested at three concentrations, namely 2.10-5 M, 10-5 M, and 10-6 M in order to find the appropriate concentration for the screening. A limit value necessary to highlight the potential active compounds, namely those able to protect the protein, was defined thanks to the screening values of the reference compounds as well as their pEC50 value. The latter represents the negative logarithm of the concentration of antioxidant compound necessary to protect 50% of ALP activity; it was determined by a dose-response curve.

-5 Then, the 146 compounds were screened at 10 M and the pEC50 value was calculated for the active compounds and for some inactive compounds. These latter were assessed to validate the screening concentration and the limit value. The presence of a possible inhibition of compounds on the ALP activity was checked for all compounds at a concentration of 10-5 M by replacing the oxidant by glycine buffer. α-tocopherol was not used for this assay as it was not soluble in these conditions.

2-2.2.2 ORAC assay

This well-known assay was used to determine the involvement of the ALP protein in the reaction between the peroxyl radicals and the antioxidant compound in the ALP assay. If an antioxidant compound possesses the same activity in both assays, a direct interaction between this one and radicals might be hypothesized. However, if the antioxidant activity differs, the protein might be involved in the protective effect of the compound.

Chapter 2

In order to compare both assays, the ORAC assay was carried out with the same conditions as the ALP assay. The only difference was that 186 μL of a fluorescein solution at 6.10-8 M was used instead of both the solutions of protein and substrate. The fluorescence was read at λEx 485 ± 20 nm, λEm 528 ± 20 nm to obtain the fluorescence value of oxidized samples (Fluosample) and oxidized controls (Fluoox) as well as non-oxidized controls (Fluonon-ox). The more efficient the antioxidant activity of a compound, the higher the percentage of unaltered fluorescein is. This last one was calculated by Equation (2.2):

Fluo− Fluo ( sample ox ) % remaining fluorescein = 100 x (2.2) ()Fluonon ox− Fluo ox

The screening was carried out with respect to the concentrations set in the ALP assay. For each compound with a percentage of fluorescein protection upper than the limit value defined by the ALP assay and some ones under it, the pEC50 value was determined by a dose-response curve. This parameter represents the concentration of antioxidant compound necessary to avoid 50% of fluorescein oxidation, namely a fluorescence loss of 50%. α-tocopherol was not used for this assay as it was not soluble in these conditions.

2-2.2.3 DPPH• assay

The capacity of compounds to scavenge the stable radical 2,2-diphenyl-1- picrylhydrazyl (DPPH•) was determined spectrophotometrically by measuring the color change of DPPH•. The purple radical becomes yellow DPPH-H upon reduction by test compounds16. The solutions of DPPH• and compounds under study were prepared in absolute ethanol (99%) from purchased products for reference compounds and from stock solutions in DMSO for some compounds

• • from the set. The initial concentration of DPPH (CDPPH initial) in the wells was

Chapter 2

-5 • • 8.6.10 M. A calibration curve (for example AbsDPPH = 7867 x [DPPH ] – 8.24.10-4) was carried out to verify the exact initial concentration. Clear polystyrene flat bottom 96-well microplates, obtained from Milian (Geneva, Switzerland) were filled with 35 μL of test compound solution or ethanol for the DPPH• blank. The blank was run to estimate spontaneous DPPH• degradation during the time of measurement. The reaction was initiated by addition of 215 μL of DPPH• solution. The plates were covered by clear polystyrene lids and shaken for 2 seconds, the decrease in absorbance (λ = 515 nm) was monitored at room temperature every 20 seconds for 90 min for the determination of ER50 values.

The parameter ER50 is obtained by a dose-response curve; it is the ratio of the concentration of the compound over the radical concentration necessary to reduce 50% of radical. To validate the screening parameters, the set was screened at a ratio of 0.5 for 10 minutes; the absorbance was recorded every 12 seconds. The scavenging activity of active and some inactive compounds was assessed by using ten concentrations to obtain the ER50 parameter.

2-2.2.4 ABTS•– assay

The free radical scavenging ability was evaluated spectrophotometrically by the ABTS•– decolourization11, the green radical becomes clear after reduction. To form the ABTS•– radical, a ABTS water solution of 7 mM final concentration was added to a potassium persulfate water solution (final concentration of 2.45 mM). The mixture was stocked in the dark at room temperature for 16h before use. In order to know the exact ABTS•– radical concentration, an aliquot of ABTS•– radical stock solution was diluted 60 times in water and 215 μl of this diluted solution was added to 35 μl of water in a clear polystyrene flat bottom 96-well microplate, obtained from Milian (Geneva, Switzerland). The absorbance was measured at λ = 734 nm and the exact concentration was calculated thanks to the Beer-Lambert law using Equation (2.3):

Chapter 2

Abs ⎡⎤ABTS i− = (2.3) ⎣⎦εl

4 ε stands for the molar extinction coefficient in water that is ε734 nm = 1.5 x 10 cm.M-1 11 and l is the measured path length of 0.8 cm. The 60-time dilution allows an absorbance about 0.8 AU that is 67 μM. For the determination of screening parameters, namely the reaction time and the appropriate ratio of concentrations of test compound over ABTS•–, the

ER50 value, representing the efficient ratio of the antioxidant compound over radical concentration producing an absorbance decrease of 50%, was determined for each reference compound thanks to a dose-response curve. This value was obtained by mixing 215 μl of stock ABTS•– radical solution diluted 60 time in absolute ethanol (99%) added to 35 μl of the ethanolic compound solution at ten concentrations or 35 μl of ethanol for the blank, which was used to estimate the radical degradation during the reaction. Microplates were covered by clear polystyrene lids, obtain from Milian (Geneva, Switzerland) and were shaken for 2 seconds before each measure. The decrease of absorbance was recorded at λ = 755 nm every 20 seconds for 90 min. Once the screening parameters defined, the 146-compound set was screened at a ratio of 0.2 for 10 minutes with an absorbance recorded every 12 seconds. The activity of the active and some inactive compounds was then assessed by using ten concentrations to obtain the value of the ER50 parameter.

For a significant comparison, a reaction time of 90 min for the DPPH• and ABTS•– assays was chosen with respect to the oxidation time in the ALP assay.

Chapter 2

2-3 Results and discussion

2-3.1 Screening parameters

2-3.1.1 ALP assay

The concentration, which will be used to screen new chemical entities, needs to be determined attentively. An appropriate screening concentration will lead to the emergence of the potential antioxidant compounds. The ability of reference compounds to protect the effectiveness of the protein has been assessed in the ALP assay. It is described by the potency parameter that is the pEC50. The higher the pEC50 is, the more potent the compound is. The potency parameter for each reference compound is displayed in Table 2-1.

Quercetin (pEC50 = 5.95) has exhibited the highest potency in protecting the protein against the peroxyl radical-induced oxidation. As concerns the inactive compounds, glutathione and mannitol have been unable to protect ALP. In Table 2-1, the protective activity of a compound toward the AAPH-induced oxidation of the protein has a similar meaning than the remaining activity of ALP.

Chapter 2

Table 2-1: Antioxidant activity and screening values of the reference compound assessed in ALP assay. (na: not active, nd: not determined, the broken lines define the classes of potency.)

Compounds pEC50 % protective activity at 10 μM

Quercetin 5.95 ± 0.04 91 ± 8

Mangiferin 5.74 ± 0.17 100 ± 9

Resveratrol 5.72 ± 0.03 82 ± 4

Chlorogenic acid 5.68 ± 0.03 100 ± 7

Caffeic acid 5.66 ± 0.10 100 ± 6

Gallic acid 5.31 ± 0.03 100 ± 2

Melatonin 5.20 ± 0.12 50 ± 7

Uric acid 4.89 ± 0.05 29 ± 9

Trolox® 4.85 ± 0.14 34 ± 11

Ascorbic acid 4.51 ± 0.17 16 ± 6

Glutathione na 14 ± 6

Mannitol na 4 ± 4

Afterwards, the reference compounds have been assessed at three concentrations in order to define a screening concentration. Figure 2-6 summarizes the percentage of remaining activity of ALP at compound concentrations of 20 μM (black columns), 10 μM (grey columns), and 1 μM (white columns). The higher the percentage of remaining activity is, the more potent the antioxidant compound is.

The reference compounds are ranked by decreasing pEC50 value in the Figure 2-6.

Chapter 2

Figure 2-6: Protective activity of ALP by reference compounds at a concentration of 20 μM (black columns), 10 μM (grey columns), and 1 μM (white columns) against AAPH- induced damage. The antioxidants are ranked by decreasing pEC50.

It has been reported that quercetin, mangiferin, resveratrol, chlorogenic acid, caffeic acid, and glutathione inhibit the enzyme activity at 25 μM4. Quercetin is the most influent, the ALP activity decreases by 35% in its presence while it decreases by 24% in the presence of mangiferin. Consequently, lower concentrations have been tested. The inhibition of the protein has been less significant at a 20-μM concentration; quercetin has induced an activity loss of 20% when 8% of ALP activity has been inhibited by mangiferin (data not shown). Therefore, this concentration is still too high to differentiate an oxidation-induced activity loss from a compound-induced protein inhibition. Furthermore, a 20-μM concentration is also too high to specifically describe the compound activity since melatonin and uric acid have shown a protective capacity as high as the most potent antioxidant compounds while their pEC50 value is lower (5.20 and 4.89, respectively, Table 2-1). On the contrary, a concentration of 1 μM has not been discriminatory enough. The percentage range of remaining activity has been contracted; it varies from 47% (with quercetin as a protective compound) to 4%

Chapter 2

(in presence of glutathione). Therefore, the range is not significant enough to provide accurate data regarding the potency of the compounds. Indeed, melatonin seems to be as potent as caffeic acid and chlorogenic acid while its pEC50 is lower. A concentration of 10 μM is then the most adequate since the range of remaining activity percentage is large (between 100% and 4%) and the compounds did not act as ALP inhibitors (data not shown). A screening limit value has been chosen in order to highlight the potential antioxidants from the inactive compounds. The entities exhibiting a screening value above the limit value will be further assessed to obtain the pEC50 value. When considering the protective capacity of ascorbic acid and glutathione, the screening boundary value has been set at 15% of remaining activity. Indeed, the pEC50 value of ascorbic acid equals 4.5 and it has protected 16% of ALP activity at a 10-μM concentration. On the other hand, glutathione has protected 14% of protein activity during the screening and it did not exhibit a higher capacity at higher concentrations. Therefore, no potency parameter could be determined for glutathione.

Before the screening, the effectiveness of ALP has been checked in presence of the 146 compounds at a concentration of 10 μM. None of them has been revealed as protein inhibitor. Thus their antioxidant properties have been assessed in the ALP method. The 146-compound set has been screened at a concentration of 10 μM for the purpose of validating the screening parameters.

Chapter 2

Table 2-2: Antioxidant activity and screening values of some of the 146 compound assessed in ALP assay. (na: not active, nd: not determined, the broken lines define the classes of potency.)

% protective % protective

Compounds pEC50 activity at Compounds pEC50 activity at 10 μM 10 μM

SO-IV-561 5.80 ± 0.27 100 ± 4 CB286 na 16 ± 7

SO-I-39 5.69 ± 0.20 100 ± 9 MC2387 na 14 ± 3

CB290 5.62 ± 0.03 100 ± 1 CB303 na 13 ± 1

SO-II-233 5.61 ± 0.13 55 ±8 CB301 na 13 ± 3

MC2385 5.41 ± 0.13 47 ± 2 VRL-43 na 11 ± 3

ML48 5.22 ± 0.04 58 ± 1 MC2415 na 10 ± 4

MC2416 5.14 ± 0.05 22 ± 2 VRL-58 nd 4 ± 4

CB392 5.00 ± 0.14 46 ± 3 VRL-31 nd 2 ± 2

ML45 4.83 ± 0.02 43 ± 6 VRL-40 nd 2 ± 2

CB393 4.57 ± 0.13 30 ± 2

The percentages confirm that the screening concentration is appropriate for a good discrimination (Table 2-2). The potency parameter has been determined for the compounds which have induced a remaining activity above 15%. The pEC50 is rather in line with the percentage obtained from the screening. Indeed, SO-IV-561 and SO-I-39 have allowed the protein ALP to efficiently catalyze the dephosphorylation of MUP despite the peroxyl radical-induced oxidation during the screening and have been characterized by a high pEC50. ML48 has been able to protect 58% and has shown a pEC50 value of 5.22. Furthermore, ML45 and CB393 have been capable of protecting 43% and 30% of protein activity, respectively, and have exhibited a low potency value after an in-depth assessment (4.83 and 4.57, respectively). The potency parameter has also been determined for some compounds which have exhibited a percentage of protective activity below the limit value. None of them have revealed a protective capacity at higher concentrations. Despite the presence of a inactive compound above 15% (CB286), the percentage limit value (15%) of protein protection is an appropriate boundary to uncover the potential active compounds. 57

Chapter 2

A concentration of 10 μM has been shown as an adequate screening concentration which involves a strong discrimination based on the protective capacity of the screened compounds. 15% of remaining activity efficiently leads to the emergence of the potential compounds capable of protecting the ALP protein against oxidation-induced enzymatic effectiveness loss. Indeed, when using these screening parameters, the set of 146 compounds has been screened and antioxidant compounds toward peroxyl radicals have been uncovered.

2-3.1.2 ORAC assay

In order to define the involvement of the protein in the ALP assay and the interactions between the protein, the substrate, the oxidant, and the antioxidant compounds, the ORAC assay was carried out using the same conditions as ALP assay as well as the same concentration of compounds for the screening that is 10

μM. The 146-compound set has been screened and the pEC50 value has been determined for the emerging potentially active compounds, namely exhibiting a percentage higher than 15% of capacity for protecting fluorescein, as well as for some compounds under the limit value (Table 2-3). The percentage of unaltered fluorescence has a similar meaning than the percentage of protective activity of compounds against the oxidation-induced fluorescence loss.

Chapter 2

Table 2-3: Antioxidant activity and screening values of some of the 146 compounds assessed in ORAC assay. (na: not active, nd: not determined, the broken lines define the classes of potency.)

% unaltered % unaltered

Compounds pEC50 fluorescein at Compounds pEC50 fluorescein at 10 μM 10 μM

Mangiferin 6.04 ± 0.06 89 ± 10 MC2387 na 18 ± 3

Quercetin 5.78 ± 0.02 95 ± 3 SO-II-233 na 14 ± 8

Caffeic acid 5.65 ± 0.08 89 ± 11 MC2415 na 10 ± 2

Chlorogenic 5.65 ± 0.08 89 ± 11 CB286 nd 3 ± 2 acid

Resveratrol 5.57 ± 0.04 90 ± 3 Glutathione na 3 ± 1

SO-IV-561 5.52 ± 0.02 75 ± 11 CB301 nd 2 ± 1

CB290 5.49 ± 0.04 76 ± 11 VRL-40 nd 3 ± 2

SO-I-39 5.34 ± 0.12 73 ± 1 VRL-31 nd 3 ± 1

Gallic acid 5.29 ± 0.06 85 ± 14 VRL-58 nd 3 ± 1

Melatonin 5.28 ± 0.05 51 ± 11 VRL-43 nd 2 ± 1

MC2385 4.74 ± 0.05 40 ± 6 Uric acid na 1 ± 1

Trolox 4.74 ± 0.12 15 ± 4 Ascorbic acid na 1 ± 1

ML45 4.61 ± 0.07 19 ± 4 Mannitol na 1 ± 1

CB393 4.56 ± 0.06 17 ± 5 CB303 nd 1 ± 1

CB392 4.55 ± 0.09 18 ± 2

First of all, the concentration for screening compounds in ORAC assay has been discriminative enough, as it was in ALP method; the range of the protective activity of the tested compounds is included between 95% and 0%. Moreover, the percentage of protective activity is in agreement with the pEC50 value. The compounds, which have exhibited a high percentage during the screening, have been characterized by a high potency parameter. For instance, mangiferin has been potent at protecting fluorescein during the screening since 89% of fluorescence has remained and it has exhibited the highest pEC50. Moreover, the potency parameter is lower for the compounds which have shown a lower percentage during the screening, such as ML45 and CB393. The limit value of 15% is also suitable in this method. The compounds under this limit such as

Chapter 2

SO-II-233, MC2415, glutathione, and uric acid did not show any protective capacity at higher concentrations.

The screening parameters defined in ALP assay are also appropriate in the ORAC assay, potential active compounds may be revealed thanks to the screening.

2-3.1.3 DPPH• assay

DPPH• radical is used in a well-known method, which readily provides the free radical-scavenging capacity of new entities16,21,22. The scavenging ability of reference compounds has been assessed toward

• the DPPH radical at different concentrations in order to obtain their ER50 value after 90 minutes. All of them have reached a steady state (annex I, figure I-1). An

ER50 value has been calculated at 5 and 10 minutes in order to define the shortest required time and the most relevant ratio for the screening (Table 2-4).

Chapter 2

Table 2-4: DPPH•-scavenging potency of reference compounds at 5, 10, and 90 minutes, and after the screening (ratio 0.5 for 10 min). (na: not active)

% DPPH• Reference ER50 at 5 min ER50 at 10 min ER50 at 90 min scavenged at ratio compound 0.5 for 10 minutes

Gallic acid 0.14 ± 0.03 0.08 ± 0.01 0.04 ± 0.01 83 ± 1

Quercetin 1.36 ± 0.42 0.21 ± 0.03 0.09 ± 0.01 66 ± 3

Chlorogenic acid 1.64 ± 0.27 0.41 ± 0.03 0.11 ± 0.03 53 ± 6

Mangiferin 1.15 ± 0.40 0.44 ± 0.09 0.12 ± 0.06 40 ± 3

Caffeic acid 0.22 ± 0.01 0.19 ± 0.01 0.17 ± 0.01 95 ± 1

α-tocopherol 0.20 ± 0.03 0.19 ± 0.02 0.20 ± 0.03 94 ± 1

Trolox 0.24 ± 0.01 0.22 ± 0.01 0.21 ± 0.02 97 ± 1

Ascorbic acid 0.28 ± 0.08 0.28 ± 0.08 0.28 ± 0.08 88 ± 8

Resveratrol 1.59 ± 0.24 1.04 ± 0.17 0.46 ± 0.08 34 ± 1

Uric acid na na na 2 ± 1

Glutathione na na na 1 ± 1

Mannitol na na na 1 ± 1

Melatonin na na na 1 ± 1

After a 5-minutes reaction time, the most rapidly and the most slowly acting antioxidant compounds have emerged. Indeed, gallic acid, caffeic acid, α- tocopherol, trolox®, and ascorbic acid have already exhibited a potent scavenging ability which is similar to that at 90 minutes. On the contrary, quercetin, chlorogenic acid, and mangiferin have been characterized by a very high ER50 value, which is 1.36, 1.64, and 1.15, respectively, while their DPPH• scavenging activity is significantly more potent at 90 min. They are slowly acting scavengers. The potency parameter for these compounds is too high and not representative enough to provide accurate data after only a reaction time of 5 minutes. Indeed, the ER50 value of resveratrol has the same order of magnitude as the slowly acting compounds while it is less potent at 90 minutes. At 10 minutes, the potency parameter of quercetin, chlorogenic acid, and mangiferin reflects the scavenging capacity better than that at 5 minutes. Moreover, the polyphenols have shown a more potent activity than that of

Chapter 2 resveratrol while at 5 minutes they were almost similar. Despite the data loss which concerns the kinetics, 10 minutes is the most appropriate time for the screening. It is the shortest time which provides an accurate perception of the DPPH• scavenging activity. In order to define the most useful ratio, which will provide data concerning potency and kinetics parameters, the compounds have been screened at ratio 0.5 for 10 minutes (Figure 2-7).

Figure 2-7: Decay of DPPH• absorbance due to the scavenging ability of reference compounds during the screening at a ratio 0.5 for 10 minutes.

The percentages of DPPH• scavenged by each compound at the end of the screening have been calculated. They are summarized in the Table 2-4. The percentages are well-shared and they cover a range from 97% (ascorbic acid) to 1% (glutathione, mannitol, and melatonin). The ratio allows the discrimination between the DPPH• scavengers and the inactive compounds toward this radical,

Chapter 2 namely glutathione, mannitol, melatonin, and uric acid among the reference compounds. The appearance of the curves, displayed by plotting the absorbance of DPPH• as function of time, provides kinetic data. Indeed, ascorbic acid, trolox® caffeic acid, and α-tocopherol have reached a steady state at the end of the screening. As concerns gallic acid, quercetin, chlorogenic acid, mangiferin, and resveratrol, the curve which displays their scavenging ability was still decreasing at the end of 10 minutes. These results indicate that ascorbic acid, trolox®, caffeic acid, and α-tocopherol are more rapidly acting scavengers than the other polyphenols. Therefore, a 10-minute screening at ratio 0.5 leads to the emergence of DPPH• scavengers and provides data concerning the scavenging rate. The curve slope reflects well the kinetics rate for 90 minutes (annex I, figure I-1). However, the screening of reference compounds have not provided an accurate boundary value since no compound has exhibited a percentage of scavenged DPPH• between 1 and 34%. The reference compounds have also been assessed at a lower ratio which was 0.25, but the kinetic information has been less accurate than that obtained with a ratio of 0.5. The most rapidly acting compounds exhibited the same behavior than the most slowly acting compounds (Figure 2-8). The ratio 0.25 has therefore been rejected.

Chapter 2

Figure 2-8: DPPH•-scavenging activity of reference compounds at ratio 0.25 (black columns) and 0.5 (white columns).

To validate the screening parameters, the set of compounds has been screened at a ratio of 0.5 for 10 minutes. The Table 2-5 summarizes the percentage of DPPH• reduced at the end of the screening and the potency parameter for the active and for some inactive compounds. Any compound from the set has not exhibited a DPPH• scavenging ability neither during the screening nor after a 90-minute reaction time. Since the compound MC2385 was able to scavenge 15% of DPPH• and it has been considered as an inactive compound, a limit value of 15% has been set. The compounds able to scavenge more than 15% of radical will be assessed in an in-depth study in order to obtain their potency parameter that is ER50.

Chapter 2

Table 2-5 : Parameters of DPPH•-scavenging capacity and screening parameters of reference compounds and some compounds from the set. (na: not active, nd: not determined, the broken lines define the classes of potency.)

% DPPH• % DPPH•

ER50 at 90 scavenged at ER50 at 90 scavenged at Compound Compound minutes ratio 0.5 for 10 minutes ratio 0.5 for 10 minutes minutes

Gallic acid 0.04 ± 0.01 83 ± 1 CB301 nd 2 ± 2

Quercetin 0.09 ± 0.01 66 ± 3 SO-II-233 nd 2 ± 2

Chlorogenic 0.11 ± 0.03 53 ± 6 SO-IV-561 nd 2 ± 2 acid

Mangiferin 0.12 ± 0.06 40 ± 3 SO-I-39 nd 2 ± 3

Caffeic acid 0.17 ± 0.01 95 ± 1 CB303 nd 2 ± 2

Trolox 0.18 ± 0.02 97 ± 1 VRL-31 nd 3 ± 3

α-tocopherol 0.20 ± 0.03 94 ± 1 VRL-40 nd 3 ± 3

Ascorbic acid 0.28 ± 0.08 88 ± 8 Uric acid na 2 ± 0

Resveratrol 0.46 ± 0.08 34 ± 1 VRL-43 nd 2 ± 2

MC2385 na 15 ± 7 CB290 nd 1 ± 2

MC2415 na 11 ± 1 VRL-58 nd 1 ± 1

MC2387 na 10 ± 1 Glutathione na 0 ± 0

CB286 nd 6 ± 1 Mannitol na 0 ± 0

CB393 nd 5 ± 2 Melatonin na 0 ± 0

ML45 nd 3 ± 0

A ratio of 0.5 during 10 minutes allows the distinction between active and inactive compounds. Furthermore, the slope of the absorbance decrease and the percentage of reduced DPPH• provide data which concern the kinetics and the potency of the scavengers.

Chapter 2

2-3.1.4 ABTS•– assay

The shorter the reaction time is, the higher the screening throughput is.

Therefore, in order to find a short screening time, the ER50 for each reference compound was calculated at 5, 10, and 90 minutes (Table 2-6).

Table 2-6: ABTS•– -scavenging potency of reference compounds at 5, 10, and 90 minutes, and after the screening (ratio 0.2 for 10 min). (na: not active)

% of ABTS•– Reference ER50 at 5 min ER50 at 10 min ER50 at 90 min scavenged at ratio compound 0.2 for 10 minutes

Gallic acid 0.08 ± 0.01 0.08 ± 0.01 0.06 ± 0.01 84 ± 2

Quercetin 0.10 ± 0.02 0.09 ± 0.01 0.07 ± 0.01 82 ± 1

Resveratrol 0.23 ± 0.08 0.18 ± 0.04 0.12 ± 0.02 50 ± 1

Mangiferin 0.14 ± 0.01 0.13 ± 0.01 0.11 ± 0.01 71 ± 1

Glutathione 0.23 ± 0.02 0.19 ± 0.01 0.17 ± 0.01 50 ± 2

Caffeic acid 0.23 ± 0.01 0.22 ± 0.02 0.17 ± 0.03 51 ± 2

Uric acid 0.33 ± 0.03 0.32 ± 0.04 0.25 ± 0.02 29 ± 1

α-tocopherol 0.27 ± 0.03 0.27 ± 0.03 0.26 ± 0.01 37 ± 2

Chlorogenic acid 0.28 ± 0.03 0.28 ± 0.02 0.26 ± 0.01 37 ± 3

Trolox 0.29 ± 0.02 0.29 ± 0.01 0.28 ± 0.04 40 ± 1

Ascorbic acid 0.30 ± 0.03 0.31 ± 0.05 0.30 ± 0.03 29 ± 5

Melatonin 0.90 ± 0.01 0.69 ± 0.05 0.32 ± 0.07 18 ± 1

Mannitol na na na 1 ± 1

Except for melatonin and resveratrol, the ER50 value of the compounds is nearly the same at the three times. The ER50 value of melatonin, which decreased from 0.9 at 5 minutes to 0.32 at 90 minutes, has shown a very slow scavenging activity. Resveratrol has had the same behavior but with a weaker variation. The kinetics aspect is given by the appearance of the absorbance decrease as function of time (Figure 2-9). A decreasing curve indicates that the potency of the compound will be higher after 90 minutes, while a curve, which has already reached a steady state during the screening, means that the compound will not 66

Chapter 2 scavenge any more radical molecules. The reaction time of 5 minutes is too short to provide an accurate perception regarding the appearance of the absorbance decrease. A compound with a slow activity such as melatonin could be considered as an inactive scavenger and be rejected. Consequently, the reaction time of 10 minutes is more appropriate.

Figure 2-9: Decay of ABTS•– absorbance due to the scavenging ability of reference compounds during the screening at a ratio 0.2 for 10 minutes..

The ER50 values of reference compounds at 10 minutes fluctuated between 0.08 for gallic acid and 0.32 for melatonin. Thus, the percentages of reduced

•– ABTS for a ratio of 0.2, situated closed to the middle of the ER50 value range, have been calculated (Table 2-7).

Chapter 2

Table 2-7: Parameters of ABTS•– -scavenging capacity and screening parameters of reference compounds and some compounds from the set. (na: not active, nd: not determined, the broken lines define the classes of potency.)

% ABTS•– % ABTS•

ER50 at 90 scavenged at ER50 at 90 scavenged at Compounds Compounds minutes ratio 0.2 for 10 minutes ratio 0.2 for 10 minutes minutes

Gallic acid 0.06 ± 0.01 84 ± 2 VRL-40 0.56 ± 0.05 13 ± 5

Quercetin 0.07 ± 0.01 82 ± 1 SO-IV-561 0.60 ± 0.02 26 ± 2

Mangiferin 0.11 ± 0.01 71 ± 1 VRL-31 0.65 ± 0.07 13 ± 3

Resveratrol 0.12 ± 0.02 50 ± 1 ML45 0.82 ± 0.13 17 ± 2

SO-I-39 0.13 ± 0.01 33 ± 2 CB393 na 10 ± 1

Glutathione 0.17 ± 0.01 50 ± 2 CB290 na 10 ± 4

Caffeic acid 0.17 ± 0.03 51 ± 2 MC2415 na 8 ± 1

α-tocopherol 0.26 ± 0.01 37 ± 2 CB303 na 6 ± 6

Uric acid 0.25 ± 0.02 29 ± 1 CB301 na 6 ± 3

Chlorogenic 0.27 ± 0.01 37 ± 3 MC2385 nd 5 ± 1 acid

VRL-43 0.27 ± 0.03 17 ± 2 CB286 nd 3 ± 2

Trolox 0.28 ± 0.04 40 ± 1 SO-II-233 nd 3 ± 1

Ascorbic acid 0.30 ± 0.03 29 ± 5 MC2387 nd 2 ± 1

Melatonin 0.32 ± 0.07 18 ± 1 Mannitol na 1 ± 1

VRL-58 0.50 ± 0.05 13 ± 1

The scavenged ABTS•– percentages are well-distributed between 84% for gallic acid and 18% for melatonin, and are representative of the ER50 value. Gallic acid and quercetin have exhibited the most potent scavenging activity during the screening with percentages of scavenged ABTS•– of 84% and 82%, respectively.

They have also shown the highest ER50 values after a more accurate assessment, which is 0.06 and 0.07, respectively. The compounds with a screening percentage around 50%, have exhibited an ER50 close to 0.2, such as glutathione (50%, ER50

= 0.17) and caffeic acid (51%, ER50 = 0.17). Furthermore, the curve appearance of some compounds, such as melatonin, indicates that they have not reached a

Chapter 2 steady state yet, contrary to ascorbic acid, which will not be capable of scavenging further ABTS•– molecules. Therefore, a ratio of 0.2 for 10 minutes provides useful data concerning the kinetics and the scavenging capacity. A first limit value of 15% has been set according to the melatonin scavenging capacity. The compounds able to scavenge more than 15% of ABTS•– radical are considered as active compounds and their activity is further assessed in order to determined their potency parameter. To validate these parameters, a screening of the set of compounds has been carried out. The percentage values of scavenged ABTS•– are summarized in the Table 2-7. The potential ABTS•– scavengers have been further assessed along with eight compounds below the screening limit value to confirm their inactivity.

•– The percentages of scavenged ABTS are rather in line with the ER50 value. However, VRL-31, VRL-40, and VRL-58, which scavenged 13% of ABTS•– after a 10-minute-reaction time, have exhibited an ER50 value of 0.65, 0.56, and 0.50, respectively, after 90 minutes. The screening boundary value for in-depth investigations has then been adjusted at 10%. The screening parameters for performing the ABTS•– methods have been chosen to provide the most accurate data at the shortest reaction time. A ratio of 0.2 for 10 minutes allows reaching this aim. The potential ABTS•– scavengers are capable of reducing at least 10% of radical. They will be then further investigated.

Compounds have been reported as more rapidly acting scavengers toward ABTS•– than they do toward DPPH•, surely due to the weaker stability of the radical ABTS•– relative to that of DPPH•23. That would explain that the ratio used for the screening is lower in ABTS•– assay than in DPPH• assay.

Chapter 2

2-3.2 Ranking of the compounds into potency classes

To more easily distinguish the antioxidant properties, the compounds have been ranked into classes. These classes have been defined thanks to the parameters obtained from the four assays. Since both ALP and ORAC assays are carried out using the same conditions and the same screening parameters, similar classes have been defined for both assays arbitrarily. Four classes (delimited by a broken line in Table 2-1, Table 2-2, and Table 2-3) have been established depending on the protective activity.

The class labeled class 3 contains the potent compounds, described by a pEC50 above 5.5. The class 2 includes the intermediate compounds (5 < pEC50 ≤ 5.5) when the poor compounds are put together into the class 1 (4.5 < pEC50 ≤ 5).

Finally the inactive compounds form the class 0 with a pEC50 lower or equal at 4.5 or with a screening percentage of protective activity of ALP and fluorescein below 15%. The compounds have also been ranked thanks to the spectroscopic assays, namely DPPH• and ABTS•– assays. Two parameters have been highlighted in these methods, which are the potency and the kinetics parameters. The values of scavenging activity have emphasized four classes of compounds based on the reducing power (delimited by a broken line in Table 2-5 and Table 2-7). Indeed, a compound with an ER50 value below 0.2 has been tagged as potent (class 3) while it is intermediate (class 2) if its ER50 value was higher or equal to 0.2 and below

0.5. Finally a poor compound (class 1) is described by an ER50 value higher or equal to 0.5 and it is inactive (class 0) if its scavenge capacity was below 15% and 10% for the screening in DPPH• and ABTS•– assays, respectively. As concerns the kinetics classes, two graphs, which display the percentage of scavenged DPPH• for the screening ratio of 0.5 versus the percentage of scavenged ABTS•– for the screening ratio of 0.2, have been plotted. The first one (Figure 2-10) has been done from the percentages at 30 seconds while the second one (Figure 2-11) has been done with the values at 60 seconds of the free radical- scavenging reaction.

Chapter 2

Figure 2-10: Percentages of free radical scavenged by the active compounds from the set at 30 seconds in DPPH• and ABTS•– methods.

Figure 2-11: Percentage of free radicals scavenged by the active compounds from the set at 60 seconds in DPPH• and ABTS•– methods.

Chapter 2

Firstly, the distribution of the compounds on both graphs shows that four classes may be defined. Their boundaries are from 100% to 75% of scavenged radical for the rapidly reducing compounds (class 3), 75% to 50% for the compounds acting with a moderate rate (class 2), 50 to 25% for the slowly reducing compounds and 25 to 0% for the very slow or inactive compounds (class 0). Then, according to the DPPH•-related percentages, 30 seconds is a too short time to provide an accurate view of the kinetics. Indeed, no compound may be used to refer to the intermediate class (from 75% to 50%) because none of them scavenges between around 45 to 70% of DPPH• radical. The distribution of the compounds from the DPPH• percentage values is better after a 60-second reaction. Trolox®, α-tocopherol, and caffeic acid appear more rapidly acting DPPH• scavengers than gallic acid, which are in line with the screening (Figure 2-7) and may be used to represent the intermediate compound class. The three other groups also refer to the reference compounds; the rapidly acting compounds include ascorbic acid, gallic acid is among the slowly reducing compounds and finally the last group with the other reference compounds being very slowly acting DPPH• scavengers. The compounds may already be classified so after the screening.

Chapter 2

Table 2-8: Boundaries of potency classes and kinetics classes. (The kinetics describes the percentage of radical scavenged in 60 seconds.)

Potency Kinetics ALP / ORAC methods DPPH• / ABTS•– methods DPPH• / ABTS•– methods

Class 3 Class 3 Potent compounds Rapidly scavenging compounds

5.5 < pEC50 ER50 < 0.2 75% ≤ % scavenged radical < 100% Class 2 Class 2 Intermediate compounds Moderately scavenging compounds

5 < pEC50 ≤ 5.5 0.2 ≤ ER50 < 0.5 50% ≤ % scavenged radical < 75% Class 1 Class 1 Poor compounds Slowly scavenging compounds

4.5 < pEC50 ≤ 5 0.5 ≤ ER50 25 ≤ % scavenged radical < 50% Class 0 Class 0 Inactive compounds Very slowly scavenging compounds

pEC50 ≤ 4.5 Screening: Screening: < 15% (DPPH• assay) % scavenged radical < 25% < 15% < 10% (ABTS•– assay)

Chapter 2

2-3.3 Similarity of ALP assay with ORAC assay

6.5

6.0

5.5

5.0 obtained by the ORAC method 50 SO-II-233 pEC 4.5 4.5 5.0 5.5 6.0 6.5

pEC50 obtained by the ALP method

Figure 2-12: Correlation between ALP and ORAC assays. (r2 = 0.58, slope = 0.89 ± 0.09, rs = 0.78, P value < 0.0001) (The dotted lines define the classes.)

Overall, the potency values are higher in ALP method than those obtained in ORAC method (Figure 2-12). This could be explained by the experimental conditions, which were chosen to carry out the method using the protein effectively. These conditions were then used in ORAC assay without any adjustment. In the ALP assay, how an antioxidant compound protects the protein from peroxyl radical-induced activity loss is unclear. Besides the radical detoxification by the antioxidant compound, the latter may also repair ALP damage or bond with the protein at ROO•-attacked sites and undergo the oxidation instead of the protein. As regards ORAC assay, two mechanisms may occur. The peroxyl radical-oxidized fluorescein is reduced by the antioxidant leading to the regeneration of fluorescein, which maintains the fluorescence level high. Trolox® has been reported as a fluoresceinyl radical reducing agent24,25. The antioxidant 74

Chapter 2 compound may also be oxidized by ROO• radicals before the fluorescein oxidation that delays the alteration of the fluorescent compound. The moderate correlation between ALP and ORAC assays (r2 = 0.58) indicates that some mechanisms involved in both methods are different. Although most compounds belong to the same class in both assays, some of them are more potent at protecting the protein than fluorescein; SO-II-233 was active in the ALP assay while it did not show any antioxidant activity in the ORAC assay. It has to be noted that the dot representing the aurone-derivative SO-II-233 has been put on the abscissa axis in order to highlight it on the graph, but the compound did not exhibit any pEC50 value. The compounds, which have exhibited a higher potency class in ALP assay than that in ORAC assay, are surely involved in protein binding hindering the protein oxidation or undergoing the oxidation themselves. This binding does obviously not take place at the active site, since the MUP dephosphorylation occurs, but elsewhere on the protein. Furthermore, the peroxyl radical-induced damage induces conformational changes leading to the ALP activity loss while the protein-binding antioxidant does not affect the conformation. The moderate

Spearman’s rank correlation coefficient (rs = 0.78) supports the distinct behavior in both assays since the potency rank of some compounds has been switched between the methods. These results emphasized the relevant interest in collecting antioxidant property-related information from both methods.

2-4 Conclusion

The screening parameters, namely the concentration of the compounds under study in the ALP and ORAC assay as well as the reaction time and the ratio of concentrations in the DPPH• and ABTS•– assays have been defined thanks to the reference antioxidants and confirmed with a set of 146 compounds. A concentration of 10 μM has been shown to be discriminative enough for a screening carried out in both ALP and ORAC assays. With the same conditions, the ALP assay is more sensitive than the ORAC assay. Both are complementary and provide the involvement of the protein in the reaction with both antioxidant

Chapter 2 and oxidant compounds. A 10-minute reaction time allows the emergence of potential DPPH• and ABTS•– scavenging compounds by using a screening ratio of 0.5 and 0.2, respectively. Limit values have been fixed at 15% in ALP, ORAC, and DPPH• assays, and at 10% in ABTS•– assay. They are useful to highlight the potentially active compounds, which will be further studied, from the inactive compounds.

Classes have been defined thanks to the pEC50 (ALP and ORAC assay) and

• •– ER50 (DPPH and ABTS assays) values to range the compounds depending on their activity against radicals. Furthermore, classes regarding the DPPH• and ABTS•– scavenging kinetics have also been characterized. They are based on the activity of the compound in reducing radical in a minute. The antioxidant activity of the 146-compound set as well as the reference compounds emphasizes that the four assays are complementary to characterize antioxidant properties. Indeed, the antioxidant power may differ depending on the radical-induced oxidation. These four assays will be used afterwards to screen a larger set of compounds in order to find new antioxidant entities. The compounds revealed as new antioxidants will be ranked into the classes, which will then be used to yield a hierarchical classification.

Chapter 2

Bibliographic references

1. Adams, C. P.; Brantner, V. V. Estimating the cost of new drug development: is it really $802 million? Health Aff 2006, 25, 420-428. 2. DiMasi, J. A.; Hansen, R. W.; Grabowski, H. G. The price of innovation: new estimates of drug development costs. J. Health Econ. 2003, 22, 151-185. 3. Agresti, J. J.; Antipov, E.; Abate, A. R.; Ahn, K.; Rowat, A. C.; Baret, J. C.; Marquez, M.; Klibanov, A. M.; Griffiths, A. D.; Weitz, D. A. Ultrahigh- throughput screening in drop-based microfluidics for directed evolution. Proc. Natl. Acad. Sci. 2010, 107, 4004-4009. 4. Bertolini, F.; Novaroli, L.; Carrupt, P. A.; Reist, M. Novel screening assay for antioxidant protection against peroxyl radical-induced loss of protein function. J. Pharm. Sci. 2007, 96, 2931-2944. 5. Levine, R. L.; Garland, D.; Oliver, C. N.; Amici, A.; Climent, I.; Lenz, A. G.; Ahn, B. W.; Shaltiel, S.; Stadtman, E. R. Determination of carbonyl content in oxidatively modified proteins. Methods Enzymol. 1994, 186, 464-479. 6. Meucci, E.; Mordente, A.; Martorana, G. E. Metal-catalyzed oxidation of human serum albumin: conformational and functional changes. J. Biol. Chem. 1991, 266, 4692-4699. 7. Hunt, J. V.; Simpson, J. A.; Dean, R. T. Hydroperoxide-mediated fragmentation of proteins. Biochem. J. 1988, 250, 87-93. 8. Halliwell, B. Oxidative stress, nutrition and health. Experimental strategies for optimization of nutritional antioxidant intake in humans. Free Rad. Res. 1996, 25, 57-74. 9. Huang, D.; Ou, B.; Hampsch-Woodill, M.; Flanagan, J. A.; Prior, R. L. High- Throughput Assay of Oxygen Radical Absorbance Capacity (ORAC) Using a Multichannel Liquid Handling System Coupled with a Microplate Fluorescence Reader in 96-Well Format. J. Agric. Food Chem. 2002, 50, 4437-4444. 10. Ancerewicz, J.; Migliavacca, E.; Carrupt, P. A.; Testa, B.; Brée, F.; Zini, R.; Tillement, J. P.; Labidalle, S.; Guyot, D.; Chauvet-Monges, A. M.; Crevat, A.; Le Ridant, A. Structure-property relationships of trimetazidine derivatives and model compounds as potential antioxidants. Free Rad. Biol. Med. 1998, 25, 113-120. 11. Re, R.; Pellegrini, N.; Proteggente, A.; Pannala, A.; Yang, M.; Rice-Evans, C. Antioxidant activity applying and improved abts radical cation decolorization assay. Free Rad. Biol. Med. 1999, 26, 1231-1237. 12. Chegaev, K.; Lazzarato, L.; Rolando, B.; Marini, E.; Tosco, P.; Cena, C.; Fruttero, R.; Bertolini, F.; Reist, M.; Carrupt, P. A.; Lucini, V.; Fraschini, F.; Gasco, A. No-donor melatonin derivatives: synthesis and in vitro pharmacological characterization. J. Pineal Res. 2007, 42, 371-385. 13. Yanez, C.; Lopez-Alarcon, C.; Camargo, C.; Valenzuela, V.; Squella, J. A.; Nunez- Vergara, L. J. Structural effects on the reactivity 1,4-dihydropyridines with alkylperoxyl radicals and ABTS radical cation. Bioorg. Med. Chem. 2004, 12, 2459-2468. 14. Leopoldini, M.; Marino, T.; Russo, N.; Toscano, M. Antioxidant properties of phenolic compounds: H-atom versus electron transfer mechanism. J. Phys. Chem. A 2004, 108, 4916-4922. 15. Meister, A. On the antioxidant effects of ascorbic acid and glutathione. Biochem. Pharmacol. 1992, 44, 1905-1915.

Chapter 2

16. Brand-Williams, W.; Cuvelier, M. E.; Berset, C. Use of a free radical method to evaluate antioxidant activity. Food Sci. Technol. 1995, 28, 25-30. 17. Hayes, J. D.; McLellan, L. I. Glutathione and glutathione-dependent enzymes represent a co-ordinately regulated defence against oxidative stress. Free Rad. Res. 1999, 31, 273-300. 18. Formica, J. V.; Regelson, W. Review of the biology of quercetin and related bioflavonoids. Fd Chem. Toxic. 1995, 33, 1061-1080. 19. Kocha, T.; Yamaguchi, M.; Ohtaki, H.; Fukuda, T.; Aoyagi, T. Hydrogen peroxide- mediated degradation of protein: different oxidation modes of copper- and iron-dependent hydroxyl radicals on the degradation of albumin. Biochim. Biophys. Acta 1997, 1337, 319-326. 20. Halliwell, B. Uric acid: an example of antioxidant evaluation. Antioxid. Health Dis. 1996, 3, 243-256. 21. Martin, F.; Hay, A. E.; Cressend, D.; Reist, M.; Vivas, L.; Gupta, M. P.; Carrupt, P. A.; Hostettmann, K. Antioxidant C-Glucosylxanthones from the Leaves of Arrabidaea patellifera. J. Nat. Prod. 2008, 71, 1887-1890. 22. Martin, F.; Hay, A. E.; Quinteros Condoretty, V. R.; Cressend, D.; Reist, M.; Gupta, M. P.; Carrupt, P. A.; Hostettmann, K. Antioxidant phenylethanoid glycosides and a neolignan from Jacaranda caucana. J. Nat. Prod. 2009, 72, 852-856. 23. Perez-Jimenez, J.; Saura-Calixto, F. Anti-oxidant capacity of dietary polyphenols determined by ABTS assay: a kinetic expression of the results. Int. J. Food Sci. Technol. 2008, 43, 185-191. 24. Bisby, R. H.; Brooke, R.; Navaratnam, S. Effect of antioxidant oxidation potential in the oxygen radical absorption capacity (ORAC) assay. Food Chem. 2008, 108, 1002-1007. 25. Huang, D.; Ou, B.; Prior, R. L. The chemistry behind antioxidant capacity assays. J. Agric. Food Chem. 2005, 53, 1841-1856.

Chapter 3

Chapter 3 Antioxidant profiles as a result of clustering analyses

3-1 Introduction – How to perform a cluster analysis

The cluster analysis is used in a broad spectrum of fields such as biology, marketing, medicine, Internet… Such diversified fields have connections with this method since it contributes to gather data according to similar features. As a result, an Internet search engine provides query-related web pages and a symptom checklist leads to a diagnosis1-3. To perform a cluster analysis in this study, a set of compounds and related variables are required. The clustering analysis gathers the compounds which have similar characteristics provided by the variables.

Figure 3-1: Compound-related data and associated clusters.

Many clustering procedures exist to analyze the data set but each of them requires three parameters to be defined. These parameters are the clustering

Chapter 3 algorithm, the type of clusters, and the distances between two compounds within the cluster. The algorithm choice depends on how large the data set is and the results expected from the clustering. Hierarchical clustering and k-means are among the clustering algorithms. The second parameter is the type of cluster that is the distance between two clusters; the types differ in the variables of the compounds. For instance, the clusters are well-separated if a compound exists in only one group and they are overlapped if a compound may reside within two groups. Finally, the distances between compounds belonging to the same cluster are computed using a metric such as Euclidean, Manhattan or Hamming distances. The smaller the distance, the more the compounds will be alike. Due to the large choice of procedures, only the procedure used in this study will be further described. Hierarchical cluster analysis (HCA) is the appropriate algorithm for a small set of compounds. Moreover, it quantifies how far apart the compounds are. This compound distribution in the space is the expected result in this investigation. Two approaches are possible for generating the hierarchical clustering (Figure 3-2). As concerns the agglomerative HCA (bottom-up), each compound is an individual cluster at the outset. The closest clusters merge at each analysis step until they reach a single cluster. Unlike this approach, the divisive HCA (top- down) leads the set of compounds, which forms a cluster unto itself, to be divided into many clusters at each analysis step. The compounds form a cluster on their own at the final step.

Chapter 3

Figure 3-2: Agglomerative (left) and divisive (right) cluster analysis steps.

The type of cluster defines the distance between two clusters and therefore the steps which two clusters will merge at. Ward’s method is one of cluster type techniques where the means of all variables of each cluster are calculated along with the error for each compound. The squared errors are then summed. The error is the distance of the compound value relative to the mean. In Ward’s method, the proximity between two clusters is defined as the smallest increase in the overall sum of squared errors when the two clusters are merged. The distance between two compounds within a cluster is computed using Euclidean distance, which is the sum of the squared differences over all of the variables. The results from the cluster analysis are displayed using a tree-like diagram, called dendrogram (Figure 3-3).

Chapter 3

Figure 3-3: Dendrogram.

A set of 177 compounds has been analyzed by clustering to form groups characterized by the scavenging rate on the one hand and by the radical-reducing ability on the other hand. A HCA has lead to a formation of clusters thanks to the potency and the kinetics assessed in the DPPH• and ABTS•– methods. The variables required to provide the reducing capacity-related cluster analysis are the potency values obtained using ALP, ORAC, DPPH•, and ABTS•– assays. The resulting groups of compounds from both clustering have been then gathered, that has yielded subgroups. These subgroups are antioxidant profiles characterized by the antioxidant properties of both cluster analyses. A strategy based on a well-ordered assays procedure has allowed the profiles to be reached. The antioxidant properties of trolox®-derivatives have been used as an application of this strategy. They have been assessed according the well-ordered method procedure to reach antioxidant profiles.

Chapter 3

3-2 Materials and methods

3-2.1 Materials

The 177-compound set was composed by aurone- and azaaurones-derivatives prepared in the group of Prof. Ahcene Boumendjel (Département de Pharmacologie Moléculaire. Grenoble, France), pyrrolidine-3,4-diol- and conduramine-derivatives synthesized in the group of Prof. Pierre Vogel by Dr Robert Lysek (Ecole Polytechnique Fédérale de Lausanne, Switzerland), pyrazolo[4,3-b]pyrrolizine-derivatives synthesized in the group of Prof. Antonello Mai (Dipartimento di Studi Farmaceutici. Università degli Studi di Roma, Italy), xanthone-derivatives either synthesized in the group of Prof. Madalena Pinto (Laboratorio de Quimica Organica. Faculdade de Farmacia. Porto, Portugal) 4 or extracted from Chironia Kresbii in the group of Prof. Kurt Hostettmann (Laboratoire de pharmacognosie et phytochimie. Université de Genève-Lausanne. Geneva, Switzerland), and mangiferin-derivatives extracted from the Leaves of Arrabidaea patellifera, phenylethanoid glycoside as well as neolignan extracted from Jacaranda caucana and isolated in the group of Prof. Kurt Hostettmann by Dr Frédéric Martin (Laboratoire de pharmacognosie et phytochimie. Université de Genève-Lausanne. Geneva, Switzerland)5,6. Phenol, aniline, and cysteine have been added as representative compounds to describe the antioxidant profiles. They were purchased from Fluka (Buchs, Switzerland). The structures and the antioxidant properties of the compounds are displayed in annex II. The trolox®-derivatives (Figure 3-9) used to validate the antioxidant assay pathway were synthesized by the group of Prof. Ahcene Boumendjel (Département de Pharmacologie Moléculaire. Grenoble, France). The group of parameters was analyzed by cluster analysis using Tsar Version 3.3 (Oxford Molecular Hunt valley Maryland USA). The other compounds, devices and software are described in the section Materials in Chapter 2 (2-2.1).

Chapter 3

3-2.2 Methods

3-2.2.1 Antioxidant potency toward the three radicals

The 177-compound set was screened using the four assays and the parameters defined in Chapter 2. Only the compounds detected as potential antioxidant were investigated in in-depth studies in order to determine their potency parameters, namely the pEC50 (from the ALP and ORAC assays) and

• •– ER50 values (from the DPPH and ABTS assays). The procedures of each assay are described in the section 2-2.2 Methods of the chapter 2. The pEC50 values, the

ER50 values, and the percentage of scavenged radical after 60 seconds during the screening in DPPH• and ABTS•– assays were then ranked into the classes previously defined.

3-2.2.2 Cluster analysis

In order to emphasize the stable radical scavenging rate and the reductive potency toward the three radicals, two cluster analyses were performed. The first one was done from the ER50 values and the percentage of scavenged radical, both ranked into the classes. Its name is the kinetics cluster analysis. The second one was done from the potency classes, namely from the pEC50 and the ER50 values. The resulting analysis is named potency cluster analysis. Both analyses based on Ward’s method and Euclidean distances were performed using Tsar Version 3.3 (Oxford Molecular Hunt valley Maryland USA). They were not standardized by the mean. The kinetics cluster analysis underlines the ability as well as the kinetics of the compound to scavenge the free radical in both DPPH• and ABTS•– assays while the potency cluster analysis emphasizes the reducing power toward the three radicals in the four assays.

Chapter 3

3-3 Results and discussion – Formation of antioxidant profiles

Only the 85 compounds exhibiting an antioxidant activity in at least one assay have been considered for the cluster analyses, the inactive compounds in all assays have been removed.

3-3.1 Kinetics cluster analysis

To further characterize the antioxidant properties, both potency and kinetic parameters obtained from DPPH• and ABTS•– assays have been analyzed by clustering. This one, the kinetics cluster analysis (Figure 3-4), has been done from the ER50 class values (from 3 to 0 for the potent compounds to the inactive ones, respectively) and from the percentage of scavenged free radical class values (from 3 to 0 referring to the rapidly to the very slowly acting radical scavengers, respectively).

Chapter 3

Figure 3-4 : Dendrogram obtained from the kinetics cluster analysis, based on Ward'method and Euclidien distances

Five clusters, labeled A, B, C, D, and E, have been formed thanks to the cluster analysis. The characteristics and the related reference compound are summarized in Table 3-1.

Chapter 3

Table 3-1 : Class values and reference antioxidant compounds represented in the kinetics clusters.

DPPH• DPPH• ABTS•– ABTS•–

efficiency kinetics efficiency kinetics Cluster A Ascorbic acid, Caffeic acid, 3 - 2 3 - 2 3 - 2 2 - 1 (10 compounds) Trolox Chlorogenic acid, Cysteine, Cluster B Gallic acid, Mangiferin, 3 - 1 1 - 0 3 - 2 2 - 1 (15 compounds) Quercetin, Resveratrol Cluster C Glutathione, Melatonin, 0 0 3 - 2 1 - 0 (28 compounds) Phenol, Uric acid Cluster D - 0 0 1 0 (13 compounds) Cluster E Aniline 0 0 0 1 - 0 (19 compounds)

The compounds which belong to the group A are potent and intermediate DPPH• and ABTS•– scavengers. They exhibit a fast or moderate DPPH• scavenging rate and reduce the ABTS•– radical with a moderate or slow activity rate. Ascorbic acid, caffeic acid, α-tocopherol, and trolox® are the antioxidant reference compounds included in this group (Table 3-1). The cluster B which contains chlorogenic acid, gallic acid, mangiferin quercetin, and resveratrol differs from the first cluster in the DPPH•-scavenging kinetics; the reduction of the radical is slow or very slow. The three other groups contain the non-DPPH• radical scavengers. Group C includes glutathione, melatonin, uric acid, and phenol. It gathers the potent and intermediate compounds against ABTS•– radical while group D is made up of the poor ABTS•– radical scavengers. None of reference compounds is related to this cluster. Aniline and the other compounds from group E fail in both ABTS•– and DPPH• assays.

Chapter 3

DPPH • screening

DPPH • Not DPPH • scavengers scavengers

Fast, Slow, intermediate very slow ABTS assay Cluster A Cluster B Efficient, Poor Inactive intermediate

Cluster C Cluster D Cluster E

Figure 3-5 : Antioxidant assay pathway leading to the clusters related to the DPPH• and ABTS•– scavenging kinetics.

Many procedures to carry out the two methods following a well-defined order are possible to split the five clusters up. One of them is the ABTS•– assay performed at the outset, which leads to put apart the cluster D with the poor ABTS•– scavengers from the cluster E containing the non-ABTS•– scavengers and from the clusters A, B, and C. Then these three groups may be divided thanks to their DPPH• kinetics behavior. Another procedure is to carry out the DPPH• assay as a first method. The cluster A and B differ from the other ones in the scavenging potency. Afterward, the DPPH• kinetics capacity may be considered to split the clusters A and B, and the ABTS•– assay may be performed to separate the clusters C, D, and E according to their ABTS•– scavenging potency. However, the most convenient assay order is to perform the DPPH• screening at the outset (Figure 3-5). It provides not only information on the kinetics, but also on the scavenging capacity at 90 minutes. Indeed, in this context, the 10-minute screening emphasizes the kinetics activity and uncovers the DPPH• scavengers while 90 minutes are required to obtain the potency against the DPPH• and ABTS•– radicals. Therefore the cluster A is detected thanks to the reaction rate. The ability to scavenge more than 15% of DPPH• after 10 minutes leads to Cluster B. It contains the DPPH• scavengers whereas the compounds of the clusters C, D, and E are inactive against this radical. Then, the ABTS•– assay allows the differentiation between the three last groups (Figure 3-5) since the compounds from Cluster C are potent or intermediate while those from Cluster D

Chapter 3 are poor toward the ABTS•– radical. Cluster E contains the inactive compounds against DPPH• and ABTS•–. The ABTS•– kinetics activity-related data are not specific enough for a first discrimination between the five clusters.

3-3.2 Potency cluster analysis

To emphasize the radical-related antioxidant behavior of the compounds, the potency values ranked into classes have been analyzed by cluster analysis. Eight groups, labeled from 1 to 8, have been formed from the potency cluster analysis (Figure 3-6). Their characteristics toward the radicals are summarized in the Table 3-2. Except for Cluster 3, at least one reference antioxidant compound is included in each cluster.

Figure 3-6: Dendrogram obtained from the potency cluster analysis, based on Ward's method and Euclidien distances.

Chapter 3

Table 3-2 : Class values and reference antioxidant compounds represented in the potency clusters.

ALP ORAC DPPH• ABTS•– Cluster 1 Chlorogenic acid, Gallic 3 - 2 3 - 2 3 - 2 3 - 2 (11 compounds) acid, Resveratrol Cluster 2 Caffeic acid, Mangiferin, 3 3 3 3 (11 compounds) Quercetin Cluster 3 - 3 - 1 1 - 0 0 0 (11 compounds) Cluster 4 Aniline 3 - 2 3 - 1 0 0 (8 compounds) Cluster 5 Melatonin, Phenol 3 - 2 3 - 2 0 3 - 1 (10 compounds) Cluster 6 Glutathione 0 0 0 3 - 1 (17 compounds) Cluster 7 Uric acid 2 - 0 1 - 0 0 2 - 1 (14 compounds) Cluster 8 Ascorbic acid, Cysteine, 1 1 - 0 2 - 1 3 - 2 (3 compounds) Trolox®

Chlorogenic acid, resveratrol, and gallic acid are among the compounds of Cluster 1. They exhibited a potent or an intermediate activity in the four assays but the reducing potency of the compounds has been ranked in the class 2 in at least one method. Unlike this group, Cluster 2 is formed by the most potent compounds; their activity was characterized by the classes 3 in each assay. The reference antioxidant compounds belonging to this group are caffeic acid, mangiferin, and quercetin. Although a few compounds included in Cluster 3 showed a poor activity in the ORAC assay, this group is constituted by the compounds only capable of protecting the ALP protein with a potency level from class 3 (potent) to class 1 (poor). It is the only group which is not reference antioxidant compound-related. The aniline-like compounds constitute the cluster 4. They solely exhibited a peroxyl radical-detoxifying ability, with an activity level ranging between 3 and 1. The cluster 5 contains the melatonin- and phenol- like compounds. Their peroxyl radical-detoxifying capacity was potent or intermediate and they failed to scavenge the DPPH• radical. The ABTS•– radical scavenging capacity is not well-defined, the compounds are ABTS•– scavengers but this activity ranges from class 3 (potent) to class 1 (poor). Unlike Cluster 4, the glutathione-like compounds (cluster 6) are only ABTS•– radical scavengers; their capacity extends from a great potency to a poor activity. The compounds 90

Chapter 3

involved in the cluster 7, such as uric acid, showed an intermediate or poor activity in the ABTS•– assay and an activity extent toward the ROO• radicals from the class 2 to the class 0. They are not DPPH• radical scavengers. The remaining group put together the ABTS•– and DPPH• radical scavengers with a poor or lacking activity toward the peroxyl radicals. The acid ascorbic-, trolox®-, and cysteine-like compounds are ranked in this cluster since their activity toward the peroxyl radicals is not as high as those belonging to the clusters 1 and 2, and their ABTS•– and DPPH• scavenging potency is higher than that of clusters 3 to 7. The potency characteristics of this group are then to be DPPH• and ABTS•– scavengers and detoxify the ROO• radical poorly or even not. Except Cluster 8 which contains only 3 compounds, the other compounds are well-shared out between the seven groups. Thanks to the specificities of each cluster, an antioxidant assay pathway may be designed (Figure 3-7). It is useful to discern each cluster according to the radical scavenging capacity of the compounds.

DPPH • screening

DPPH• scavengers Not DPPH• scavengers

ORAC or ORAC assay Class 1 - 0 Cluster 8 Class 2 Class 3 ALP assay Class 1 Class 0 -2 Class 3 Class 2 ABTS•-assay ABTS•- assay ABTS•- assay Cluster 1 ALP, ORAC or Class 2 Class 3 ABTS•- assays Class 3 - 1 Class 0 Class 0 Class 0 -1 -1 Class 3 Class 3 Class 2 Class 3 Cluster 5 Cluster 4 -2 ALP Cluster 3 ALP assay ABTS•-, ALP Class 1 Cluster 2 assay Class 2 Class 0 or ORAC -1 assays Cluster 7 Cluster 6

Figure 3-7 : Antioxidant assay pathway leading to the clusters related to the antioxidant potency. (The colored arrows are described in the text.)

Chapter 3

Indeed, the DPPH• screening is an adequate method to split the clusters into two groups, namely the DPPH• radical scavengers, which are the compounds belonging to the clusters 1, 2, and 8, and those DPPH• scavengers

• which are not able to scavenge the DPPH radical, ORAC or Class 1 - 0 Cluster 8 Class 2 ALP assay namely the groups 3, 4, 5, 6, and 7. Among the Class 3 Class 2 Cluster 1 ALP, ORAC or DPPH• scavengers, the ascorbic acid-like ABTS•- assays Class 3 Class 2 Class 3 compounds turned out to be inactive against the ABTS•-, ALP Cluster 2 or ORAC peroxyl radicals. Therefore, the cluster 8 is put assays apart from the clusters 1 and 2 by carrying out either the ALP assay or the ORAC assay. The clusters 1 and 2 may then be disjoined if an intermediate activity is revealed in the DPPH• or ORAC assay (blue solid arrow on Figure 3-7); hence the cluster 1 is detected. However, if a potent activity is uncovered, so a third or even the fourth assay is necessary as long as an intermediate activity does not appear, which means that the compound belongs to the cluster 1. Therefore, the cluster 2 is only detected after the four assays are carried out.

The ORAC assay leads the non-DPPH• radical scavengers to be split into three groups of clusters (green arrows). Indeed, the potent and intermediate compounds reside in the clusters 4 and 5, the poor peroxyl radical scavengers may belong to the clusters 3, 4, and 7 and the inactive compounds are shared into the clusters 3, 6, and 7.

The compounds which emerge as potent and Not DPPH• scavengers intermediate antioxidants from the ORAC assay (green ORAC assay Class 3 solid arrow on Figure 3-7), can reach the clusters 4 or 5 -2 ABTS•-assay thanks to a third method. Indeed, they differ from each Class 3 - 1 Class 0 other in their ABTS•– scavenging potency; the phenol- Cluster 5 Cluster 4 like compounds have the ability to scavenge the ABTS•– radical (cluster 5), while the aniline-like compounds do not (cluster 4).

Chapter 3

One of the compounds of the cluster 4 showed Not DPPH• scavengers a poor activity in the ORAC assay as some ORAC assay compounds of the cluster 3 and 7 did (green dotted Class 1 ABTS•- assay •– Class 2 arrow on Figure 3-7). The ABTS method is then Class 0 -1 Class 3 Cluster 4 -2 Cluster 3 useful to rank these compounds into the three ALP Class 1 assay •– groups. The intermediate and poor ABTS Cluster 7 scavengers reach the cluster 7 (purple dotted arrow on Figure 3-7) while a further assay is required to discern the inactive compounds toward this free radical (purple solid arrow on Figure 3-7). The ALP method allows the distribution into two groups. The cluster 3 contains the poor ALP protectors (purple dashed arrow on Figure 3-7) while the cluster 4 contains the potent and intermediate protein protectors (purple dotted-dashed arrow on Figure 3-7).

Regarding the compounds failing in the ORAC Not DPPH• scavengers assay (green dashed arrow on Figure 3-7), they need ORAC assay Class 0 to be assessed by a further method. Since the ABTS•- assay Class 3 •– Class 0 compounds belonging to the cluster 3 are not ABTS -1 Cluster 3 ALP assay radical scavengers, contrary to the clusters 6 and 7, Class 2 Class 0 -1 the ABTS•– assay can be used in order to disjoin the Cluster 7 Cluster 6 group 3 from the glutathione- and uric acid-related clusters. Finally, the ALP assay leads to the ranking of the compounds capable of protecting the ALP protein, namely those of the cluster 6 and the inactive ones, namely the glutathione-like compounds (cluster 7).

3-3.3 Antioxidant profiles

Both kinetics and potency cluster analyses have been gathered yielding combinations of clusters. Each combination provides an antioxidant profile which is labeled by the combination of the letter from the kinetics clustering (A to E) and the number from the potency cluster analysis (1 to 8).

Chapter 3

Table 3-3 displays the characteristics of the profiles and the reference compounds-related. The number written in the cells is the number of compounds qualified by the profile and the cells with a cross match to a not possible case. The compounds included in the cluster combinations are detailed in annex I table I-1.

Table 3-3 : Characteristics of antioxidant profiles. (The number is the number of compounds characterized by the profile.)

Kinetics C D E A B Slow and Slow and poor Not ABTS•– Fast DPPH• Slow DPPH• potent ABTS•– ABTS•– and DPPH• scavengers scavengers Potency scavengers scavengers scavengers Chlorogenic 1 acid Potent compounds 2 Gallic acid in all assays Resveratrol 9 2 Mangiferin The most potent Caffeic acid Quercetin compounds in all 6 5 assays 3 ALP protecting 11 compounds 4 Aniline Peroxyl radicals 8 scavengers 5 Melatonin Potent ROO• and Phenol 1 ABTS•– radicals 9 6 Glutathione ABTS•– 5 12 scavengers 7 Intermediate or Uric acid poor ROO• and 7 7 ABTS•– scavengers 8 ABTS•– and Ascorbic acid Cysteine DPPH• Trolox 1 scavengers, poor 2 ROO• scavengers

Most of the antioxidant profiles are related to a well-known compound. Caffeic acid is representative of the profile A2. It is a rapidly acting DPPH• scavenger

Chapter 3 and a very potent compound against the three radicals. It only differs from the mangiferin- and quercetin-like compounds (B2) in the DPPH• scavenging kinetics. The compounds characterized by the antioxidant profile B1, such as chlorogenic acid, gallic acid, and resveratrol are slower than the caffeic acid-like compounds to scavenge the DPPH• radical and less potent compounds than those included in the cluster combination B2 but with the same order of magnitude as concerns the kinetics. The profiles A8 and B8 depict a drastically poorer activity than that of the profiles A2 and B2 as regards the peroxyl radicals and cysteine even shows a weaker DPPH• scavenging activity. The cluster combination A8 provides characteristics for this antioxidant profile which are a rapid DPPH• scavenging rate, an ability to scavenge the ABTS•– radical and a weak ROO• radical detoxifying capacity. The cysteine antioxidant profile is characterized by the same potency activities as the profile A8 but the DPPH• scavenging rate is slower than that of A8. The melatonin- and phenol-like compounds (C5) are described by a potent or an intermediate activity toward the ROO• (in both methods ALP and ORAC) and ABTS•– radicals unlike the uric acid-like compounds (C7) which are poorly active or inactive in the ORAC assay. The ALP protective capacity of the profile C7 is intermediate or poor, and the ABTS•– scavenging ability is intermediate. The antioxidant profile C6 differs from all other profiles in its potent and intermediate capacity to solely trap the ATBS•– radical. Glutathione is included in this cluster combination. The compound, which forms the profile D5 on its own, owes its difference to its greater activity toward the peroxyl radicals than that of the compounds of the other profiles from the cluster D and to its poorer ABTS•– radical trapping ability than that of the groups included in the cluster C. As regards the five compounds characterized by the profile D6, they differ from the glutathione-like compounds in their poorer ABTS•– scavenging ability and from the other cluster combinations in the column D in their lack of activity toward the peroxyl radicals. The antioxidant profile E3 defines compounds which solely protect the protein ALP while the profile E4 concerns the peroxyl radical detoxifying compounds. Both profiles share the inability to scavenge the DPPH• and ABTS•– radicals.

Chapter 3

Each antioxidant profile provides specific characteristics either on the potency or on the kinetics toward the radicals or both.

3-4 Antioxidant assay procedure as a strategy to reach the antioxidant profiles

A procedure of well-ordered assays may be useful to rapidly describe the antioxidant properties of a new entity thanks to an antioxidant profile related to a reference compound.

According to both kinetics and potency cluster analyses, the DPPH• assay appears as the most discriminative method. Indeed, the cluster A is readily detected thanks to its rapid DPPH• scavenging kinetics (Table 3-1) and the cluster 1, 2, and 8 differ from the other ones in their DPPH• scavenging potency (Table 3-2). Furthermore, the procedure defined from the kinetics cluster analysis is more convenient to sort the clusters than that from the potency cluster analysis (Figure 3-5 and Figure 3-7, respectively). As a result, the 10- minute screening in the DPPH• method as a first assay yields the most relevant and discriminative outcomes. The rapidly acting compounds (kinetics classes 3 and 2) are ranked in Cluster A while a slow DPPH• scavenging rate leads to Cluster B. The very slow DPPH• scavengers may belong to the clusters B, C, D, or E. The cluster B is then distinguished by an activity to scavenge the DPPH• radical while the compounds from the clusters C, D, and E are inactive. If a compound turns out to be a DPPH• scavenger after the screening, it is ranked in the cluster B. If it is inactive against the radical, it belongs to a cluster among the clusters C, D, or E. Therefore, the cluster B is differentiated. Finally, the three remaining groups, namely C, D, and E, differ from each other in their potency to scavenge the ABTS•– radical. Indeed, the compounds from the cluster C exhibit a potent or an intermediate activity while the poor ABTS•– scavengers are ranked in the cluster D. Cluster E put together the compounds without any ABTS•– trapping capacity. Hence, the ABTS•– assay, carried out after the

Chapter 3 screening of the DPPH• method, leads the non DPPH• scavengers to be ranked into the three clusters thanks to the potency toward ABTS•– radical. The five clusters from the kinetics cluster analysis have now been differentiated (Figure 3-8).

Chapter 3

E3 or ALP assay Class 0 inactive inactive compound compound Mannitol-like Mannitol-like Class Class 3, 2, 1 3, 2, assay Class 1 ORAC D6 assay •– D7 assay Class 1 ORAC Class 1 Cluster C, D, E ABTS C5 D5 E4 E3 CD E assay ORAC ALP ALP assay C6, C7 C7 C6 • B8 DPPH screening scavengers • B1 to readily reach the antioxidant profiles. profiles. the antioxidant reach readily to Slow or very or Slow very slow B DPPH Class 2 assay ORAC Class 2 Class 2 •– B2 ALP assay assay Class 3 Class 3 B1, B2 B1, B2 ABTS A8 A1 A Class 2 assay ORAC Class 2 Class 2 •– Figure 3-8 : Well-ordered assay procedure assay procedure Well-ordered : Figure 3-8 A2 ALP assay Class 3 assay Class 3 A1, A2 A1, A2 ABTS

Chapter 3

The clusters A and B contain the potent compounds against the DPPH• radical. That means that all compounds included in the cluster A belong either to the clusters 1, 2, or 8. The same is true of the compounds included in the cluster B. These ones only differ from the compounds of Cluster A in the DPPH• scavenging kinetics rate. Therefore, the same pathway may be used to rank the compounds into the three groups (1, 2, and 8). To make the explanation clearer, only the cluster A will be considered herein.

The ORAC assay appears to be discriminatory enough to share out the DPPH• scavengers according to their antioxidant power. Indeed, some compounds among those of the cluster A have a poor or no activity against the peroxyl radicals. Consequently, thanks to the poor and null A capacity recorded in the ORAC assay, the compounds from ORAC assay Cluster 8 are distinguishable from those of the cluster 1 and 2; Class 2 A1, A2 A8 the antioxidant profile A8 is discerned. The ALP assay might ABTS•– assay Class 2 A1 be carried out to make this distribution, but the ORAC assay is Class 3 more appropriate since it is more convenient to perform. The A1, A2 ALP assay Class 2 compounds with a class-2 activity in this same method can also Class 3 be put apart at this step. Indeed they belong to the cluster 1 A2 since the cluster 2 only contains class-3 compounds (Table 3-2); as a result, they are defined by the antioxidant profile A1. A further antioxidant parameter is required to differentiate the compounds with a potent capacity (class 3) in the ORAC method. The scavenging potency assessed in the ABTS•– assay allows a ranking in Cluster 1 for the class-2 compounds. As regards the potent compounds, the results recorded in the ALP assay need to be taken into account to share them. As previously described, the compounds are included in Cluster 1 if they exhibit an intermediate activity or into Cluster 2 (A2) if a great potency is uncovered. Indeed, as long as an intermediate activity has not been recorded, the compounds are susceptible to belong to Cluster 2 (A2). This latter contains only the compounds with the most potent activity in the four assays while those from the cluster 1 possess an intermediate potency in at least one assay. Thus, thanks to this pathway, three subgroups have been formed within Cluster A namely A1,

Chapter 3

A2, A8. The same procedure is applicable to divide the cluster B into three subgroups, which leads to the antioxidant profiles B1, B2, and B8.

The clusters C, D, and E have been separated from the groups A and B by the DPPH• screening and then by the ABTS•– assay. The compounds from the clusters C and D are shared out among the clusters 5, 6, and 7. Only the clusters 3 and 4 are included in the cluster E.

As regards Cluster C, the parameters obtained from Cluster C, D, E the ORAC assay lead Cluster 5 (C5), with potent or ABTS•– assay intermediate compounds in this assay, to be detached C from the clusters 6 and 7 (classes 1 or 0). These latter are ORAC assay finally separated thanks to the results from the ALP C6, C7 C5 assay. The compounds from Cluster 7 (C7) have the ability to protect the protein contrary to those from ALP assay C7 C6 Cluster 6 (C6).

Regarding Cluster D, only the ORAC method provides the Cluster C, D, E compound differentiation. Indeed, the compounds from the cluster 5 ABTS•– assay are potent for the protection of fluorescein; therefore the profile D5 Class 1 D is highlighted. The clusters 6 and 7 are likewise set apart thanks to ORAC assay the poor activity (D7) and the inability (D6) to protect the Class 1 D7D5 D6 fluorescent compound.

Cluster C, D, E As concerns Cluster E, the compounds from Cluster ABTS•– assay 4 are put apart thanks to their potent and intermediate E ORAC activity in the ORAC assay (E4). The compounds from assay Class 1

Clusters 3 (E3) are highlighted since they exhibit a poor E4 E3 E3 or inactive activity in the ALP method. compound Class ALP 3, 2, 1 assay Class 0 Mannitol-like compound

100

Chapter 3

After the screening in DPPH• method, which has already been carried out, a screening may be done before each assay, using the parameters defined before. It highlights the active compounds and allows the removal of the inactive ones. This procedure saves an amount of product, since only one concentration is required, and an amount of time for the ABTS•– scavenging capacity (10 minutes instead of 90 minutes for the screening and the assay, respectively). The merger of both cluster analyses allows a more specific procedure relative to the procedure leading to the potency-related cluster (Figure 3-7) thanks to the contribution of the kinetics-related data.

3-4.1 Determination of the antioxidant properties of trolox®- derivatives as an application of the method

Trolox® is a well-known antioxidant capable of scavenging the DPPH•7 and ABTS•–8 radicals. However, its activity is poor against the peroxyl radicals9. To enhance the trolox® properties including the ROO• detoxification, new derivatives (Figure 3-9) have been synthesized. They have been assessed by following the described antioxidant assay procedure and then evaluated in each assay to confirm the antioxidant profile.

101

Chapter 3

O O O N O N H HO H N HO N

H 1 2 O O O OH N O N OH N HO O H HO O 3 4 O O O O O O OH O N S N O H O N O H HO H HO HO 6 5 7 Figure 3-9 : Structure of trolox®-derivatives.

The seven compounds have been screened against the DPPH• radical and the kinetic rate has been ranked in the defined classes. Compounds 1, 4, 5, 6, and 7 have turned out to be included into the kinetics class 1 (Table 3-4). Therefore, this ability leads them into the cluster B. The compounds 2 and 3 have exhibited a very slow DPPH• kinetics but the scavenged DPPH• percentages at the end of the screening has resulted in their ranking among the active compounds. Hence, both compounds have been included into the cluster B along with the other trolox®-derivatives. According to the procedure, the ORAC assay has been then carried out to assess their antioxidant potency toward the peroxyl radicals. Compounds 1 and 4 have exhibited an intermediate activity in protecting fluorescein against the peroxyl radical-induced oxidation; hence they are belonging to Cluster 1. Consequently, they have a chlorogenic acid-, gallic acid-, and resveratrol-like antioxidant profile (B1). Compounds 3, 5, 6, and 7 have exhibited a poor activity against the peroxyl radicals and compound 2 has been inactive. The behavior of these compounds in the DPPH• screening assay and in the ORAC method leads to the cluster combination B8. They are defined as cysteine-like compounds.

102

Chapter 3

Table 3-4 : Kinetics- and potency-related classes of trolox®-derivatives and some reference compounds in DPPH• and ORAC assays, respectively.

Screening in DPPH• assay ORAC assay % of Kinetics Potency Compounds scavenged pEC50 Profile class class DPPH• a

1 46 ± 2 1 5.29 ± 0.09 2 B1 2 18 ± 5 0 < 4.5 0 B8 3 24 ± 3 0 4.93 ± 0.02 1 B8 4 32 ± 3 1 5.43 ± 0.06 2 B1 5 27 ± 4 1 4.96 ± 0.22 1 B8 6 29 ± 0 1 4.82 ± 0.03 1 B8 7 42 ± 1 1 4.78 ± 0.03 1 B8 Chlorogenic 15 ± 2 0 5.65 ± 0.08 3 B1 acid Gallic acid 44 ± 7 1 5.29 ± 0.06 2 B1 Resveratrol 16 ± 4 0 5.57 ± 0.04 3 B1 Cysteine 19 ± 3 0 < 4.5 0 B1 Trolox® 59 ± 7 2 4.74 ± 0.12 1 A8 a: % of scavenged DPPH• in 60 seconds at a ratio of 0.5

To confirm these profiles, the trolox®-derivatives have been assessed in the DPPH•, ABTS•–, and ALP assays. Compounds 1 and 4 are supposed to be slow DPPH• scavengers; this has been confirmed by the DPPH• screening, and to be capable of interacting with the three radicals. The intermediate activity against the peroxyl radicals has already been uncovered in the ORAC assay. The outcomes from the last three assays have indicated that compounds 1 and 4 act as potent DPPH• scavengers, intermediate ABTS•– scavengers (class 2) and had an intermediate and a great protein protective potency. These results are in agreement with the chlorogenic acid profile (Table 3-5). As regards compounds 2, 3, 5, 6, and 7, the great DPPH• scavenging potency has been detected for each of these compounds. This high DPPH• trapping activity differs from that of cysteine, which is poor. According to the ER50 parameters (Table 3-5), about five times more cysteine is needed than the trolox®- derivatives. The opposite tendency is observed with the ABTS•– radical. Cysteine 103

Chapter 3 is about five-fold more active toward this radical than the derivatives. The poor peroxyl radical reduction by the derivatives is in agreement with that of cysteine. Compounds 2 and 3 are ranked in a higher class, but from the dose-response curves, they are unable to protect the ALP protein over 30% and 70%, respectively, even at high concentrations (data not shown). Thus, they have an intermediate protecting potency but the efficiency is intermediate and even low.

Table 3-5 : Potency and kinetics data of trolox®-derivatives and some reference compounds obtained in the DPPH•, ABTS•– and ALP methods.

DPPH• assay ABTS•– assay ALP assay

Potency % of Kinetics Potency Compounds ER50 scavenged ER50 pEC50 Class class class class ABTS•– a 0.12 ± 0.24 ± 5.29 ± 1 3 36 ± 4 1 2 2 0.02 0.02 0.11 0.18 ± 0.34 ± 5.50 ± 2 3 26 ± 1 1 2 2 0.02 0.04 0.05 b 0.13 ± 0.25 ± 5.19 ± 3 3 30 ± 2 1 2 2 0.01 0.04 0.10 b 0.14 ± 0.24 ± 5.70 ± 4 3 38 ± 1 1 2 3 0.02 0.06 0.15 0.16 ± 0.25 ± 4.87 ± 5 3 35 ± 1 1 2 1 0.02 0.03 0.08 0.15 ± 0.23 ± 4.89 ± 6 3 37 ± 1 1 2 1 0.03 0.03 0.06 0.16 ± 0.34 ± 4.91 ± 7 3 36 ± 2 1 2 1 0.03 0.04 0.04 Chlorogenic 0.11 ± 0.26 ± 5.68 ± 3 36 ± 2 1 2 3 acid 0.03 0.01 0.03 0.04 ± 0.06 ± 5.31 ± Gallic acid 3 71 ± 2 2 3 2 0.01 0.01 0.03 0.46 ± 0.12 ± 5.72 ± Resveratrol 2 38 ± 1 1 3 3 0.08 0.02 0.03 1.51 ± 0.07 ± 4.54 ± Cysteine 1 66 ± 5 2 3 1 0.14 0.01 0.08 0.21 ± 0.27 ± 4.85 ± Trolox® 2 39 ± 1 1 2 1 0.02 0.01 0.14 a: % of scavenged ABTS at 60 sec at ratio 0.2, b: the compounds 2 and 3 did not protect the ALP protein over a percentage of 30% and 70%, respectively

104

Chapter 3

The antioxidant pathway has lead these new entities to reach the antioxidant profiles thanks to only two assays, which are the screening of the DPPH• method and the ORAC assay. Compounds 1 and 4 have been well-ranked among the chlorogenic acid-like compounds. The rank of compounds 2, 3, 5, 6, and 7 among the cysteine-like compounds defined by the antioxidant profile B8 confirm that the main characteristics are a poor ROO• radical detoxifying ability and a capacity to scavenge the DPPH• and ABTS•– radicals with a slow rate toward the DPPH•. However, the antioxidant profile B8 does not provide a specific potency level toward the DPPH• and ABTS•– radicals. The antioxidant profiles of the trolox®-derivatives emphasize the slower DPPH• scavenging rate than that of trolox® and the greater ROO• detoxifying activity of compounds 1 and 4 relative to that of trolox®. All trolox®-derivatives have exhibited a slightly higher DPPH• scavenging activity than that of Trolox® and a similar ABTS•– trapping capacity.

3-5 Conclusion

A 177-compound set including diverse structures has been screened thanks to four methods using three radicals. The 85 active compounds have been further assessed in each assay and every potency or kinetics parameter has been ranked in four classes which extend from potent to null capacities. Cluster analyses of these classes have yielded antioxidant property-related groups. The potency cluster analysis has been made from the ability of the compound to reduce the peroxyl, DPPH• and ABTS•– radicals. On the other hand, the kinetics cluster analysis emphasizes the scavenging rate toward the stable radicals, namely the DPPH• and the ABTS•– radicals. Merging both cluster analyses has provided 14 antioxidant profiles characterized by antioxidant properties related to both the reducing potency toward the radicals and the kinetics to scavenge the stable radicals. Most of the profiles are well-known antioxidant compound-referred. A well-ordered assay procedure has been described to reach each profile. Only two assays, namely the screening in the DPPH• method and the ORAC assay are necessary to provide the characteristics of some profiles. The other ones need a 105

Chapter 3 third or the fourth assay. Thanks to this antioxidant assay pathway, a new entity may be related to an antioxidant profile. The antioxidant properties of trolox®-derivatives have been assessed as an application of the procedure. They have reached an antioxidant profiles thanks to the screening in the DPPH• assay and to the ORAC method. Two of them are defined by the chlorogenic acid-related antioxidant profile while the remaining derivatives are cysteine-like compounds. The antioxidant behavior in the last assays has confirmed the profile and has provided some specificity regarding the cysteine-related profile. The characteristics of the profile B8 are a DPPH• and ABTS•– scavenging activity and a poor ROO• detoxifying capacity. The addition of a substituent on the trolox® scaffold induces an antioxidant behavior which differs from that of trolox® since no trolox®-derivative has been described by the trolox®-related antioxidant profile.

106

Chapter 3

Bibliographic references

1. Chiu, C. Y.; Douglas, J.; Li, X. Cluster analysis for cognitive diagnosis: theory and applications. Psychometrika 2009, 74, 633-665. 2. Vogt, W.; Nagel, D. Cluster-analysis in diagnosis. Clin. Chem. 1992, 38, 182-198. 3. Zhang, K.; Chai, Y.; Yang, S. X. Self-organizing feature map for cluster analysis in multi-disease diagnosis. Expert Systems with Applications 2010, 37, 6359- 6367. 4. Gnerre, C.; Thull, U.; Gaillard, P.; Carrupt, P. A.; Testa, B.; Fernandes, E.; Silva, F.; Pito, M.; Pinto, M. M. M.; Wolfender, J. L.; Hostettmann, K.; Cruciani, G. Natural and synthetic xanthones as monoamine oxidase inhibitors: biological assay and 3D-QSAR. Helv. Chim. Acta 2001, 84, 552-570. 5. Martin, F.; Hay, A. E.; Cressend, D.; Reist, M.; Vivas, L.; Gupta, M. P.; Carrupt, P. A.; Hostettmann, K. Antioxidant C-Glucosylxanthones from the Leaves of Arrabidaea patellifera. J. Nat. Prod. 2008, 71, 1887-1890. 6. Martin, F.; Hay, A. E.; Quinteros Condoretty, V. R.; Cressend, D.; Reist, M.; Gupta, M. P.; Carrupt, P. A.; Hostettmann, K. Antioxidant phenylethanoid glycosides and a neolignan from Jacaranda caucana. J. Nat. Prod. 2009, 72, 852-856. 7. Capitani, C. D.; Carvalho, A. C. L.; Rivelli, D. P.; Barros, S. B. M.; Castro, I. A. Evaluation of natural and synthetic compounds according to their antioxidant activity using a multivariate approach. Eur. J. Lipid. Sci. Technol. 2009, 111, 1090-1099. 8. Re, R.; Pellegrini, N.; Proteggente, A.; Pannala, A.; Yang, M.; Rice-Evans, C. Antioxidant activity applying and improved abts radical cation decolorization assay. Free Rad. Biol. Med. 1999, 26, 1231-1237. 9. Trygg, J.; Wold, S. O2-PLS, a two-block (X-Y) latent variable regression (LVR) method with an integral OSC filter. J. Chemometrics 2003, 17, 53-64.

107

108

Chapter 4

Chapter 4 Estimation of formal potential thanks to ferrocene-derivatives using cyclic voltammetry

4-1 Introduction

4-1.1 Cyclic voltammetry, a relevant method

Cyclic voltammetry (CV) is a more and more commonly used electrochemical method thanks to its contribution toward a wide field extent. It has a great interest, for instance, in characterizing the oxidation mechanism of antioxidant compounds1-4, in characterizing the components and their antioxidant properties in wine5,6, teas, and coffee7, in detecting amino acids in proteins8,9, and in analyzing the DNA oxidation products10. Cyclic voltammetry is also a very useful technique in characterizing electron transfers (ET) and mechanisms in which chemical reactions are coupled with ET, such as proton transfer11. CV is one of the numerous voltammetric methods based on the electrochemical analyses where the current is recorded as the function of the potential applied to the electrode, which induces the oxidation or the reduction of electroactive species. In the case of CV, the forward scan is recorded over a potential range from an initial potential (Ei) to a switching potential (Es). The reverse scan is generally recorded from the latter until the final potential (Ef) that is usually the initial potential; hence the presence of both oxidation and reduction on the voltammogram. A second switching potential may be reached before the final one (Figure 4-1).

109

Chapter 4

Potential Potential

Es Es1

Ei Ef Ei = Ef Es2

Time Time

Figure 4-1 : Linear variation of potential as function of time in cyclic voltammetry

Some concepts concerning the cyclic voltammetry need to be described for a better understanding of the parameters used in this investigation.

4-1.2 Formal standard potential

The formal standard potential Eθ’ is the main parameter for the characterization of an electrochemical system displayed by Equation (4-1).

Ox+ ne− Red (4-1)

where Ox and Red stand for the oxidized and the reduced species, respectively. Eθ’ is defined by the Nernst equation12 (4-2):

θ ' RT ⎡ COx ⎤ EE=+ ln ⎢ ⎥ (4-2) nF⎣ C Red ⎦

where E is the rest potential (in V), R is the ideal gas constant (8.3144 J.mol-1.K-1), T is the temperature (in K), n is the number of transferred electron,

4 -1 F is the Faraday constant (9.6485.10 C.mol ), Cox and Cred are the bulk concentrations (in mol.cm-3) of the oxidized and reduced species, respectively. The

110

Chapter 4 rest potential is the potential reached at equilibrium when the working electrode is immersed in the solution. This rest potential is measured from a reference electrode. When the applied potential is higher than the rest potential, the oxidation occurs. Conversely, a lower potential induces the reduction of the species. The formal standard potential Eθ’ and the standard potential Eθ should not be confused. The formal standard potential depends on factors such as the solvent, the electrolytes, and the electrodes. It is defined by Equation (4-3):

RT γ EEθθ' =+ lnOx (4-3) nF γ Red

where Eθ’ is the formal standard potential, Eθ is the standard potential (at the temperature of 298 K, pressure of 1 bar, and at pH = 0), γox and γred stand for the activity coefficient of oxidized and reduced species, respectively. The other terms have been defined before.

4-1.3 Electrochemical reactions

Many steps are involved during an electrochemical reaction (Figure 4-2).

First of all, the mass transport of the electroactive species from the solution (Ab) to the electrode surface takes place. Then, the species are adsorbed (Aad) and the electron transfer occurs (Bad). Finally, after desorption from the electrode, another step of mass transport takes place from the electrode to the solution (Bb).

111

Chapter 4

kdif Bad Bb

kET kdif Aad Ab

Figure 4-2: Steps of the electrochemical reaction. Ab (Bb) is the species A (B) in the bulk, Aad (Bad) is the species A adsorbed on the electrode, kdif is the diffusion rate and kET is the electron transfer rate.

4-1.3.1 Mass transport

Diffusion, migration and convection are the three modes involved in the mass transport. Diffusion is the movement of particles under the influence of the concentration gradient between the bulk solution and the electrode surface. Migration is the movement of the charged species through an electrostatic field. This parameter may be minimized thanks to a large concentration of electrolytes in the bulk solution. Convection is the movement of particles induced by mechanical forces such as agitation with a magnetic stirring bar. This parameter is cancelled by keeping the bulk solution unstirred. Hence, only the diffusion is taken into account and is described by the Fick’s first law12, Equation (4-4)

⎛⎞∂Cx,t( ) J =−D ⎜⎟Ox (4-4) Ox Ox ⎜⎟∂ ⎝⎠x

-1 -1 where Jox is the flux (in mol.cm .s ) of Ox, Dox is the diffusion coefficient (in ⎛⎞∂Cxt(), cm2.s-1) of Ox and ⎜⎟Ox is the concentration (in mol.cm-3) gradient of Ox. ⎜⎟∂ ⎝⎠x

112

Chapter 4

4-1.3.2 Electron transfer

The rate of the electron transfer at the electrode|solution surface is described in terms of reversible, quasi-reversible and irreversible. The three mechanisms are easily recognized according to their characteristic current- potential curve as well as peculiar parameter values. The parameters used to qualify the electron transfers are figured out from the voltammogram. These are the anodic (ipa) and cathodic (ipc) peak currents, the anodic (Epa) and cathodic (Epc) peak potentials, the half-wave potential (E1/2) and the anodic and cathodic half- peak potential (Epa/2) and (Epc/2), respectively (Figure 4-3).

E1/2 Epa Oxidation

Epa/2

ipa ipa/2 0

i ipc E pc/2 Current (Ampere) pc/2

Reduction Epc

Ei Potential (Volt) Ef

Figure 4-3 : Cyclic voltammogram and electrochemical parameters.

The definition of the terms reversible, quasi-reversible, and irreversible is not similar to those of the reversibility of a homogeneous chemical reaction.

4-1.3.3 Reversible electron transfer

A reversible electron transfer occurs when the rate of the transfer of an electron at the electrode|solution interface is faster than the diffusion of the

113

Chapter 4 electroactive species. Therefore, the limiting kinetics process is the diffusion. The net current i (in A) at the electrode is defined by Equation (4-5)12,13:

⎛⎞∂∂Cx( ) ⎛⎞Cx( ) inFAD==−⎜⎟RedOnFAD ⎜⎟x (4-5) RedO⎜⎟∂∂x ⎜⎟ ⎝⎠xxxx=00 ⎝⎠=

where n stands for the number of transferred electron, F for the Faraday

4 2 constant (9.6485.10 C.mol-1), A for the electrode surface area (in cm ), DRed and

2 -1 DOx for the diffusion coefficient (in cm .s ) of reduced and oxidized species ⎛⎞∂Cx() ⎛⎞∂Cx( ) respectively and ⎜⎟Red and ⎜⎟Ox for the initial concentration (in ⎜⎟∂ ⎜⎟∂ ⎝⎠x x =0 ⎝⎠x x =0 mol.cm-3) of reduced and oxidized species at the electrode surface, respectively. Laplace transformation for linear diffusion is applied to first Fick’s equation and then solved with Nernst equation. Randles-Sevcik equation is obtained as the result of these mathematic methods and after numerical integrations the following equation is obtained:

nFν D inFAC= 0.4463 Red (4-6) pdRe RT

where ip is the peak current (in A), n is the number of transferred electron, F is

2 the Faraday constant, A is the electrode surface area (in cm ), CRed is the reduced species concentration in the bulk solution (in mol.cm-3), ν is the scan rate (in

-1 2 -1 V.cm ), DRed is the diffusion coefficient (in cm .s ) of reduced species, R is the ideal gas constant (8.3144 J.mol-1.K-1), T is the temperature (in K). This equation may also be used for the reduction reaction, the concentration and the diffusion coefficient of the oxidized species will be used instead of those of oxidized species. By rearrangement, the Randles-Sevcik equation (4-7) may be written

114

Chapter 4

i p = constant (4-7) ν

It shows that the peak current over the square root of the scan rate is a constant. Therefore, a reversible electrochemical reaction is characterized by a linear variation of peak current as a function of the square root of the scan rate.

A second characteristic parameter of a reversible electron transfer is the potential peak separation. Firstly, the peak potential is defined by Equation (4-8)11

RT EEp =±1 ()1.109 ± 0.002 (4-8) 2 nF

At a temperature T = 25°C, the peak potential separation11 is defined as

57 Δ=EE − E = mV (4-9) ppapcn

Secondly, by combining the equations of the current and the diffusion- limiting current with Nernst equation12, the following equation (4-10) is obtained

RT ⎡Diδ ⎤⎡⎤RT − i E=+ Eθ ' ln ⎢ RedOx⎥⎢⎥ln dc (4-10) nF Dδ nF i− i ⎣⎢ OxRe d⎦⎣⎦⎥⎢⎥ da

where δOx and δRed are the diffusion layer thickness formed by the oxidized and the reduced species, respectively, idc and ida are the cathodic and anodic diffusion- limiting current respectively and i is the current, the other terms have been

ii+ defined before. When i = ()dc da , the half-wave potential is defined by 2

115

Chapter 4

RT ⎡D δ ⎤ EE=+θ ' ln⎢ RedOx⎥ (4-11) 1 nF D δ 2 ⎣⎢ OxRe d ⎦⎥

Usually, DRed = DOx, hence δRed = δOx, then the half-wave potential is

θ’ identical with the formal standard potential, E1/2 = E .

E1/2 is readily calculated from the voltammogram since it is exactly located between the anodic and the cathodic peak potential.

EE+ θ ' pa pc EE==1 (4-12) 2 2

where Epa and Epc stand for the anodic and cathodic peak potentials, respectively.

Finally, the ipa/|ipc| ratio equal to 1 means that no homogeneous chemical reaction is associated to the heterogeneous electron transfer and no adsorption takes place at the electrode surface. Afterwards, |ipc| will be written ipc and the absolute values will be considered. The characteristic parameters for a reversible electron transfers are the peak potentials and the ip.ν−1/2 which are constant as the scan rate increases and the ΔEp equal to 57/n mV (Table 4-1).

4-1.3.4 Quasi-reversible electron transfer

In a quasi-reversible heterogeneous reaction, the electron transfer rate is intermediate. Consequently, the limiting process for the current in this mechanism is both the diffusion and the electron transfer which are rather of same magnitude. Nernst equation cannot be used and the diffusion equations are solved with the Butler-Volmer equations12. This mathematical treatment will not be discussed here but some characteristics factors may be mentioned such as the increasing peak potential separation as the scan rate increased and the peak

116

Chapter 4 current does not vary linearly with the square root of the scan rate over the all range but still increases as a function of ν1/2 (Figure 4-4).

4-1.3.5 Irreversible electron transfer

A very low electron transfer rate is characteristic of an irreversible reaction and the limiting process for the current in this mechanism is the electron transfer:

Ox+→ ne− Red (4-13)

Like for the quasi-reversible reaction, the Bulter-Volmer equation is used since the Nernst equation cannot. However, the peak current increases linearly within the square root of the scan rate but the peak current remains lower than that expected for a reversible process (Figure 4-4).

i (A)

υ1/2

Figure 4-4 : Graph of the variation of the current with the square root of the scan rate for a reversible (dashed line), quasi-reversible (solid line) and irreversible (dotted line) electron transfer reaction.

117

Chapter 4

Table 4-1: Characteristic parameters of electron transfer reactions.

Reversible Quasi-reversible Irreversible

ΔEp 57/n mV > 57/n mV >> 57/n mV

Ep constant as ν increases varies with ν varies with ν

ip.ν−1/2 linear not linear linear

ipa/ipc 1 1 -

θ E ’ identical to E1/2 - -

4-1.4 Coupled homogeneous chemical reactions

Besides the heterogeneous reactions, chemical reactions may occur prior or after the electrochemical process. In order to simplify the notation, E stands for a heterogeneous electron transfer and C for a homogeneous chemical transfer in the following. The oxidation and the reduction of organic molecules are usually coupled to a chemical reaction, which mainly involves a proton or hydrogen atom transfer. These molecular changes appear on the cyclic voltammogram and are scan-rate- dependant. Here are some processes of electron transfers coupled to chemical reactions. Z represents the electrochemical inactive species, r, q, and i means reversible, quasi-reversible, and irreversible, respectively, and indicate the reversibility of both heterogeneous and homogeneous reaction. ErCr mechanism: the reversible electron transfer is followed by a reversible chemical reaction.

Ox+ ne− Red Z (4-14)

CiEi mechanism: an electron transfer takes place after the chemical reaction. Both are irreversible.

118

Chapter 4

ZOxneRed→+− → (4-15)

ErCiEi mechanism: a reversible heterogeneous reaction takes place, followed by an irreversible chemical reaction and then by an irreversible electron transfer.

−− Ox+→+→ ne Z Ox2 ne Red (4-16)

Other mechanisms may also be studied such as catalytic reactions (ECcat):

Ox+ ne− Red (4-17) Red+ A Ox+ B

or the disproportionation (EE-Disp) as represented by the following equation:

Ox+ ne− Ox i− (4-18) i− 2− 2Ox Ox+ Ox

The voltammograms undergo some variations induced by the coupled homogeneous reactions. The peak potential and / or the peak current for the heterogeneous reaction may be shift and the peak current ratio may decrease due to the effect of the chemical reaction14.

119

Chapter 4

4-1.5 Materials

4-1.5.1 Electrochemical cell

In order to minimize the ohmic drop (iR drop) that is the uncompensated solution resistance, the electrochemical reaction is carried out in a cell with three electrodes, namely the working electrode (WE), the reference electrode (RE), and the counter electrode (CE). A potentiostat generates a current between the WE and the CE corresponding to a potential being measured between the WE and the RE. Oxygen presents in the bulk solution undergoes an oxidation at high potential. To reduce the oxygen concentration, the solution is purged with argon or nitrogen.

4-1.5.2 Working electrode (WE)

The WE is the site where the reaction under investigation takes place. The current flows through it according to the applied potential. The more positive the applied potential generated by the potentiostat is, the higher efficiency in the oxidation process. On the contrary, the more negative the applied potential, the higher efficiency in the reduction process. Mercury (Hg) (such as the dropping mercury electrode (DME) and the hanging mercury drop electrode (HDME)), platinum (Pt), gold (Au), silver (Ag), glassy carbon (GC), and carbon past are among the wide variety of electrode materials. The choice of the electrode material depends on the potential range and on the electrolyte-solvent systems (Figure 4-5). Indeed, the redox potential of the molecule under investigation has to be in the window of the WE potential range in order to avoid the electrochemical reaction of the electrode itself.

120

Chapter 4

0.1 M KCl C 1 M HClO4

0.1 M Et4NOH pH = 14 pH = 7 Hg pH = 0 pH = 14 pH = 7 Pt pH = 0

-3 -2 -1 0 1 2 E [E vs. SCE]

Figure 4-5: Appropriate WE materials for some potential ranges under specific conditions.

Pt and Au electrodes are used with organic and inorganic solvents because of their high overpotential for oxygen evolution as well as their low overpotential for hydrogen evolution. Hg electrode is appropriate to measure the reduction of metal ion since its overpotential for hydrogen evolution is high. The carbon past electrode is usually mixed with enzymes or other compounds to make modified electrodes. The potential window of the GC electrode is very large and this material is chemically stable. These disk electrodes are encased in an inert material toward the solvent. These are PEEK (PolyEther Ether ketone), Teflon, and polypropylene. Both macroelectrodes and microelectrodes (order of magnitude of μm) are used. The advantages of the last ones are the minimized iR-drop and they lead to steady state current from unstirred solutions. Rotating disk electrodes and stationary electrodes are available for voltammetry measurements.

Unlike DME and HDME, a pre-treatment is necessary for the solid electrode before a new voltammogram is recorded. Indeed, a film may be formed at the electrode surface due to the oxidized and reduced species adsorption, impurities and product polymerization. This film leads to a main problem in electrochemical studies that is the lack of reproducibility. The pre-treatment consists in rinsing

121

Chapter 4 the electrode with distilled water and polishing it manually with aluminum oxide-water dispersion on a tissue with a light pressure. Then the electrode is rinsed and immersed in water and then in ethanol in an ultrasound bath.

4-1.5.3 Reference electrode (RE)

Reference electrode (RE) is used to control and to measure the interfacial potential at the WE|solution interface. To be efficient, the RE potential needs to be constant so, the electrode must be non-polarizable. The most common reference electrodes are the silver/silver chloride and the calomel reference electrodes15. The first one is a silver wire coated with silver chloride while the second one is made of mercury covered by a layer of mercury chloride (calomel). They are often placed in the Luggin capillary, close to the working electrode in order to reduce the iR drop that could shift the peak potential toward more positive one in oxidation and more negative one in reduction. Specific reference electrodes are used in non-aqueous solutions. These are pseudo-reference electrodes, namely a platinum or silver wire immersed in the sample solution or a silver/silver ion electrode, which is a silver wire immersed in a solution of silver nitrate or perchlorate.

4-1.5.4 Counter electrode (CE)

The counter electrode is the remaining part of the three-electrode system. It is the second site, with WE, where the current passes through. By this way, only a low current amount goes through the reference electrode, so that the interfacial potential does not vary with the current. The metals widely used as counter electrode are platinum and glassy carbon since they are chemically and electrochemically inert.

122

Chapter 4

4-1.5.5 Solvents and electrolytes

The choice of the solvent depends on the nature of the electroactive species. Both aqueous and organic solvents may be used to make the species soluble. They need to be stable in the potential range and not to interfere with the electrochemical species13. The role of the electrolytes is to promote the anionic and cationic species in the solution. Their concentration is in 100-fold excess over the concentration of the species under investigation to minimize the migration and reduce the resistance between the electrodes. In aqueous solutions, the common electrolytes are NaCl, KCl, KNO3, HCl, HNO3, H2SO4, NaOH, etc. Tetraalkylamonium salts such as (C4H9)4NBF4, (C4H9)4NClO, and (C4H9)4NPF6 are the electrolytes used in the organic systems.

4-1.6 Double layer

When the electrode is immersed in the electrolyte solution, the applied potential induces electric charges at the electrode surface which attract ions of opposite charge. This forms the double layer which behaves like an electrical capacitor and stores charge. These charges do not cross the electrode-solution interface; therefore they are not involved in the electrochemical reaction under investigation but induce a charging current. The measured current is then the sum of the charging current and the faradaic current, which is the movement of electrons across the electrode-solution interface. The charging current is scan- rate-dependent since it obeys to Equation (4-19):

iCc =ν (4-19)

where iC stands for the charging current, ν for the scan rate and C for the double layer capacitance.

123

Chapter 4

4-1.7 Chronoamperometry

Chronoamperometry is another electrochemical method based on the monitoring of the current as a function of time. The applied potential is chosen so that the electroactive species undergoes either the oxidation process or the reduction process. The concentration of the species, oxidized or reduced, can be determined since the faradaic current is described by the Cottrell equation (4-20):

DC inFA= Ox (4-20) t π t

where t is the time in second and the other terms have been defined before. The charging current also takes place but decreases exponentially with time.

Charging Faradaic current current

Figure 4-6 : Charging and faradaic current recorded during a chronoamperometry study.

As a result, the recorded current is the sum of the faradaic and charging current.

124

Chapter 4

4-1.8 Aim of the study

Cyclic voltammetry is one of the most common methods to characterize the electron transfers as well as the chemical reactions coupled to these transfers. The formal standard potentials, which are specific to each electroactive species, provide the electrochemical behavior of species and define the most favorable reactions. The oxidized form of a species with a high formal potential readily reacts with the reduced form of a species with a low formal potential. However, most of the antioxidant compounds do not exhibit any accessible formal potential since they usually undergo an irreversible chemical reaction and the reduced form is not oxidizable (or beyond the potential range)5,16, or undergo a polymerization17. In order to electrochemically characterize the antioxidant compounds, the formal standard potential is required. Since this parameter cannot usually be determined for antioxidant compounds, it could be defined by a range on a potential scale, the steps of which are fixed by well-known electrochemical species. The CV study of an antioxidant mixed with a range-limiting species will provide the ability of the compound to reduce the well-known species. A regeneration of the well-known species means that the compound is capable of reducing it; consequently the formal standard potential of the antioxidant is lower than that of the species. On the contrary, if the compound is unable to reduce the species, then its formal standard potential is higher than that of the species. In this study, ferrocene and some of its derivatives are used as boundary markers. O

NH2 OH

Fe Fe Fe

Figure 4-7 : Structure of aminoferrocene (left), ferrocene (middle) and ferrocenecarboxylic acid (right).

125

Chapter 4

Ferrocene is a reference species, known to electrochemically undergo a reversible one-electron reaction in both aqueous and organic solvents18,19. Several ferrocene- derivatives have been synthesized and have shown a great interest in the scientific fields20,21. Their electrochemical behavior in many solvents lead to a reversible one-electron process22. Aminoferrocene, ferrocene and ferrocenecarboxylic acid (Figure 4-7) have been chosen as reference electroactive species. Their formal potential will be determined in the study conditions and used to mark the boundary of five potential ranges on the scale. Afterwards, the electrochemical reaction of caffeic acid, melatonin and p-hydroquinone added to each ferrocene-derivative will be studied by cyclic voltammetry in order to assess their formal standard potential.

4-2 Materials and methods

4-2.1 Materials

Caffeic acid, ferrocene, ferrocenecarboxylic acid, were purchased from Sigma (Buchs, Switzerland), boric acid, hydroquinone, melatonin, potassium chloride, sodium chloride were purchased from Fluka (Buchs, Switzerland), aminoferrocene was purchased from TCI Europe (Zwijndrecht, Belgium). Water solutions were prepared in demineralised and purified water obtained with the Elix 3 Millipore water purifying system. The electrochemical measurements were performed in a three-compartment cell with a glassy carbon working electrode (diameter 3mm, purchased from Metrohm (Zofingen, Switzerland)), an Ag/AgCl reference electrode and a platinum wire counter electrode. The polishing kit was purchased from Metrohm (Zofingen, Switzerland). The potentiostat was a Autolab PGSTAT30 (Metrohm) used with the GPES version 4.9 software.

126

Chapter 4

4-2.2 Methods

Each compound was dissolved in a (50/50) borate buffer/ethanol (99%) solvent. The borate buffer (0.12 M, pH = 8.3) contained KCl (2.7 mM) and NaCl (0.137 M) as supporting electrolytes. The cyclic voltammograms were recorded for each solution over a potential range of -0.5 mV to 1 V with a potential step of 0.1 V. The scan rate varied over a range from 0.1 mV.s-1 to 100 mV.s-1. The glassy carbon electrode was polished before each measurement with 0.3 μm alumina powder, immersed in ethanol in an ultrasonic bath for 10 minutes, and then rinsed with distilled water. A voltammogram was recorded for each ferrocene-derivative at a concentration of 1 mM, 0.1 mM, and 0.01 mM as well as for the three compounds under investigation at a concentration of 1 mM. Then, the electrochemical reaction was studied for the mixture of each compound (1 mM) with each ferrocene-derivative at the three concentrations. The study by chronoamperometry was performed at a potential of 0.25 V. The voltammograms of ferrocene (1 mM), caffeic acid (1 mM), and the mixture of both were recorded each second for 10 minutes.

4-3 Results and discussion

4-3.1 System validation

In order to rank the studied compounds with respect to the formal potential of ferrocene-derivatives, the validation of the system is required. This validation involves the ferrocene-derivative behavior being electrochemically reversible, namely defined by specific parameters (Table 4-1). Aminoferrocene is one of the ferrocene-derivatives used to assess the formal standard potential of a compound. The electrochemically reversible behavior of aminoferrocene will lead to a step on the potential scale. The plot of the peak

127

Chapter 4 current versus the square root of the scan rate provides a proportionality between both parameters (annex I Figure I-2). Therefore, both oxidation and reduction processes are diffusion-controlled (r² = 0.99 for oxidation and reduction reactions). The ratio ipa/ipc remained constant as ν increased. It equals 1.5.

According to the voltammogram displayed in Figure 4-8, the large ipa/ipc ratio is due to a contribution of the capacitive current. Indeed, the reduction peak is not large enough for an irreversible reaction and the height between forward and reverse signal confirms the capacitance. As concerns the peak potentials, they remained constant, when the scan rate increased and the ΔEp (67 mV) is closed to that expected for a Nernstian value. The electrochemical behavior of aminoferrocene is reversible and the Eθ’, found at 2 ± 9 mV, is the formal standard potential used as a limit value of a potential range.

1e-005

5e-006 i (A) -1.0 -0.5 0.5 1.0 1.5 E (V) -5e-006

-1e-005 Figure 4-8 : Voltammogram of aminoferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

The electrochemical behavior of ferrocene is also diffusion-limited since the peak currents increased linearly with the square root of the scan rate (annex I Figure I-3). The square of the correlation coefficient is 0.99 for oxidation and reduction processes. The ratio ipa/ipc remained constant and equal to 1 as the scan rate increased which indicates a reversible charge transfer in the absence of adsorption or coupled chemical reactions. The well-defined oxidation and reduction peak potentials provide a formal standard potential Eθ’ of 235 ± 6 mV (ΔEp = 64 mV) (Figure 4-9). Eθ’ stayed constant as the scan rate increased. These parameters are consistent with a reversible reaction.

128

Chapter 4

1e-005

5e-006 i (A) -1.0 -0.5 0.5 1.0 1.5 E (V) -5e-006

-1e-005 Figure 4-9 : Voltammogram of ferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

The cyclic voltammograms recorded for ferrocenecarboxylic acid showed well-defined oxidation and reduction peaks at 340 mV and 274 mV, respectively (Figure 4-10).

1e-005

5e-006 i (A) -1.0 -0.5 0.5 1.0 1.5 E (V) -5e-006

-1e-005 Figure 4-10 : Voltammogram of ferrocenecarboxylic acid (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

The anodic (ipa) and cathodic (ipc) peak currents increase linearly with the square root of the scan rate with a squared correlation coefficient of 0.99 for the oxidation and the reduction processes. Therefore, the oxidation-reduction reaction for the ferrocenecarboxylic acid is also diffusion-controlled. The ratio ipa/ipc equals around 1.3, which means that a capacitance took place. However, the ratio ipa/ipc remained constant as the scan rate increased. The same is true for the oxidation and reduction peak potentials. These parameters as well as a ΔEp

129

Chapter 4

(66 mV) close to the Nernstian value indicate a reversible reaction. The formal standard potential, located at 313 ± 18 mV may be used as step on the potential scale.

Table 4-2 summarizes the parameters obtained for the ferrocene and its derivatives.

Table 4-2 : Electrochemical data obtained from cyclic voltammograms of ferrocene and its derivatives.

FcNH2 Fc FcCOOH linear linear linear ipa = f(ν1/2) (r² = 0.99) (r² = 0.99) (r² = 0.99) linear linear linear ipc = f(ν1/2) (r² = 0.99) (r² = 0.99) (r² = 0.99)

ipa / ipc 1.5 1.0 1.3

Variation of ipa/ipc constant constant constant with increasing ν1/2

Variation of Ep constant constant constant when ν increased

E1/2 = E0’ (mV) 2 235 313

ΔEp (mV) 69 68 67

The parameters obtained from the voltammograms of ferrocene and its derivatives in the mixture borate buffer (pH = 8.3) and ethanol (50/50) provide electrochemically reversible behavior for each species. The system is validated; hence a potential scale with three boundary values which are 2 mV, 235 mV, and 313 mV. The ranking of the formal standard potentials of the ferrocene-derivatives is consistent with their structure. Indeed the electron-withdrawing group, namely COOH, increases the Eθ’ value relative to that of ferrocene. On the contrary, the electron loss is enabled thanks to the electron-donating NH2 residue, which induces a Eθ’ value decrease.

130

Chapter 4

4-3.2 Electrochemical behavior of studied compounds alone

Before the addition of an antioxidant to the ferrocene-derivative solution, the cyclic voltammogram of caffeic acid, melatonin, and para-hydroquinone was recorded in the mixture borate buffer (pH = 8.3)/EtOH (50/50).

4-3.2.1 Cyclic voltammetry study of caffeic acid

The voltammogram (Figure 4-11) depicts an oxidation and a reduction peak at 576 mV and -58 mV, respectively, as well as a shoulder around 360 mV. The voltammogram shows that the double layer capacitance is obviously present. However the capacitive current did not vary linearly with the increasing scan rate; the shoulder is not solely induced by the capacitance. The two oxidation peaks likely stand for the two ionized forms of caffeic acid (CA). Indeed, at pH = 8.3, 58% is in its form CA–, namely -COO–, and 42% is in its form CA2–, namely – COO–, 3-OH, 4-O–16. The form CA2– is oxidized at a lower potential than that of CA– in aqueous solution16 but the peak potentials shifts to more positive value in a mixture water/MeOH (1/1) when pH increases3. Therefore the oxidation of CA– may have occurred at around 360 mV when the form CA2– was oxidized at 576 mV. The oxidation of caffeic acid is a two-electron-two-proton process4,16,17, thus the second electron loss from CA2– likely occurs at the same potential value as the electron losses from CA–; hence a large peak current at 576 mV.

The process is diffusion-controlled since ipa and ipc varied linearly with the square root of the scan rate (r² = 0.99 for both oxidation and reduction process).

Peak potentials remained constant as the scan rate increased. The large ΔEp and proportionality between the peak currents and ν1/2 describe an apparently irreversible charge transfer for caffeic acid.

131

Chapter 4

2e-005

1e-005 i (A)

-1.0 -0.5 0.5 1.0 1.5 E (V)

-1e-005 Figure 4-11 : Voltammogram of caffeic acid (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

The electrochemical behavior of the 3,4-dihydroxycinnamic acid is pH- dependant. The process is reversible at pH acid (pH = 2)16 and becomes

17 irreversible at pH > 9 since ΔEp increases as pH increases. At pH = 8.5, caffeic acid follows slow heterogeneous electron transfer kinetics16. The pH-dependant electrochemical behavior is due to the neutral and ionized forms of caffeic acid and the capacity for losing electrons. This is consistent with the results obtained in the study conditions.

The ratio ipa/ipc increased until reached a limit value of 2.7 as the scan rate increased. This is consistent with a mechanism ECE23,24 which means a one- proton-two-electron process. The first oxidation leads to the semi-quinone through an electron loss which is further oxidized by the loss of a second electron and proton to yield the quinone4,16,17. The drastic low cathodic peak current relative to ipa is due to the decomposition of the quinone which occurs at pH above 7.4 while it is stable at pH under 5.53,16. The two-electron-one-proton transfer rapidly occurs; hence only one oxidation peak for CA2–. The lower difference between the peak currents of the shoulder (360 mV) and the anodic peak (576 mV) at low scan rate than that at higher scan rate highlights a rearrangement of the structure. Indeed, at 0.5 and 1 mV.s-1 (annex I Figure I-4 and I-5), a steady state appeared in the forward scan while no reduction peak was recorded in the back scan. The oxidized product of CA– undergoes further reactions, likely with the contribution of CA2–, which leads to a structure not

132

Chapter 4 electrochemically reducible. On the contrary, high scan rates do not allow a rearrangement; hence the presence of a reduction peak. Besides the irreversible chemical reaction, adsorption obviously contributes to the large ratio of peak currents. It is visible on the voltammogram and caffeic acid is known to be adsorbed at the GC electrode surface4,17. The film resulting from the adsorption is used as a surface-immobilized redox couple25.

4-3.2.2 Cyclic voltammetry study of melatonin

The voltammogram (Figure 4-12) shows an oxidation peak in the forward scan and no reduction peak during the reverse scan (considering the potential window under study). The electrochemical behavior of melatonin is obviously irreversible. The plot of the anodic peak current versus the square root of the scan rate displays proportionality between them (r² = 0.99), which means a diffusion-controlled process for this compound. The oxidation peak potential (664 mV) did not significantly vary with an increasing scan rate.

2e-005

1e-005

5e-006 i (A)

-1.0 -0.5 0.5 1.0 1.5 -5e-006 E (V)

-1e-005 Figure 4-12 : Voltammogram of melatonin (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

A pH-dependant electrochemical behavior has been established for melatonin with a potential shift to less potential values; hence in favor of the oxidation26,27. In protic media (pH < 4), the voltammogram provides two oxidation peaks (0.76 and 0.93 mV) and a reduction peak (0.38 mV). A third oxidation peak

133

Chapter 4 appears on the second scan (0.42) which is due to the oxidized product from the first oxidation26,27. The melatonin electrochemical behavior is a one-proton-two- electron process. After the first electron loss, the radical cation is formed and then the loss of the second electron and proton leads to the quinonemine. It is highly reactive and undergoes further reactions such as nucleophile attack and dimerisation26,27.

4-3.2.3 Cyclic voltammetry study of para-hydroquinone

The quinone/hydroquinone couple is one of the most studied redox couples28- 33. Its well-known electrochemical properties support the use of these compounds to develop new strategies. Electrogenerated benzoquinone was revealed as selective tag for L-cysteine residue and is therefore efficient in detecting and quantifying the cysteine containing proteins8,9. Para-hydroquinone is also the initial product of the electrochemical synthesis of hydroxylated benzofuranes, known to have therapeutic properties34. The acidity or the basicity of unknown compounds and the electrochemical behavior of non electroactive species may be assessed thanks to p-benzoquinone and p-hydroquinone35. The oxidation of para- hydroquinone (p-HQ) to benzoquinone is known to be a two-proton-two-electron process36,37. This well-known compound is used as a proof of the principle; its formal standard potential is 0.209 V at pH = 8.3 (the calculation is detailed in annex I eq. I-1)38. However, this Eθ’ is surely higher in the system under study due to the 50% ethanol-containing media which makes the pKa of p-HQ higher. Para-hydroquinone undergoes an oxidation and a reduction during the electrochemical process as displayed in Figure 4-13.

134

Chapter 4

2e-005

1e-005

-1.0 -0.5i (A) 0.5 1.0 1.5 E (V) -1e-005

-2e-005 Figure 4-13 : Voltammogram of para-hydroquinone (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

Both anodic and cathodic peak currents linearly increase with the increasing square root of the scan rate, which is related to a diffusion-controlled process.

The ratio ipa/ipc is closed to 0.9 and remained constant as the scan rate increased.

On the contrary, ΔEp increased with higher scan rate, which is specific of an irreversible heterogeneous electron transfer.

4-3.3 Determination of the formal standard potential of caffeic acid in the ferrocene system

4-3.3.1 Caffeic acid and aminoferrocene

The electrochemical behavior of aminoferrocene in presence of caffeic acid will provide a formal potential for CA which will be lower than that of FcNH2 (2

+ mV) if a reaction occurs, namely FcNH2 is regenerated by caffeic acid. On the contrary, no change on the voltammogram will mean that no reaction between both species take place; hence a higher formal potential of caffeic acid than 2mV. The voltammogram of the mixture caffeic acid/aminoferrocene in equimolar ratio is displayed in Figure 4-14. Two oxidation peaks and a shoulder are visible at 69 mV, 687 mV and around 400 mV, respectively. The only reduction peak is located at 7 mV. These potential values are shifted to more positive values

135

Chapter 4 relative to those of the species alone (Table 4-3). A probably specific adsorption (wave around 0.4V) that partially blocks the electrode should hinder the electron transfer involving a potential shift toward more positive values.

2e-005

1e-005

-1.0 -0.5i (A) 0.5 1.0 1.5 E (V) -1e-005

-2e-005 Figure 4-14 : Cyclic voltammogram of caffeic acid (1 mM) and aminoferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

The current peak values calculated from the voltammogram in Figure 4-14 are similar to those from the voltammograms of the species alone. Furthermore, the ratio of ipa1 over ipc is equal to that of FcNH2. The electrochemical behavior of both aminoferrocene and caffeic acid is visible on the voltammogram. The broad reduction peak includes the cathodic peak of both FcNH2 and CA which are close.

Caffeic acid is obviously unable to regenerated FcNH2, consequently, its formal potential is higher than 2 mV.

136

Chapter 4

Table 4-3 : Electrochemical data of caffeic acid (1 mM), aminoferrocene (1 mM) and the mixture caffeic acid (1 mM) / aminoferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

Caffeic acid (1 mM) / Caffeic acid (1 mM) FcNH2 (1 mM) FcNH2 (1 mM)

Epa1 = 69 mV Epa 576 mV 35 mV Epa2 = 687 mV

Epc -58 mV -31 mV Epc = 7 mV

ipa1 = 10.5 μA ipa 16 μA 8.4 μA ipa2 = 11.9 μA

ipc 6 μA 5.7 μA ipc = 7.7 μA

ipa1/ipc = 1.4 ipa/ipc 2.7 1.5 ipa2/ipc = 1.5

4-3.3.2 Caffeic acid and ferrocene

Ferrocene showed specific characteristics of a reversible charge transfer and is the second step on the potential scale used to assess formal standard potential of antioxidant compounds. The voltammogram recorded from the addition of caffeic acid to ferrocene in an equimolar ratio shows an oxidation peak and two reduction peaks (Figure 4-15).

137

Chapter 4

3e-005

2e-005

1e-005 i (A)

-1.0 -0.5 0.5 1.0 1.5 -1e-005 E (V)

-2e-005 Figure 4-15 : Cyclic voltammogram of caffeic acid (1 mM) and ferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

The anodic peak from the caffeic acid oxidation is not visible and the oxidation peak value is close to that of ferrocene. The reduction peaks are located at the same potential value than that of ferrocenium and oxidized caffeic acid

(Table 4-4). Furthermore, Epa, Epc1, and ratio of ipa over ipc1 remained constant with increasing scan rate which is in line with the electrochemical behavior of ferrocene. On the other hand, Epc2 shifted toward more negative values as scan rate increased, which is consistent with the cathodic potential values oxidized CA. Therefore, the anodic peak and the first cathodic peak depict the oxidation- reduction process of ferrocene when the second anodic peak is the signal of the reduction of oxidized caffeic acid. The process is diffusion-controlled. Each peak current increased linearly with the increasing square root of the scan rate (r² = 0.99).

138

Chapter 4

Table 4-4 : Electrochemical data of caffeic acid (1 mM), ferrocene (1 mM) and the mixture caffeic acid (1 mM) / ferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

Caffeic acid (1 mM) / Caffeic acid (1 mM) Fc (1 mM) Fc (1 mM)

Epa 576 mV 271 mV 265 mV

Epc1 = 206 mV Epc -58 mV 203 mV Epc2 = -55 mV

ipa 16 μA 10.5 μA 26 μA

ipc1 = 9.5 μA ipc 6 μA 10.1 μA ipc2 = 10 μA

ipa/ipc 2.7 1.0 ipa/ipc1 = 2.7

The voltammogram evidences the regeneration of Fc by caffeic acid. Indeed, more ferrocene undergoes the oxidation at the electrode surface, hence the larger anodic peak current and caffeic acid is oxidized by the reaction with ferrocenium. The anodic peak of CA has not probably disappeared but it is hidden by the oxidation peak of Fc. The reduction peak of oxidized CA is 1.5-fold higher than that when CA is alone. It is due to the sum of caffeic acid oxidized by the reaction with Fc+ and by the CA oxidation at the electrode surface. This is a further evidence of the regeneration of ferrocene by caffeic acid. The presence of the cathodic peak of Fc+ shows that Fc was oxidized despite the reducing ability of caffeic acid. At lower scan rate, the reduction peak of Fc+ has disappeared while the oxidation peak of CA is visible (Figure 4-16). The reaction between ferrocenium and caffeic acid has more time to occur. Fc is mostly regenerated by caffeic acid, hence the absence of reduction peak of Fc+. The reduction peak on the voltammogram is the cathodic peak of oxidized CA which shifts to more positive potential values as the scan rate decreases.

139

Chapter 4

4e-006

2e-006 i (A)

-1.0 -0.5 0.5 1.0 1.5 E (V) -2e-006 Figure 4-16 : Cyclic voltammogram of caffeic acid (1 mM) and ferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 5 mV.s-1

A large excess of caffeic acid toward ferrocene provides the same results (annex I Figure I-6). Caffeic acid is obviously able to reduce the ferrocenium produced from the heterogeneous oxidation of ferrocene. This ability provides a formal standard potential lower than that of ferrocene, namely 235 mV. The electrochemical process is a catalytic mechanism, described by Equation (4-17).

4-3.3.3 Caffeic acid and ferrocenecarboxylic acid

The electrochemical processes of the addition of caffeic acid to the aminoferrocene and ferrocene solution have provided a formal potential included in a potential range from 2 to 235 mV. Consequently, caffeic acid should be capable of regenerating FcCOOH as it did with Fc.

An oxidation peak (Epa = 333 mV) and two reduction peaks (Epc1 = 273 mV and Epc2 = -57 mV) are visible on the voltammogram (Figure 4-17).

140

Chapter 4

3e-005

2e-005

1e-005 i (A)

-1.0 -0.5 0.5 1.0 1.5 -1e-005 E (V)

-2e-005 Figure 4-17 : Cyclic voltammogram of caffeic acid (1 mM) and ferrocenecarboxylic acid (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

The process is diffusion-controlled since the peak current of the three peaks increased linearly with the increasing scan rate (r² = 0.99). The anodic peak potential slightly varied to more positive potential value when the scan rate increases, while Epc1 remained constant. The second cathodic peak potential Epc2 significantly decreased to less positive potential value as the scan rate increased. The potential values at 100 mV.s-1 and the potential variation with the scan rate allow the peak identification. Indeed, the oxidation peak and the first reduction peak Epc1 depict the carboxyferrocene electrochemical behavior when the second cathodic peak is the reduction peak of caffeic acid (Table 4-5). As with the reaction with ferrocene, the oxidation peak of CA is not visible and the anodic peak current of oxidized CA is 2-fold higher in the mixture than that when CA is alone. The regeneration of FcCOOH by caffeic acid is proven.

141

Chapter 4

Table 4-5 : Electrochemical data of caffeic acid (1 mM), carboxyferrocene (1 mM) and the mixture caffeic acid (1 mM) / carboxyferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

Caffeic acid (1 mM) / Caffeic acid (1 mM) FcCOOH (1 mM) FcCOOH (1 mM)

Epa 576 mV 340 mV 333 mV

Epc1 = 273 mV Epc -58 mV 274 mV Epc2 = -57 mV

ipa 16 μA 9.8 μA 25 μA

ipc1 = 9.1 μA ipc 6 μA 7.5 μA ipc2 = 11.3 μA

ipa/ipc 2.7 1.3 ipa/ipc1 = 2.7

The absence of the first reduction peak at low scan rate (annex I Figure I-7) and when CA is in large excess (annex I Figure I-8) emphasizes the ability of caffeic acid to regenerate the oxidized ferrocenecarboxylic acid, since the produced FcCOOH+ is readily reduced by CA. The catalytic reaction is described by Equation (4-17). These results are consistent with those obtained with ferrocene. The formal standard potential of caffeic acid is included in a potential range from 2 mV to 235 mV.

4-3.3.4 Electrochemical study of the caffeic acid-ferrocene solution by chronoamperometry

The study by CV of the caffeic acid added to a ferrocene solution has provided a catalytic process. In order to confirm the ferrocenium-reducing ability of caffeic acid, the mixture was studied by chronoamperometry.

142

Chapter 4

1.0x10-5 i (A)

3.0x10-6

1.0x10-6 0 0 200 400 600 t (s)

Figure 4-18 : Chronoamperometry study of caffeic acid (dashed line), ferrocene (dotted line) and the caffeic acid-ferrocene solution (solid line).

At a potential of 250 mV, ferrocene undergoes an oxidation reaction. The concentration of Fc at the electrode surface decreases while that of Fc+ increases. The monitored current as function of time provides the oxidized species concentration. The plot displayed in Figure 4-18 depicts a low current for caffeic acid and ferrocene, due to the charging current, and a larger current for the mixture which indicates a catalytic steady state. Indeed, the reduction of ferrocenium by caffeic acid keeps the Fc concentration steady. The results are then consistent with the regeneration of ferrocene by caffeic acid provided by cyclic voltammetry.

4-3.4 Determination of the formal standard potential of melatonin in the ferrocene system

4-3.4.1 Melatonin and aminoferrocene

Melatonin was characterized by only one oxidation potential at 664 mV in the studied conditions, which does not allow the determination of its formal standard potential. The ability of the indolamine to regenerate aminoferrocene is

143

Chapter 4 assessed and will provide clues concerning the position of its E0’ on the formal standard potential scale.

2.0×10 -5

1.0×10 -5 i (A)

-1.0 -0.5 0.5 1.0 1.5 E (V)

-1.0×10 -5 Figure 4-19 : Cyclic voltammogram of melatonin (1 mM) and aminoferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

Two anodic peaks and a cathodic peak are visible on the voltammogram displayed in Figure 4-19. The first oxidation potential and the reduction potential are shifted of around 50 mV toward more positive values relative to those of aminoferrocene (Table 4-6). The gap between the second anodic peak and that of melatonin is even much larger (about 150 mV). These potential shifts are attributed to the adsorption at the electrode surface which partially blocks the electron transfer. The similarity of the peak currents tends to the attribution of the first anodic peak and the reduction peak to aminoferrocene when the second oxidation peak describes melatonin behavior. Furthermore, the ratio ipa1/ipc in the mixture is close to that of FcNH2. Therefore, the oxidized aminoferrocene is not regenerated by the indolamine, which means that the E0’ of melatonin is above 2 mV. Proportionality between the peak currents and increasing the scan rate has been verified, the process is diffusion-limited (r² = 0.99).

144

Chapter 4

Table 4-6 : Electrochemical data of melatonin (1 mM), aminoferrocene (1 mM) and the mixture melatonin (1 mM) / aminoferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

Melatonin (1 mM) / Melatonin (1 mM) FcNH2 (1 mM) FcNH2 (1 mM)

Epa1 = 87 mV Epa 664 mV 35 mV Epa2 = 812 mV

Epc - -31 mV Epc = 19 mV

ipa1 = 10 μA ipa 10.2 μA 8.4 μA ipa2 = 12.7 μA

ipc - 5.7 μA ipc = 7.7 μA

ipa/ipc 1.5 ipa1/ipc = 1.3

The voltammograms recorded at low scan rate and with a large excess of melatonin provided similar results.

4-3.4.2 Melatonin and ferrocene

The electrochemical process of melatonin added to ferrocene solution will lead to the localization of the formal potential of the antioxidant either above or under 235 mV.

1.0×10 -5 i (A)

-1.0 -0.5 0.5 1.0 1.5 E (V)

-1.0×10 -5 Figure 4-20 : Cyclic voltammogram of melatonin (1 mM) and ferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 80 mV.s-1

145

Chapter 4

The voltammogram recorded from the cyclic voltammetry study of the mixture melatonin and ferrocene depicts three well-defined peaks (Figure 4-20). The diffusion-controlled process (r² = 0.99) reveals a first oxidation peak, located at 274 mV, which is similar to that of ferrocene (Table 4-7). The second oxidation peak has the same potential value as that of melatonin as well as a similar peak current. Finally, the reduction peak of the mixture reflects the reduction of Fc+ since the potential values and the peak current are close to those of ferrocene

(Table 4-7). Besides, the ratio ipa1/ipc of the mixture is close to that of ferrocene.

Table 4-7 : Electrochemical data of melatonin (1 mM), ferrocene (1 mM) and the mixture melatonin (1 mM) / ferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 80 mV.s-1

Melatonin (1 mM) / Fc Melatonin (1 mM) Fc (1 mM) (1 mM)

Epa1 = 274 mV Epa 664 mV 271 mV Epa2 = 667 mV

Epc - 203 mV Epc = 209 mV

ipa1 = 8.2 μA ipa 9.0 μA 9.8 μA ipa2 = 11.2 μA

ipc - 7.6 μA ipc = 7.4 μA

ipa / ipc 1.3 ipa1 / ipc = 1.1

The relevant electrochemical characteristics of the mixture provide a formal standard potential of melatonin higher than that of ferrocene. This is also supported by the voltammograms recorded at low scan rate and with melatonin in a large excess.

146

Chapter 4

4-3.4.3 Melatonin and ferrocenecarboxylic acid

Carboxyferrocene is the remaining electroactive species used to rank the formal potential of melatonin. The voltammogram of the mixture is displayed in Figure 4-21.

2.0×10 -5

1.5×10 -5

1.0×10 -5

-6

i (A) 5.0×10

-1.0 -0.5 0.5 1.0 1.5 -5.0×10 -6 E (V)

Figure 4-21 : Cyclic voltammogram of melatonin (1 mM) and ferrocenecarboxylic acid (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

Two oxidation peaks are present in the forward scan and a reduction peak is present on the reverse scan. Proportionality has been proven between the peak current and the increasing scan rate for each peak (r² = 0.99) which indicates a diffusion-controlled process. The electrochemical data of the three peaks are gathered in Table 4-8.

147

Chapter 4

Table 4-8 : Electrochemical data of melatonin (1 mM), ferrocenecarboxylic acid (1 mM) and the mixture melatonin (1 mM) / ferrocenecarboxylic acid (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

Melatonin (1 mM) / Melatonin (1 mM) FcCOOH (1 mM) FcCOOH (1 mM)

Epa1 = 352 mV Epa 664 mV 340 mV Epa2 = 680 mV

Epc - 274 mV Epc = 282 mV

ipa1 = 9.2 μA ipa 10.2 μA 9.8 μA ipa2 = 13.3 μA

ipc - 7.5 μA ipc = 7.7 μA

ipa / ipc 1.3 1.2

Although the potential value are slightly shifted toward more positive values, the first oxidation peak and the reduction peak obviously describe the electrochemical behavior of carboxyferrocene. Besides, this is confirmed by the similar anodic and cathodic currents. The second oxidation peak corresponds with the melatonin voltammogram. Therefore, the addition of the antioxidant in the FcCOOH solution does not affect the electrochemical behavior of the two compounds; the indoleamine is unable to regenerate FcCOOH+. The voltammograms of the mixture using low scan rates and melatonin in excess provided the same results. The formal standard potential of melatonin is obviously higher than that of ferrocenecarboxylic acid, namely 313 mV.

148

Chapter 4

4-3.5 Para-hydroquinone as a proof of the principle

4-3.5.1 Para-hydroquinone and aminoferrocene

Para-hydroquinone is a widely used compound in electrochemistry studies thanks to its well-known behavior. Its oxidation follows a two-proton-two-electron process and leads to para-benzoquinone. When added to aminoferrocene in equimolar ratio, the cyclic voltammetry study provides a voltammogram with two seeming oxidation peaks and a reduction peak (Figure 4-22).

2.0×10 -5

1.0×10 -5

-1.0 i (A) -0.5 0.5 1.0 1.5 E (V) -1.0×10 -5

-2.0×10 -5 Figure 4-22 : Cyclic voltammogram of p-hydroquinone (1 mM) and aminoferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

The oxidation peak of aminoferrocene and p-hydroquinone shifted toward more positive values and the peak current of aminoferrocene is larger in the mixture

+ (Table 4-9). This would indicate that p-HQ is capable of regenerating FcNH2 which would lead to a larger amount of FcNH2; hence the larger peak current.

However, as regards the reverse scan, the either the reduction peak of FcNH2 or that of p-HQ has disappeared. The peak potential is in line with the Epc of the ferrocenium-derivative while the peak current tends to indicate that the

+ reduction peak is that of oxidized p-HQ. By assuming that FcNH2 is reduced by

+ p-HQ, the current of the reduction peak of FcNH2 would be much lower than

149

Chapter 4 that of aminoferrocenium recorded without p-HQ. Moreover, the peak current of oxidized p-HQ would be much higher than that recorded alone due to the oxidation undergone by the FcNH2 regeneration and by the electrochemical process. Para-benzoquinone is susceptible to undergo nucleophile addition which may produce non-reducible species. Such an addition has already been reported with L-cysteine and para-hydroquinone8. However a pKb of 10.35 has been determined in a 80% ethanol-containing media39 that implies that aminoferrocene is mainly in its protonated form in the media under study. Then, aminoferrocene does likely not act as a nucleophilic agent and other reactions seem take place. Therefore, the study related to the mixture p-hydroquinone and aminoferrocene cannot provide specific results.

Table 4-9 : Electrochemical data of p-hydroquinone (1 mM), aminoferrocene (1 mM) and the mixture p-hydroquinone (1 mM) / aminoferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

para-hydroquinone para-hydroquinone (1 FcNH2 (1 mM) (1 mM) mM) / FcNH2 (1 mM)

Epa1 = 80 mV Epa 279 mV 35 mV Epa2 = 294 mV

Epc -66 mV -31 mV Epc = -34 mV

ipa1 = 12 μA ipa 13.2 μA 8.4 μA ipa2 = 13.2 μA

ipc 15.1 μA 5.7 μA ipc = 14.9 μA

ipa1 / ipc = 0.80 ipa / ipc 0.87 1.5 ipa2 / ipc = 0.89

4-3.5.2 Para-hydroquinone and ferrocene

The CV study of each compound showed that both of them have a close oxidation peak value (Table 4-10). The voltammogram of the mixture depicts a broad anodic peak with a shoulder and two cathodic peaks (Figure 4-23).

150

Chapter 4

2.0×10 -5

1.0×10 -5

-1.0 i (A) -0.5 0.5 1.0 1.5 E (V) -1.0×10 -5

-2.0×10 -5 Figure 4-23 : Cyclic voltammogram of p-hydroquinone (1 mM) and ferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 80 mV.s-1

The second cathodic peak is obviously the reduction peak of p-hydroquinone since potential and current values are similar. As concerns the first cathodic peak (283 mV), the current value is close to that of ferrocene but the potential value is significantly higher than that of Fc (Table 4-10). This implies that the anodic peak at 333 mV is the signal of the oxidation of ferrocene and the oxidation- reduction process of ferrocene has undergone an around-70-mV shift.

151

Chapter 4

Table 4-10 : Electrochemical data of p-hydroquinone (1 mM), ferrocene (1 mM) and the mixture p-hydroquinone (1 mM) / ferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 80 mV.s-1

para-hydroquinone para-hydroquinone Fc (1 mM) (1 mM) (1 mM) / Fc (1 mM)

Epa1 = 265 mV Epa 279 mV 271 mV Epa2 = 333 mV

Epc1 = 283 mV Epc -66 mV 203 mV Epc2 = -79 mV

ipa1 = 14.2 μA ipa 15.0 μA 9.8 μA ipa2 = 15.9 μA

ipc1 = 7.6 μA ipc 13.0 μA 7.6 μA ipc2 = 13.5 μA

ipa1/ipc2 = 1.1 ipa/ipc 1.2 1.3 ipa2/ipc2 = 2.1

The maximal current was recorded for a potential value of 333 mV, which is a higher oxidation potential value than that of p-HQ and Fc. The anodic currents at 265 mV and 333 mV are close to that of p-HQ alone and larger than that of Fc. By assuming that the p-HQ reduces the produced ferrocenium, the peak current ratio of p-HQ should be higher than 1.1 since more p-benzoquinone is produced from the ferrocenium reduction. However, the ratio ipa/ipc of the couple benzoquinone/hydroquinone is equal to 1.1 as it was for the CV study of p-HQ alone. Consequently, no catalytic reaction takes place between Fc and p-HQ; the formal potential of p-hydroquinone is above 235 mV. The larger anodic current of Fc is surely due to the oxidation current effect of p-HQ and the 70-mV shift is attributed to the oxidized p-HQ adsorption at the electrode surface. The absence of regeneration of ferrocene by p-HQ confirms that the

+ hydroquinone is not capable of reducing FcNH2 and by-products are obviously formed from reactions between the species involved.

152

Chapter 4

4-3.5.3 Para-hydroquinone and ferrocenecarboxylic acid

The voltammogram recorded from the CV study of the mixture para- hydroquinone and ferrocenecarboxylic acid displays two oxidation and two reduction peaks (Figure 4-24).

2.0×10 -5

1.0×10 -5

-1.0 i (A) -0.5 0.5 1.0 1.5 E (V) -1.0×10 -5

-2.0×10 -5

Figure 4-24 : Cyclic voltammogram of p-hydroquinone (1 mM) and carboxyferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

The oxidation and reduction process of p-hydroquinone appears on the voltammogram with very close potential and current values relative to those of p- HQ alone (Table 4-11). However, the electrochemical signal of FcCOOH has undergone a 100 mV shift toward more positive values due to the adsorption of oxidized p-HQ. The peak currents did not vary linearly with the increasing square root of the scan rate but increased linearly with the scan rate. This implies a non-faradaic adsorption. The larger anodic current of FcCOOH in the mixture relative to that of FcCOOH alone is likely due to the p-HQ-related current which induces a higher baseline.

153

Chapter 4

Table 4-11 : Electrochemical data of p-hydroquinone (1 mM), ferrocenecarboxylic acid (1 mM) and the mixture p-hydroquinone (1 mM) / ferrocenecarboxylic acid (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1

para-hydroquinone para-hydroquinone (1 FcCOOH (1 mM) (1 mM) mM) / FcCOOH (1 mM)

Epa1 = 275 mV Epa 279 mV 340 mV Epa2 = 426 mV

Epc1 = 370 mV Epc -66 mV 274 mV Epc2 = -63 mV

ipa1 = 12.7 μA ipa 13.2 μA 9.8 μA ipa2 = 16.5 μA

ipc1 = 8.6 μA ipc 15.1 μA 7.5 μA ipc2 = 17.7 μA

ipa1/ipc2 = 0.7 ipa/ipc 0.87 1.3 ipa2/ipc1 = 1.9

The reduction peak current of p-benzoquinone demonstrates that no more oxidized p-HQ was produced in presence of FcCOOH+. As a result, para- hydroquinone did not induce a regeneration of carboxyferrocene, which means that its formal standard potential is above 313 mV. This is relatively in line with its Eθ’ in the media under study. Indeed, the presence of ethanol (50%) makes the Eθ’ higher than that in aqueous solution, which is 209 mV.

4-4 Conclusion

The formal standard potential is a main parameter to rank species with respect to the ability to transfer electrons. However, it is not always easy to determine it for antioxidant compounds. Ferrocene and its derivatives have been used to assess the formal standard potential of three antioxidant compounds.

154

Chapter 4

E0’ (mV)

E0’ of melatonin and p-hydroquinone 313 FcCOOH+ FcCOOH

235 Fc+ Fc

E0’ of caffeic acid 2 + + FcNH2 Fc2

Figure 4-25 : Potential scale built with formal potential of ferrocene-derivatives and localization of the formal potential of caffeic acid, melatonin, and p-hydroquinone.

Aminoferrocene, ferrocene, and ferrocenecarboxylic acid have shown a reversible electron transfer reaction in a 50/50 borate buffer (pH = 8.3)/ethanol. Their formal standard potentials have been calculated at 2 mV, 235 mV, and 313 mV, respectively. These three values have provided a potential scale (Figure 4-25). The CV studies of the antioxidants in presence of the ferrocene-derivatives have lead to catalytic reactions and to the absence of interaction between the species. The formal standard potential of caffeic acid has been assessed between 2 mV and 235 mV since catalytic reactions occurred with ferrocene and carboxyferrocene. Concerning melatonin and p-hydroquinone, their formal standard potentials have been located above 313 mV since they were unable to regenerate the ferrocene-derivatives. The pKa of para-hydroquinone will be determined in the study media to confirm its behavior in the ferrocene system. Other electrochemical species, the formal standard potential of which is known, will be studied in this system to validate it. Furthermore, the Eθ’ of caffeic acid and melatonin will be calculated thanks to simulation software. To further assess the formal standard potentials, more electroactive species which exhibit a reversible one-electron reaction, will be used to contract the potential ranges22. One of them is N,N- dimethylaminoferrocene, which may avoid the side reactions which took place when aminoferrocene was in presence of para-hydroquinone.

155

Chapter 4

Bibliographic references

1. Born, M.; Carrupt, P. A.; Zini, R.; Brée, F.; Tillement, J. P.; Hostettmann, K.; Testa, B. Electrochemical behaviour and antioxidant activity of some natural polyphenols. Helv. Chim. Acta 1996, 79, 1147-1158. 2. Firuzi, O.; Lacanna, A.; Petrucci, R.; Marrosu, G.; Saso, L. Evaluation of the antioxidant activity of flavonoids by "ferric reducing antioxidant power" assay and cyclic voltammetry. Biochim. Biophys. Acta 2005, 1721, 174-184. 3. Hapiot, P.; Neudeck, A.; Pinson, J.; Fulcrand, H.; Neta, P.; Rolando, C. Oxidation of caffeic acid and related hydroxycinnamic acids. J. Electroanal. Chem. 1996, 405, 169-176. 4. Kallel Trabelsi, S.; Belhadj Tahar, N.; Abdelhedi, R. Electrochemical behavior of caffeic acid. Electrochim. Acta 2004, 49, 1647-1654. 5. Kilmartin, P. A.; Zou, H.; Waterhouse, A. L. A cyclic voltammetry method suitable for characterizing antioxidant properties of wine and wine phenolics. J. Agric. Food Chem. 2001, 49, 1957-1965. 6. Makhotkina, O.; Kilmartin, P. A. Uncovering the influence of antioxidants on polyphenol oxidation in wines using an electrochemical method: Cyclic voltammetry. J. Electroanal. Chem. 2009, 633, 165-174. 7. Kilmartin, P. A.; Hsu, C. F. Characterisation of polyphenols in green, oolong, and black teas, and in coffee, using cyclic voltammetry. Food Chem. 2003, 82, 501-512. 8. Roussel, C.; Rohner, T. C.; Jensen, H.; Girault, H. H. Mechanistic aspects of on-line electrochemical tagging of free L-cysteine residues during electrospray ionisation for mass spectrometry in protein analysis. Chem. Phys. Chem. 2003, 4, 200-206. 9. Roussel, C.; Dayon, L.; Jensen, H.; Girault, H. H. On-line cysteine modification for protein analysis: new probes for electrochemical tagging nanospray mass spectrometry. J. Electroanal. Chem. 2004, 570, 187-199. 10. Liu, J. F.; Roussel, C.; Lagger, G.; Tacchini, P.; Girault, H. H. Antioxidant sensors based on DNA-modified electrodes. Anal. Chem. 2005, 77, 7687-7694. 11. Nicholson, R. S.; Shain, I. Theory of stationary electrode polarography single scan and cyclic methods applied to reversible, irreversible, and kinetic systems. Anal. Chem. 1964, 36, 706-723. 12. Girault, H. H. Electrochimie Physique et Analytique; Presses polytechniques et universitaires romandes: Lausanne, 2007; pp 1-572. 13. Pletcher, D. First Courses In Electrode Processes; The Royal Society of Chemistry: Cambridge, 2009; pp 1-298. 14. Bott, A. W. Characterization of chemical reactions coupled to electron transfer reactions using cyclic voltammetry. Curr. Sep. 1999, 18, 9-16. 15. Bott, A. W. Practical problems in voltammetry 3: reference electrodes for voltammetry. Curr. Sep. 1993, 14, 64-68. 16. Giacomelli, C.; Ckless, K.; Galato, D.; Miranda, F. S.; Spinelli, A. Electrochemistry of caffeic acid aqueous solutions with pH 2.0 to 8.5. J. Braz. Chem. Soc. 2002, 13, 332-338. 17. Hotta, H.; Ueda, M.; Nagano, S.; Tsujino, Y.; Koyama, J.; Osakai, T. Mechanistic study of the oxidation of caffeic acid by digital simulation of cyclic voltammograms. Anal. Biochem. 2002, 303, 66-72. 18. Bond, A. M.; Mclennan, E. A.; Stojanovic, R. S.; Thomas, F. G. Assessment of conditions under which the oxidation of ferrocene can be used as a standard

156

Chapter 4

voltammetric reference process in aqueous-media. Anal. Chem. 1987, 59, 2853-2860. 19. Tsierkezos, N. G. Cyclic voltammetric studies of ferrocene in nonaqueous solvents in the temperature range from 248.15 to 298.15 K. J. Sol. Chem. 2007, 36, 289-302. 20. Liu, J.; Castro, R.; Abboud, K. A.; Kaifer, A. E. Novel ferrocenyl polyene derivatives and their binding to unmodified cyclodextrins. J. Org. Chem. 2000, 65, 6973- 6977. 21. Hudson, R. D. A. Ferrocene polymers: current architectures, syntheses and utility. J. Organomet. Chem. 2001, 637, 47-69. 22. Batterjee, S. M.; Marzouk, M. I.; Aazab, M. E.; El-Hashash, M. A. The electrochemistry of some ferrocene derivatives: redox potential and substituent effects. Appl. Organometal. Chem. 2003, §17, 291-297. 23. Namazian, M.; Zare, H. R. Electrochemistry of chlorogenic acid: experimental and theoretical studies. Electrochim. Acta 2005, 50, 4350-4355. 24. Nicholson, R. S.; Shain, I. Theory of stationary electrode polarography for a chemical reaction coupled bewteen two charge transfers. Anal. Chem. 1965, 37, 178-189. 25. Zare, H. R.; Golabi, S. M. Caffeic acid modified glassy carbon electrode for electrocatalytic oxidation of reduced nicotinamide adenine dinucleotide (NADH). J. Solid State Electrochem. 2000, 4, 87-94. 26. Corujo-Antuna, J. L.; bad-Villar, E. M.; Fernandez-Abedul, M. T.; Costa-Garcia, A. Voltammetric and flow amperometric methods for the determination of melatonin in pharmaceuticals. J. Pharm. Biomed. Anal. 2003, 31, 421-429. 27. Radi, A.; Bekhiet, G. E. Voltammetry of melatonin at carbon electrodes and determination in capsules. Bioelectrochem. Bioenerg. 1998, 45, 275-279. 28. Laviron, E. Electrochemical reactions with protonations at equilibrium. Part X. The kinetics of the p-benzoquinone/hydroquinone couple on a platinum electrode. J. Electroanal. Chem. 1984, 164, 213-227. 29. Laviron, E. Electrochemical reactions with protonations at equilibrium .12. the 2E-, 2H+ homogeneous isotopic electron exchange-reaction (9-member square scheme). J. Electroanal. Chem. 1984, 169, 29-46. 30. Laviron, E. Electrochemical reactions with protonations at equilibrium .13. Experimental-study of the homogeneous electron exchange in quinone dihydroquinone systems. J. Electroanal. Chem. 1986, 208, 357-372. 31. Desbenemonvernay, A.; Lacaze, P. C.; Cherigui, A. Uv-visible spectroelectrochemical study of some para-benzoquinoid and ortho- benzoquinoid compounds - Comparative-evaluation of their electrochromic properties. J. Electroanal. Chem. 1989, 260, 75-90. 32. Bailey, S. I.; Ritchie, I. M. A cyclic voltammetric study of the aqueous electrochemistry of some quinones. Electrochim. Acta 1985, 30, 3-12. 33. de Astudillo, L. R.; Rivera, L.; Brito-Gomez, R.; Tremont, R. J. Electrochemical study of 1,4-benzoquinone on gold surface modified. J. Electroanal. Chem. 2010, 640, 56-60. 34. Nematollahi, D.; Amani, A.; Tammari, E. Electrosynthesis of symmetric and highly conjugated benzofuran via a unique ECECCC electrochemical mechanism: Evidence for predominance of electrochemical oxidation versus intramolecular cyclization. J. Org. Chem. 2007, 72, 3646-3651. 35. Rafiee, M.; Nematollahi, D. Voltammetry of electroinactive species using quinone/hydroquinone redox: A known redox system viewed in a new perspective. Electroanalysis 2007, 19, 1382-1386.

157

Chapter 4

36. Yamamoto, K.; Asada, T.; Nishide, H.; Tsuchida, E. The preparation of poly(dihydroxyphenylene) though the electro-oxidative polymerization of hydroquinone. Bull. Chem. Soc. Jpn. 1990, 63, 1211-1216. 37. Carter, M. K. Correlation of electronic transitions and redox potentials measured for pyrocatechol, resorcinol, hydroquinone, pyrogallol, and gallic acid with results of semi-empirical molecular orbital computations - A useful interpretation tool. J. Mol. Struct. 2007, 831, 26-36. 38. Zoski, C. G. Handbook of Electrochemistry; Elsevier: Amsterdam, 2007; pp 1-913. 39. Nesmeyanov, A. N.; Romanenko, V. I.; Sazonova, V. A. Basicity constants of ferrocene amines. Izvestiya Akademii Nauk SSSR 1966, 2, 357-358.

158

Chapter 5

Chapter 5 Structure activity relationship

5-1 Introduction

The radical reducing mechanism is widely studied in order to work out the behavior of an antioxidant compound. It provides the agent abstracted from the structure which is used to reduce the radical and the resulting rearrangement. This new structure is of great interest since it must not be a toxic compound and therefore not induce other damages. For instance, quercetin is a well-known antioxidant compound and ranked among the best ones. However, its oxidized products may induce toxic effects. Indeed, besides acting as an antioxidant, the flavonoid may undergo not only an auto-oxidation but also enzymatic reactions. The resulting o-semiquinone and o-quinone have been revealed to be involved in pro-oxidant effect-induced cellular damages1. O-quinone is also involved in the depletion of glutathione through the formation of an adduct which cannot be prevented by ascorbate. Furthermore it is involved in the interaction with other thiol groups of proteins which leads to structural disorders2,3. The mode of action of antioxidant compounds also depends on the media and pH. Bond dissociation enthalpy (BDE) and ionization potential (IP) are two relevant quantum chemical concepts when concerning the mode of action of antioxidants. The lower the BDE is, the more the hydrogen atom transfer (HAT) is enabled. On the other hand, the lower the IP is, the more the single electron transfer (SET) occurs4,5. It has been established that a HAT is the main mechanism in aqueous solution for a ΔBDE of about 10 kcal/mol and a ΔIP equal to -36 kcal/mol (ΔBDE and ΔIP are given relative to the BDE and the IP of phenol). A larger ΔIP (about -45 kcal/mol) leads to a SET predominance5. However, these calculations are most often performed in gas-phase and in benzene and do not provide specific results when it mimics an aqueous media

159

Chapter 5 reaction. The association with in vitro assays such as the ALP, ORAC, DPPH•, and ABTS•– methods will support the predicted BDE and IP parameters. The four assays used for the investigation of the compounds provide clues about the antioxidant mechanism of the compounds. Indeed, DPPH• is supposed to be scavenged by a hydrogen atom transfer (HAT) from the antioxidant compound (Figure 5-1). However, a second mechanism has been described when the reaction takes place in a polar solvent such as methanol and ethanol. The mechanism is named sequential proton loss electron transfer (SPLET)6. The solvent induces the compound ionization (usually of a hydroxyl group), which leads to an electron transfer to the DPPH• radical. Finally, the resulting DPPH– is protonated (Figure 5-1). The rate constant of SPLET mechanism is much faster than that of HAT mode of action. A compound may be revealed as a DPPH• scavenger, while it would not be active without the solvent ionizing effect. SPLET mechanism may be avoided by performing the reaction in acidic medium since the ionization will not occur7.

HAT AH + DPPH• A• + DPPH-H Slow

H+ -H+ H+ SPLET A– + DPPH• A• + DPPH– Fast

Figure 5-1: DPPH• scavenging through HAT and SPLET mechanisms6.

The detoxification of the peroxyl radicals mainly follows the same mechanism as that to scavenge DPPH• radicals. They are detoxified by hydrogen atom transfer or by a proton transfer-concerted electron transfer5,8-10. The methods which use the peroxyl radicals will provide further clues concerning the mechanism since they are performed at a set pH. The pKa values of the compounds are then a useful parameter. The reaction ABTS•– reduction is not well-defined; both HAT11-14 and ET10,15- 18 hypothesizes are put forward. It has been mentioned19 that the reaction rate and the behavior of DPPH• and ABTS•– scavengers differs according to the

160

Chapter 5 radical due to the steric hindrance. The scavenging activity of caffeic acid was more potent toward DPPH• than that against ABTS•–, whereas ferulic acid (a methoxylated form of caffeic acid) interacts more forcefully with ABTS•– than it does with DPPH•. This has been attributed to the methoxy group which increases the steric hindrance of ferulic acid. However, the lack of the hydroxyl residue on the structure of this compound may lead to a HAT ability drop without producing any change in the electron transfer capacity. This study would therefore be in favor of an electron transfer-scavenged ABTS•–. By performing the ABTS•– assay in ethanol and then in a mixture of ethanol/acetic acid, the results will provide data concerning the mode of action necessary to reduce the ABTS•– radical. The diversified compound structures and the mechanisms, which vary in the methods and according to the solvent conditions, provide well-defined structure- activity relationships (SAR). The mode of action of the compounds is a further characteristic for the antioxidant profiles.

5-2 Materials and methods

5-2.1 Materials

Acetic acid was purchased from Sigma-Aldrich (Buchs, Switzerland). The set was completed with N,N-diethylaniline purchased from Fluka (Buchs, Switzerland) and quercetin-derivatives extracted and isolated from Flavo Psidium Cattleianium in the group of Prof. Jean-Luc Wolfender and Prof. Muriel Cuendet (Laboratoire de Pharmacognosie, Phytochimie et Produits Naturels Bioactifs. Geneva, Switzerland). The other compounds and materials are described in Chapter 2 (section 2-2.1 Materials).

161

Chapter 5

5-2.2 Methods

To emphasize the scavenging mechanism toward DPPH• and ABTS•–, the active compounds have been assessed in acidic conditions defined as appropriate to delete the SPLET mechanism7.

5-2.2.1 DPPH• assay

The reaction was performed as described before (section 2-2.2.3) by replacing ethanol by a mixture of EtOH/acetic acid (10 mM, 1%).

5-2.2.2 ABTS•− assay

The formation of ABTS•– radical was carried out in water as described in the section 2-2.2.4. Since the ABTS•– reducing mechanism is unclear, a SPLET may occur if the mechanism follows a hydrogen transfer. Therefore, the determination of the ABTS•– radical anion concentration and the following steps assessing the free radical scavenging activity were performed in a mixture of EtOH/acetic acid (10 mM, 1%).

5-3 Results and discussion

The structure-activity relationship is described for the compounds according to their antioxidant behavior assessed in the four assays performed in this study. It should be reminded that the ALP and ORAC methods were carried out at pH = 8.3 which is the optimal pH of the ALP protein to conserve the catalytic activity. Therefore, the peroxyl radical detoxifying activity of some compounds may differ from that reported in the literature, since the assays are most often performed at the physiological pH (pH = 7.4). Depending on the pKa values, a compound may not have the same behavior at both pH’s.

162

Chapter 5

• •– In addition to that, the ER50 values obtained from DPPH and ABTS assays may also differ from those reported in the literature, since the reaction time, the radical concentration, and the solvent, among other conditions, often vary with the performed methods.

5-3.1 Reference compounds

5-3.1.1 Caffeic acid (profile A2)

The intermediate DPPH• scavenging rate of caffeic acid (CA) makes it be characterized by the antioxidant profile A2, which is composed of both the most rapidly acting DPPH• scavengers (including class 3 and 2) and the most potent antioxidant compounds (class 3). The 3,4-dihydroxycinnamic acid does not have the best potency values but all of them are ranked in the class 3. The pEC50 value equals 5.66 and 5.65 in the ALP and ORAC assays, respectively, and the ER50 value is of 0.17 in both the DPPH• and ABTS•– methods. The ability of caffeic acid to donate electron and hydrogen atoms is related to the low oxidation potential of the catechol moiety (Epa = 0.23V (vs SCE) at pH = 7.420) as well as to the low OH bond dissociation energy (BDE = 73.95 – 78.7 kcal/mol)9,21 and the low ionization potential (IP = 125.87 kcal/mol for the parent molecule in ethanol and IP = 126.33 kcal/mol for the anion in buffer21). These abstractions yield the semi-quinone radical as a first intermediate and then the quinone structure (Figure 5-2)22,23.

O O

OH OH HO O O O

Figure 5-2: Semi-quinone radical (left) and quinone (right) structures of caffeic acid.

163

Chapter 5

The free radical scavenging capacity is consistent with these features and with the literature10,19. The deprotonated structure of caffeic acid contributes to the great potency toward peroxyl radicals in the ALP and ORAC assays. At pH = 8.3, the carboxylic group is deprotonated and the 4-OH residue is partially deprotonated, 58% is in the form CA– and 42% is in the form CA2– (Figure 5-3)22.

– The pKa’s of caffeic acid determined by potentiometry are 4.83 (-COOH/-COO ), 8.9 (4-OH/4-O–) and 10.98 (3-OH/3-O–)24.

HO O

O O HO HO O O

Figure 5-3: Structure of caffeic acid at pH = 8.3.

Cyclic voltammetry studies20,25 have reported that the oxidation potential of caffeic acid shifted along more negative potential when the pH increased, which means that the antioxidant activity was enhanced. Same results were found in a study about the caffeic acid behavior in acidic and basic media9. The phenolate anion was more potent than the parent molecule because the 4-O– activated the electron or hydrogen atom transfer from the 3-OH, that leads to the quinone byproduct. The double bond on the side chain is a further structural feature which enhances the antioxidant activity since it reduces the OH-bond strength9,26. Besides the antioxidant residues, oxidized caffeic acid seems to undergo polymerization reactions yielding products with antioxidant capacities10,23,27. Dimers of caffeic acid have been identified in a mixture of acetate buffer pH = 3.5 and 12% ethanol, but did not exist at pH higher than 4.520. The fact that caffeic acid is capable of scavenging the DPPH• radical more rapidly than the other potent compounds might be attributed to its smaller structure; it enables the interaction with the hindered radical. The potency of caffeic acid in acidic conditions does not significantly differ from that in ethanol but it reacts more slowly with the DPPH• radical. Therefore, caffeic acid is potent enough to scavenge the free radical even in acidic conditions, which emphasizes a HAT

164

Chapter 5 mechanism. In ethanol, both HAT and SPLET mechanism are concerted, the latter induces a faster DPPH• scavenging activity. The weaker activity, but not relevant, toward ABTS•– radical in acidic media may highlight an ET mode of reducing action in this assay. But, no evidence concerning the mechanism can be drawn from the ABTS•–-related results.

5-3.1.2 Ascorbic acid (profile A8)

Ascorbic acid and trolox® turned out to have the same behavior toward the three radicals. Therefore, they both are defined by the profile A8, the properties of which are a rapid and intermediate DPPH• scavenging rate and an intermediate DPPH• trapping potency as well as a weak capacity for detoxifying the peroxyl radicals. Vitamin C is required for a good health. Besides regenerating vitamin E and

28 •28 other compounds, it has a key role in the detoxification of H2O2 , OH , and ONOO–29 among other ROS and RNS. Ascorbic acid was the most rapidly acting DPPH• scavenger in both ethanol and acidic media. It mostly achieved the radical trapping within the first minute; 96% and 82%, respectively, of DPPH• radical was reduced in a minute. The reaction rate of ascorbic acid toward DPPH• was expected. It has been described as a more rapidly acting scavenger than other compounds and even than polyphenols30 and α-tocopherol31. The rapid activity has been attributed to the only two oxidizable OH residues, which lead the compound to reach equilibrium in a minute10. Besides, a cyclic voltammetry study32 has reported the low oxidation potential (Epa = 0.282V at pH = 7 vs. SHE, Epa = 0.211V vs Ag/AgCl at pH = 3.6). This confirms that ascorbic acid oxidizes before polyphenols (for example, Epa of caffeic acid is equal to 0.410V in the same study (pH = 3.6)). However, despite the rapid reaction, ascorbic acid showed weaker activity than reference polyphenols (except resveratrol) with an

33,34 ER50 value of 0.28, which is close to the literature . It suggests that 1 mole of ascorbic acid scavenges 3 or 4 moles of DPPH•. Hydrogen abstraction from ascorbic acid to DPPH• leads to the formation of semi-dehydroascorbic acid

165

Chapter 5 followed by the formation of the dehydroascorbic acid. At this step, two hydrogen atoms are transferred to the free radical (Figure 5-4).

OH OH

HO O HO O O O

HO O O O

Figure 5-4: Structure of semi-dehydroascorbic acid (left) and dehydroascorbic acid (right).

A rearrangement occurs in alcoholic solvent inducing the formation of the bicyclic dehydroascorbic acid by the solvation of C-2 carbonyl group31. Two further hydrogen atoms may be abstracted from hydroxyl residues in position 2 and 3 leading to a 1,2-dioxethane residue (Figure 5-5).

HO O HO O O O

OEt OEt O OH O HO O O

Figure 5-5: Structures of the solvated bicyclic dehydroascorbic acid (left) and bearing the 1,2-dioxethane residue (right)

The higher activity in acidic media (ER50 = 0.14) has been reported before and has been attributed to the regeneration of ascorbic acid in protic media34. The

•– ER50 value toward ABTS in ethanol (ER50 = 0.30) also suggests that 1 mole of ascorbic acid scavenge 3 mole of free radical. However, the similar ABTS•– reductive activity in acidic condition (ER50 = 0.33) implies that the scavenging mechanism is not the same toward both free radicals. Therefore, the electron transfer appears to be the main mechanism for the ATBS•– reduction. The weak peroxyl radical-reducing activity is likely due to the pH used in the ALP and ORAC assays. A pH-dependant antioxidant ability has been highlighted34,35. The capacity is reduced when pH increases. This is consistent with the results obtained from the DPPH• assay. However, pH does not affect the ABTS•–

166

Chapter 5 scavenging ability13 which is in line with the results from the ABTS•– assay in this study. The relevant behavior of ascorbic acid toward the three radicals and in three media (buffer pH = 8.3, EtOH, and EtOH/AcOH) emphasizes two distinct mode of action with ABTS•– on one hand and with DPPH• and peroxyl radicals on the other hand. Moreover, it validates the pH-dependant HAT mechanism in the interaction with DPPH• and ROO• since vitamin C is a good peroxyl radicals scavenger at pH = 7.429. The obvious electron transfer mechanism involved in the ABTS•– reduction is consistent with the electron-donating ability of ascorbic acid36.

5-3.1.3 Trolox® (profile A8)

Trolox® is the water soluble analog of vitamin E. Its antioxidant properties have been proven toward ROO• and ONOO–29 as well as a rapid DPPH• scavenging activity12,37,38. It turned out to be an ascorbic acid-like compound. Their potency parameters and their antioxidant behavior in acidic conditions are similar. The only difference is the slower DPPH• scavenging rate of trolox® (59% of scavenged free radical beside 96% with ascorbic acid). The results from DPPH• and ABTS•– methods in ethanol and in a mixture ethanol/acetic acid depict two

• distinct trapping mechanisms. Indeed, the ER50 value in DPPH assay (ER50 =

•– 0.21) is close to that in ABTS assay (ER50 = 0.28) in ethanolic solvent. However,

• the activity toward DPPH was enhanced in acidic conditions (ER50 = 0.11) when

•– • it was not affect with ABTS radical (ER50 = 0.27). As concerns the DPPH scavenging, the good results emphasize a HAT mechanism. The resulting structure after the abstraction of a hydrogen atom is the phenoxyl radical (Figure 5-6) which reacts with a second DPPH• molecule or may undergo a diproportionation or a dimer formation38-40. Moreover, the stabilization of the phenoxyl radical is drastically enhanced in basic media and reduced in acidic media40; hence the accelerated disproportionation in presence of protons. This reaction releases a molecule of trolox® capable of scavenging another DPPH• radical.

167

Chapter 5

O O

Figure 5-6: Structure of the phenoxyl radical of trolox®.

A low pH would be in favor of the HAT mechanism. The weak activity in the ALP

(pEC50 = 4.84) and in the ORAC assay (pEC50 = 4.74) validates the pH-dependant HAT mechanism41,42. Trolox® is known not only as a hydrogen atom donor41 but also as an electron donor38,42. This latter ability may be the main mechanism with ABTS•– radical since the mechanism involved in this reaction appears

• ® different from that with DPPH . Indeed, the pH did not affect the ER50 of trolox . The reaction rate12,37 and the behavior in acidic conditions13 are in agreement with the literature. An interaction through an electron transfer from trolox® to ABTS•– would lead to the formation of the phenoxyl radical cation which readily loses a proton40. The phenoxyl radical is capable of transferring a second electron and yields the phenoxonium ion. This one has been reported to react with OH– in basic media to form a chromanone or a quinone moiety after the heterocyclic ring opening (Figure 5-7)40.

OH O O O O OH O O

O O

Figure 5-7: Structures of the chromanone (left) and quinone (right) forms of trolox® in basic media after the hydrogen atom loss.

These oxidation products are not capable of interacting with ABTS•–37. In alcoholic conditions, a nucleophile addition occurs from an attack by the solvent. The addition is favored in acidic conditions. The by-product is not able to reduce

168

Chapter 5 the ABTS•– either. The trolox® behavior points to a pH-dependant HAT mechanism involved in the detoxification of DPPH• and peroxyl radical, and a pH-independent ET mechanism scavenging the ABTS•– radical. The former leads to the formation of a phenoxyl radical intermediate while the second yields a phenoxonium ion intermediate. The results from ORAC method are in agreement with the literature43 relative to that of caffeic acid, quercetin, and resveratrol.

44 Despite its low oxidation potential (Epa = 0.11V vs Ag/AgCl) , the lack of several hydroxyl groups on the structure contributes to a weaker antioxidant activity

® than that of polyphenols. The similar pEC50 in the ALP assay shows that trolox interacts with peroxyl radicals without any involvement with the protein.

5-3.1.4 Chlorogenic acid (profile B1)

The three reference compounds included in the profile B1, namely chlorogenic acid, gallic acid, and resveratrol, are defined by a high and intermediate potency toward the radicals and a scavenging rate which is slow and very slow against DPPH•, and intermediate and slow with ABTS•–. Chlorogenic acid is an ester of caffeic acid; it bears a quinic acid moiety (Figure 5-8). Their antioxidant parameters in the four assays are close and even

• •– equal in the ALP and ORAC assays. The ER50 in both the DPPH and ABTS assays are in line with reported studies for chlorogenic acid10,45,46.

OH OH

HO HO O O O O

HO HO O O O O OH OH HO O

Figure 5-8: Structures of chlorogenic acid at pH = 8.3 (~50/50).

However, chlorogenic acid and caffeic acid do not belong to the same profile. This is due to the slower DPPH• scavenging rate of chlorogenic acid (15% of trapped

169

Chapter 5

DPPH•) relative to that of caffeic acid (50% of DPPH• scavenged in a minute). The quinic acid moiety on the chlorogenic acid structure induces a more significant steric hindrance which reduces the interaction with DPPH• radical, hindered itself. This hypothesis has already been put forward in the reaction with ABTS•– radical47. Despite the steric hindrance, chlorogenic acid is a potent

• DPPH scavenger. Besides, the results in acidic conditions (ER50 = 0.16) are close to that in ethanol (ER50 = 0.11). This similarity points that the interaction with DPPH• radical is mainly induced by a HAT mechanism. However, the rather

• •– significant gap between the ER50 obtained from the DPPH and ABTS methods suggest either two modes of action or the steric hindrance involvement. A slow rearrangement of the oxidized chlorogenic acid has been proven in a cyclic voltammetry study while nothing occurs for caffeic acid44. The newly formed

•– product may be less potent toward ABTS radical, hence the higher ER50 value than that of caffeic acid. This hypothesis suggests an electron transfer from chlorogenic acid to ABTS•– and then two distinct mechanisms toward both stable radicals. The great potency of chlorogenic acid is attributed to the catechol moiety and to its ability to form dimer with an o-hydroquinone moiety or polymers after interaction with free radicals10,48.

5-3.1.5 Gallic acid (profile B1)

Gallic acid (3,4,5-trihydroxybenzoic acid) had an intermediate capacity for protecting ALP protein and fluorescein against peroxyl radical-induced damages

• •– and it showed the highest DPPH and ABTS scavenging potency (ER50 = 0.04 and 0.06, respectively). The activity rate toward DPPH• radical is slower than that of compounds from kinetics cluster A (38% of radical scavenged over a minute, which corresponds to the class 1) and it reacts more rapidly with ABTS•– radical (71%). This ability has been reported before10,19,49.

170

Chapter 5

O O

HO HO O O

HO O OH OH

Figure 5-9: Structures of gallic acid at pH = 8.3 (~80/20).

Besides the three OH groups, the small structure makes the interaction with DPPH• and ABTS•– easier. It has been proven that gallate derivatives are less potent toward these free radicals due to the steric effect50,51. Furthermore, oxidation products from gallic acid undergo a polymerization, which yields more trapped DPPH• than the number of free OH residue on the structure52. In acidic

• •– media, ER50 values in both the DPPH and ABTS methods are similar to that in ethanol (ER50 = 0.05). These results confirm a HAT mechanism involvement in the DPPH• (and thus ROO•) detoxification, without any concerted SPLET mechanism, that is in line with literature4,49. Gallic acid owes its high scavenging activity to the three hydroxyl groups in position 3, 4, and 5 as well as to the COOH moiety, an electron-withdrawing group, and finally to the small structure (Figure 5-9). The first hydrogen atom is abstracted from the 4-OH group. The freshly produced symmetric species is stabilized by the electron-donating OH in ortho position (3 and 5) and by the COOH in para position4. Then the quinone is formed and further oxidation reactions occur, which lead to the polymerization. Gallic acid is not only a great hydrogen atom donor but also an electron donor4. These abilities are involved in the ABTS•– reduction. The potency in both conditions does not provide any clue as concerns the mechanism; however, the trihydroxybenzoic acid is the most rapidly acting ABTS•– scavenger (71% scavenged ABTS•–) and even much faster in acidic conditions (93% scavenged ABTS•–). These results imply two hypotheses. The first one is that the mechanism is a hydrogen atom transfer and the weaker stability of ABTS•– and its less hindered structure than those of DPPH• enable the interaction. The second hypothesis is that the mode of reduction differs from that with DPPH•, this hypothesis therefore suggests an electron transfer. The ionization potential

171

Chapter 5

(138.69 kcal/mol4) is not low enough to produce such a potent and rapid interaction. However, the acidic conditions would enable the proton loss after the electron transfer which rapidly yields further oxidation reactions. A relevant distinction between HAT and ET is not possible from these results, but the difference between both free radicals in the stability and in the steric hindrance is obviously involved in the reaction rate. As expected according to the free DPPH• radical scavenging activity, gallic acid is also a potent peroxyl radical scavenger. The pEC50 values obtained from the ALP and ORAC assays are similar; gallic acid therefore directly reacts with ROO• radicals without any interaction with the ALP protein. The location of the trihydroxybenzoic acid inside the kinetic cluster B is due to its mechanism through a HAT which is slower than the SPLET. The latter occurs as the first mechanism for the compounds which belong to the cluster A, such as caffeic acid.

5-3.1.6 Resveratrol (profile B1)

The remaining reference compound characterized by the profile B1 is resveratrol.

OH OH

HO O OH OH

Figure 5-10: The two major form of resveratrol at pH = 8.3.

This polyphenol (3,4’,5-trihydroxy-trans-stilbene) is the less potent DPPH•

• scavenger of this group (ER50 = 0.46 and 15% of DPPH scavenged in a minute) but it is among the most potent antioxidant compounds assessed with the ALP, ORAC, and ABTS•– methods, with an intermediate ABTS•– scavenging rate. Resveratrol is a known free radical scavenger19,28,53. The activity decrease toward

• DPPH radicals in EtOH/AcOH (ER50 = 0.7 when ER50 = 0.46 in EtOH) reveals

172

Chapter 5 that a SPLET mechanism mainly occurs to trap the radical. On the other hand, the similar ABTS•– reducing ability in both media highlights an ET mechanism. Both conjugated resorcinol and phenol moieties are involved in the reducing activity. They induce a planar structure which enables the electron delocalization. The phenol moiety is oxidized before the resorcinol residue54,55. Indeed, the most stable radical is formed by a hydrogen loss from the 4’-OH group of the p-monophenol core4 (Figure 5-10). These results and the very low ionization potential (IP = 124.18 kcal/mol4,5) confirm that the antioxidant activity toward the three radicals studied here is mainly due to an ET mechanism and confirms the electron transfer-induced ABTS•– reduction. A hydrogen atom transfer may also occur but more slowly. The high potency toward peroxyl and ABTS•– radicals is caused by the oxidation products which undergo a polymerization28,54,55. Furthermore, the oxidation product from the reaction between resveratrol and DPPH• has been identified as a dimer of oxidized resveratrol53 (Figure 5-11).

O HO OH

HO OH OH

Figure 5-11: Dimer of oxidized resveratrol.

5-3.1.7 Mangiferin (profile B2)

The antioxidant profile B2 refers to mangiferin and quercetin; the characteristics of which are to be potent DPPH• and ABTS•– scavengers and to be capable of protecting both ALP protein and fluorescein against peroxyl radical- induced damage. The high potency of mangiferin has been revealed in the four assays.

Besides, it is among the most potent studied compounds since its pEC50 value is

173

Chapter 5

5.74 in the ALP assay, 6.04 in the ORAC assay, and its ER50 value is 0.12 and 0.11 in the DPPH• and ABTS•– assays, respectively. However, its interaction with DPPH• is slow, it scavenges only 15% of this stable radical in a minute at a ratio of 0.5. The scavenging rate is faster toward ABTS•– since 64% of ABTS•– is scavenged in a minute.

HO O OH O 8 1 OH HO 7 2

6 HO O O 3 OH 5 4 OH

Figure 5-12: Major form of mangiferin at pH = 8.3.

The catechol moiety in position 6,7 is known to play an important role in the antioxidant activity, thanks to its low oxidation potential at 0.32 V (vs Ag/AgCl) at pH = 7.444. Furthermore, the hydrogen atom abstraction from the mangiferin structure also requires a low energy (80.41 kcal/mol) attributed to the catechol moiety as well as the 3-OH group56. Acidic conditions do not affect the free radical scavenging activity of the xanthone-derivative. Indeed, the ER50 values toward DPPH• equal 0.12 and 0.09 in ethanol and in ethanol/acetic acid mixture, respectively. These results imply that a hydrogen atom transfer mainly occurs for the DPPH• reduction from mangiferin. This agent transfer is also confirmed by the slow interaction between mangiferin and DPPH•. A SPLET takes place more rapidly than a HAT7, hence the faster rate in ethanol if a SPLET occurred, especially with an active group such as catechol. However, the peroxyl as well as ABTS•– radicals seem to be reduced by electron transfer. Indeed, the ALP and

ORAC assays are carried out at pH = 8.3. The pKa value equal to 6.52 (6-OH), 7.97 (3-OH), 9.44 (7-OH), and 12.10 (1-OH) have been reported57,58. Thus, 6-OH is deprotonated and 3-OH as well as 7-OH groups are partially deprotonated at this pH value. According to the great pEC50 values, a hydrogen atom transfer from mangiferin structure may not be the main mechanism and the peroxyl radicals are more likely detoxified through a proton-transfer-concerted electron transfer.

174

Chapter 5

This mechanism is in agreement with the efficient ABTS•– scavenging activity. This radical is supposed to be mainly reduced by ET; despite it is still under discussion. Furthermore, the faster reductive ability toward ABTS•– radical compared with that of DPPH• radical tend to emphasize the electron transfer from mangiferin prior to HAT. This is in agreement with a study which reported that mangiferin mainly reacts by electron transfer with ABTS•–, OH•, and

• 58 CCl3O2 (a peroxyl radical model) which leads to a phenoxyl radical structure . As a result, mangiferin has the ability to scavenge free radicals through both HAT and ET.

5-3.1.8 Quercetin (profile B2)

Quercetin (3,3’,4’,5,7-pentahydroxyflavone) is the second reference compound of the profile B2. It is a flavonol mostly found in plants, onions, berries, cherries, tea, and apples, among many other flavonoids. Many studies have reported its antioxidant capacities such as the radical-scavenging properties10,12,19, the ability to inhibit the lipid peroxydation44,59, the metal ion chelating power60-62.

5' 4' OH 6' B 7 1 3' O O 29 1' OH 6 2' A C 8 5 4 3 OH OH O

Figure 5-13: Predominant form of quercetin at pH = 8.3.

Quercetin owes its great antioxidant power to the catechol moiety on the B-ring, to the OH group on C3 as well as the 2,3-double bond (Figure 5-13). These three structural residues involve a conjugation over both B- and C-rings which stabilizes the produced radical60,61,63,64. It belongs to the profile B2, namely it possesses an efficient antioxidant activity toward DPPH•, ABTS•–, and peroxyl

175

Chapter 5 radicals, but slowly scavenge the first radical. Quercetin is known to undergo disproportionation, rearrangement, and reaction with the alcoholic solvent after the first H-abstraction by DPPH• radical which leads to an ability to scavenge many radical molecules per molecule of quercetin (Figure 5-14). Indeed, besides parent flavonol, by-products from quercetin oxidation are capable of transferring a hydrogen atom to DPPH•. The final products are 2,4,6-trihydroxybenzaldehyde and 3,4-dihydroxybenzoic acid (protocatechuic acid)63,64. This product set explains the low ER50 value of 0.07 as well as its high potency in detoxifying the peroxyl radicals through hydrogen atom transfer (pEC50 value equals 5.95 and 5.78 in the ALP and ORAC assays, respectively). This ability has been proven before26,43,65,66.

O OH

OEt HO O HO O O 2 EtOH OH OEt

OH OH OH O OH O

Figure 5-14: One of the rearrangements of oxidized quercetin63.

5-3.1.9 Cysteine (profile B8)

The only compound characterized by the antioxidant profile B8 is cysteine. It differs from the other compounds of the kinetic cluster B in its weak activity toward the DPPH• and peroxyl radicals, and from ascorbic acid and trolox® (A8) in its very slow DPPH• scavenging capacity. Since all remaining compounds failed to scavenge DPPH• radical, cysteine is not ranked among them, even

• though it is a slow (19% of DPPH scavenged in a minute) and weak (ER50 = 0.76) stable radical scavenger. Thiol-compounds have been reported as DPPH• and ABTS•– reducing compounds67,68. The higher potency toward ABTS•– (close to that of polyphenol) relative to that toward DPPH• may be due to the better stability of this latter but also implies two mechanisms. DPPH• is known to abstract a hydrogen atom. The thiol group would be the hydrogen atom donor. As

176

Chapter 5 regards ABTS•– radical, the main mechanism would be an electron transfer. Both thiol and amine groups have the ability to act as electron donor. When pH is decreased, the scavenging activity drastically decreased toward ABTS•– but enhanced with DPPH• radical, which yields the same activity in both methods

(ER50 = 0.42). The protonated nitrogen atom of the amino acid residue loses its electron donating ability. Only the thiol is able to play this role. A proton- concerted electron loss seems more appropriate, actually. Once oxidized, after hydrogen atom (or electron and proton) abstraction from the thiol residue, two molecules of cysteine form a disulfide bridge (Figure 5-15). This bond is susceptible to scission in presence of nucleophile such as EtOH and therefore yields a new molecule of cysteine69.

NH2 NH2 O 2 HO S HO S S OH

O O NH2

Figure 5-15: Formation of a disulfide bond.

The weak ALP and fluorescein protecting capacity may be attributed to the lack of antioxidant reactive groups such as a polyphenol structure. This is in agreement with the weak DPPH• scavenging activity. Cysteine does not seem to be involved with the protein since the pEC50 values are similar in the ALP and ORAC assay.

+ + NH3 NH3 O SH O S

O O

Figure 5-16: Ionized forms of cysteine at pH = 8.3 (~40/60).

177

Chapter 5

5-3.1.10 Melatonin (profile C5)

Melatonin has been widely under investigations which have provided

• • – 70 evidences of its potent ability to interact with OH , H2O2, NO , ONOO , HOCl , and peroxyl radicals15,71, even though its ROO• scavenging activity is controversial72. In this study, Melatonin (C5) has revealed an intermediate activity in protecting the protein and fluorescein against the peroxyl radicals. Its pEC50 value equals 5.20 and 5.28 in the ALP and ORAC assay, respectively. Electrochemical13 and theorical56,73 studies among others15 have emphasized that the indolamine, which is an electron-rich structure, scavenge reactive species through electron transfer. A stable radical cation is formed after the electron

56 •– loss . This capacity makes it able to reduce ABTS radical. The ER50 value of 0.32 points out that one melatonin molecule interacts with three ABTS•– molecule that is in agreement with studies performed before13,15 but it acts very slowly. Even though they are less potent13,17, the by-products formed from the melatonin oxidation enhance the antioxidant activity. These products are the cyclic 3-hydroxymelatonin (c3OHM) and N1-acetyl-N2-formyl-5- methoxykynuramine (AFMK). A deformylated metabolite is produced from AFMK; it is named the N1-acetyl-5-methoxykynuramine (AMK) (Figure 5-17).

H H O N N

HN N O O O HO Melatonin c3OHM

H N NH2 CHO O O

O N O N H H AFMK O AMK O

Figure 5-17: Structures of melatonin and of its by-products resulting from the melatonin oxidation.

178

Chapter 5

Melatonin is slightly less active toward ABTS•– molecules in acidic condition as it was reported before13. The lone pair of nitrogen is involved in a hydrogen bond; the electron transfer is therefore reduced. Despite a high peroxyl radical detoxification, the indolamine failed to scavenge the DPPH• radical. This is due not only to the fact that it acts through electron transfer but also to the radical- related stability. ROO• radicals are more reactive than the DPPH•; their

0 74 reduction potential range is high (0.77 V < E ROO•/ROOH < 1.44 V ) when the reduction peak potential of DPPH• is around 0.22V75. The cyclic voltammetry study using the ferrocene derivatives have shown that the formal potential of melatonin is higher than that of ferrocenecarboxylic acid (313 mV). A cyclic voltammetry carried out in phosphate buffer (pH = 7.4) yields an oxidation peak for ABTS of 0.498 V and a reduction peak for ABTS•+ of 0.4 V (vs calomel electrode (3.5 M KCl))76 as well as a formal potential of 0.6 V (vs Ag/AgCl) in acidic conditions42. These potentials are higher than that of ferrocenecarboxylic acid and are obviously higher than that of melatonin since this one is capable of reducing the ABTS•– and peroxyl radicals. Therefore, the antioxidant mechanism of melatonin goes through electron transfer but its formal potential does not allow the interaction with a wide free radical extent.

5-3.1.11 Phenol (profile C5)

As melatonin, the phenol antioxidant profile is the C5-labeled one. The monohydroxyl structure failed to scavenge the DPPH• radical, thus this does not allow its classification among the clusters A and B, but the compound showed an intermediate reducing activity toward ABTS•– and peroxyl radicals. This is due to its rather low bond dissociation energy and ionization potential (97.2 kcal/mol and 145.14 kcal/mol, respectively)4. Phenol is a hydrogen atom donor; this ability leads to the reactive-peroxyl-radical detoxification. However, the structure is not suitable for a widespread electron delocalization; therefore the stability of the radical is not high enough to be capable of interacting with the DPPH•, which is much more stable. Unlike this radical, phenol showed an intermediate activity for the ABTS•– reduction. But the acidic conditions induced a drastically activity 179

Chapter 5

decrease (from ER50 = 0.27 in EtOH to ER50 = 0.95 in EtOH/AcOH). In ethanol, phenol seems to be in a form which enables an electron transfer. The stability of the resulting phenol radical cation is higher than that of the ABTS•– radical anion; hence, the ability of the phenol to reduce the radical. This ability is inhibited in acidic conditions due to the neutral form of phenol. This is in line with the results obtained with the ALP and ORAC assays. Besides the high ROO• radical reactivity, pH = 8.3, which the methods were performed at, induced the phenol ionization (pKa-OH = 9). Phenol is therefore capable of readily interacting with the peroxyl radicals through a SPLET mechanism.

5-3.1.12 Glutathione (profile C6)

Glutathione (GSH) is the entity which refers to the compounds solely capable of scavenging ABTS•– radical (C6). Its activity is the greatest of the group

11,12 (ER50 = 0.17) and is in line with reported studies . After a hydrogen atom loss (or proton-concerted electron loss) from the thiol group, an intramolecular rearrangement may take place and then stabilizes the radical by a hydrogen bond bearing by the carbon of the amino acid moiety. A tautomeric form, in which the free electron is born by the carbon atom, is produced77.

O O O O HO HO NH NH NH NH O O O H2N H S H N O H 2 HS HO HO

Figure 5-18: Tautomeric forms of glutathione radical.

This structure may transfer a second electron and leads to the formation of a double bond –C=NH after the hydrogen atom loss. Besides, the abstraction of hydrogen atom from thiol residue yields to a disulfide bridge by interaction of two molecules GS•47. The peroxyl radical detoxifying ability of glutathione has been

180

Chapter 5 proven29 but pH, which the ALP and ORAC assays are carried out at, induces

78 both ionized and non-ionized form of the compound (pKa-SH = 8.6) . The presence of the protonated structure does not enhance the nucleophilicity and then the antioxidant activity. This is confirmed with the ABTS•– assay performed in acidic condition. The free radical scavenging is inhibited (ER50 >> 1). In ethanol, EtOH molecules form hydrogen bonds with SH group, which makes the SH bond weaker. That is in favor of an interaction with ABTS•– molecule through proton- concerted electron loss. Besides, as it has been shown for cysteine, the amine residue may be involved in the ABTS•– reduction and loses this ability in acidic conditions. The weaker ABTS•– scavenging activity of glutathione relative to that of cysteine is attributed to the most hindered structure of GSH.

5-3.1.13 Uric acid (profile C7)

The antioxidant properties of uric acid are of great interest, since it interacts with a wide radical extent. Moreover, besides detoxifying ONOO–, OH•, and O2 among other ROS and RNS, it prevents the lipid peroxidation and the oxidation-induced DNA damages. Due to its intermediate reducing activity

•– toward ABTS (ER50 = 0.25) and to its weak ALP protecting ability against peroxyl radicals (pEC50 = 4.89), uric acid belongs to the profile C7. The high stability of DPPH• as well as the lack of highly active antioxidant residues may explain that uric acid failed to scavenge the free radical. Furthermore, it has been established that the compound acts as antioxidant compound through electron donation79,80. This is in line with its activity toward ABTS•–. In the ALP and ORAC assays, the structure of uric acid is in the monoanion form (pKa = 3.1)79, the negative charge is mainly located on the N3 site36,80 (Figure 5-19).

O H 6 H O N H 1 5 N 7 N HN 8 2 O O H2N 4 9 O 3 N N N H O H

Figure 5-19: Structures of uric acid at pH = 8.3 (left) and of allantoin (right). 181

Chapter 5

Uric acid interacts with ABTS•– and peroxyl radicals by transferring two electrons (from N3 and N7) and a proton (from N7)80. The oxidized uric acid undergoes further reactions which leads to the formation of allantoin81,82 (Figure 5-19). In acidic conditions, the neutral form of uric acid is not capable of reducing

•– ABTS radical (ER50 >> 1). Besides, the triketo form is the most stable structure in gas phase and in acidic conditions80. This is consistent with the electron transfer since the lone pair on nitrogen atoms is involved in a hydrogen bond. Then, the electron transfer from uric acid is inhibited. The activity recorded in

ORAC assay (pEC50 = 4.39) is not very different from that in ALP assay. Thus, an involvement with the ALP protein cannot be proven. The weaker activity than that of polyphenols is attributed to two main reasons. The first one is the less widespread electron delocalization over the structure. The second one is the weak or even the lack of antioxidant properties of the oxidation-produced by-products from uric acid, although a hydroxyl radical scavenging activity has been reported for allantoin83.

5-3.1.14 Aniline and N,N-diethylaniline (profile E4)

NH2 N

Figure 5-20: Structures of aniline (left) and N,N-diethylaniline (right).

Aniline is characterized by the antioxidant profile E4. The properties of this antioxidant profile are a potent and intermediate activity toward peroxyl radicals (in both the ALP and ORAC assays), and a scavenging inability toward DPPH•

•– and ABTS . The activity of aniline in ALP assay is among the best one (pEC50 =

6.04). The lower potency value in ORAC assay (pEC50 = 5.18) suggests that aniline may directly interact with the protein to protect it against the radical attack. Hydrogen bonds, located on the peroxyl radical-attacked site, may be 182

Chapter 5 involved between aniline and the protein. Therefore, the binding affinity of aniline would hinder the oxidation of the protein. As regards the poorer potency in ORAC assay than that in ALP method, it may be due to a weak stability of the aniline radical cation. Nevertheless, the phenyl electron-withdrawing group allows the radical formation and its stabilization. Indeed, a ET mechanism seems the main mechanism involved in the peroxyl radical detoxification since N,N- diethylaniline is also capable of detoxifying them. However, such an activity difference between the ALP and ORAC assays was not recorded for N,N- diethylaniline. This emphasizes the hydrogen bond-induced binding affinity between aniline and the ALP protein. Furthermore, unlike aniline, the two ethyl groups, which induce a positive inductive effect, enhance the electron density on nitrogen atom that is in favor of an electron transfer. This effect contributes to the great potency toward the ROO• radicals. However, the aniline radical cation and the N,N-diethylaniline radical cation are less stable than the ABTS•– radical anion. Indeed they showed no activity in the assay.

5-3.1.15 Mannitol

Mannitol failed in all assays. These results were expected when considering its structure (Figure 5-21). Although it is known as a hydroxyl radical scavenger84,85, the lack of electron rich scaffold makes it inactive toward most radicals.

OH OH

HOH2C CH2OH OH OH

Figure 5-21: Structure of mannitol

This emphasizes the key role of electrons in the antioxidant activity. They are not only necessary to be transferred to the radical but also to be involved in the rearrangement and the stabilization of the newly oxidized byproduct.

183

Chapter 5

5-3.1.16 α-tocopherol

Vitamin E (α-tocopherol) was not taken into account in the cluster analyses since it was not soluble in the ALP and ORAC assays. However, it showed intermediate activity toward the DPPH• and ABTS•– radicals. The results are similar to those of trolox®. This points out that the phytyl residue does not affect the scavenging activity toward these free radicals19.

CH3 HO CH 3 CH3 CH3 CH3

H3C O

CH3

Figure 5-22: Structure of α-tocopherol.

5-3.2 Aurone- and azaaurone-derivatives

Aurones belong to a subclass of flavonoid family and are widely present in plants and fruits. The therapeutic potential of aurone-derivatives covers a broad spectrum since an inhibitory activity has been reported toward rat liver type I iodothyronine deiodinase86, MAO-B87, H1N1 virus neuramidase88, and tyrosinase89. Some derivatives also have a potent activity against Leishmania parasite infections90, prevent the cancer disease91, and are involved in the cancer therapy by reducing the resistance of the cancer cells to the drugs92. Antiproliferative, and apoptotic abilities have been revealed for azaaurone analogs93 as well as a potent MAO-B inhibitory activity87. So far, any antioxidant ability has not been reported for aurone- and azaaurone-derivatives.

184

Chapter 5

5-3.2.1 Aurone-derivatives (4 antioxidants among the 13 compounds)

R4 5' 4'

6' 3' 7 R2 6 8 O 1' 2'

R3 5 2 9 4 3 O R1

R1 R2 R3 R4 Profile CB284 –OMe –OMe –H –Br - CB285 –OMe –OMe –H –Cl - CB286 –OMe –OMe –H –F -

CB290 –OMe –OMe –H –N(Me)2 E4 CB303 –OMe –OMe –H –tBu - CB304 –OMe –OMe –H –Et - SO-I-39 –OH –OH –H –Et C5 SO-II-233 –H –OH –Et –H E3 SO-IV-561 –H –OH –H –OH D5

Figure 5-23: Structures of some aurone-derivatives.

Only four aurone-derivatives showed an antioxidant activity toward either ABTS•– radical or peroxyl radical or both, but none of them turned out to be DPPH• scavengers (annex II-B). Once more, the antioxidant activity is caused either by hydroxyl group born by a benzene ring (SO-II-233, SO-IV-561, SO-I-39) or by a trisubstituted amine residue (CB290). Indeed, CB304, which is the methoxylated form of SO-I-39, did not show any antioxidant activity. The same is true of CB303 which has a ter-butyl moiety instead of the dimethylamine group (CB290). Electron-withdrawing atoms such as fluorine (CB286), chlorine (CB285), and bromine (CB284) inhibit the electron transfer by their negative

185

Chapter 5 inductive effect. Thus, as expected, these three compounds did not reveal any antioxidant activity in the studied conditions. SO-I-39 is a phenol-like compound (profile C5). It reduced ABTS•– radical, forcefully protected the ALP protein against the radicals (ER50 = 0.13 and pEC50 =

5.69, respectively) and recorded a slightly lower activity in ORAC assay (pEC50 = 5.34) than that in the ALP method. The compound bears two hydroxyl groups in position 4 and 6 (the atom numbering is displayed in Figure 5-23). The 4-OH residue is involved in a hydrogen bond with the oxygen of the carbonyl moiety. This feature enables the electron transfer and leads to the stabilization of the radical cation produced from the interaction with ABTS•– and peroxyl radicals. As regards SO-IV-651 and SO-I-39, the only two aurone-derivatives capable of scavenging ABTS•–, it appears that the 4-OH group enhances the reducing activity. The free electron delocalization spreads over the structure of SO-I-39 from the 6-OH group to the oxygen of carbonyl which is stabilized by the hydrogen bond from the 4-OH group. Despite the most widespread free electron delocalization produced after the transfer of an electron from the 4’-OH group, SO-IV-561 is less potent toward ABTS•– radical. It is the only compound with the profile D5. It differs from the other ones of the cluster D in its potent peroxyl radical detoxifying ability and from those of the cluster 5 in its low ABTS•– scavenging activity. The free radical scavenging activity of both compounds, SO-I-39 and SO-IV-561, undergoes a drastic decrease in acidic conditions. This drop suggests that these compounds act through the HAT mechanism or the rearrangement capacity of the oxidized aurone derivative is inhibited by the protons. A hydrogen atom transfer to ABTS•– cannot be excluded even though the behavior of the reference compounds implies an electron transfer. However, the protonation of the oxygen of the carbonyl residue would be in line with an ET. Indeed, the hydrogen bond from the 4-OH group is then hindered (SO-I-39). The antioxidant properties of CB290, which are a peroxyl radical detoxifying activity and a scavenging inability toward DPPH• and ABTS•–, leads this compound to be ranked into the profile E4 with aniline. This classification is in agreement with the structure which bears a tertiary amine with two alkane chains (methyl) and a benzene cycle. Despite the widespread electron

186

Chapter 5 delocalization, CB290 did not show a better activity than N,N-diethylaniline. The tri-substituted amine is the reactive group and its activity toward peroxyl radicals is similar to that of diphenol aurone derivatives (SO-IV-561 and SO-I-39). SO-II-233 solely showed an antioxidant activity in protecting the ALP protein, which is among the best one (pEC50 = 5.61). The monophenol does not seem to detoxify the peroxyl radicals since it failed to protect fluorescein in similar conditions. It appears that the compound directly interacts with the protein. SO-II-233 might bind to the amino acids of the protein by hydrogen bonds since it possesses a hydrogen donor group (OH) and two hydrogen-bond accepting oxygen atoms. This interaction might take place at the peroxyl radical- induced damage site leading to the oxidation of the compound instead of the protein. SO-II-233 belongs to the profile E3, which is not referred to a reference compound.

187

Chapter 5

5-3.2.2 Azaaurone-derivatives (11 antioxidants among the 12 compounds)

R3 5' 4' 6' 3' R4 R2 7 H MeO 6 8 N 1' 2'

R1 5 2 9 4 3 O OMe

R1 R2 R3 R4 Profile ML40 –H –H –Et –H C7 ML44 –H –H –Cl –H D7 ML45 –H –H –H –H D7 ML48 –H –H –SMe –H C7 CB391 –H –H –CN –H D7 CB392 –OMe –H –H –H C7 CB393 –Cl –H –H –H E3 CB395 –Me –H –Me –H C7 CB396 –H –F –H –H D7 CB397 –H –OMe –H –H C7 CB398 –Cl –H –H –F D7 CB400 –Cl –H –H –Cl -

Figure 5-24: Structures of some azaaurone-derivatives.

ML48, ML40, CB392, CB397, and CB395 are uric acid-like compounds, namely they belong to the profile C7. They failed to scavenge DPPH• radical but are able to reduce ABTS•– with an intermediate activity, which is slightly lower than that of the reference compound (annex II-C). The very similar potency value in acidic conditions relative to that in ethanol highlights an ET mechanism that is in line with the reducing ability of uric acid. The substituents on the benzene linked to the double bond do not affect the activity toward ABTS•– radical. However, a 4’-thioether as well as a 3’,5’-dimethyl groups slightly enhance the

188

Chapter 5 protein protective activity that leads the activity of ML48 and CB395 into the class 2 (pEC50 = 5.22 and 5.31, respectively). The activity difference between the ALP and ORAC assay is not large enough to suggest an interaction of the compounds with the protein. A carbonitrile (CB391) in position 4’ induces an activity drop toward

•– ABTS (ER50 = 0.70) and in the protein protective activity against peroxyl radical-induced damages (pEC50 = 4.23) relative to those of the azaaurone derivatives of the profile C6. This is caused by the negative inductive effect of the substituent, which reduces the electron transfer capacity. Thanks to its weak activity in the ABTS•– and ORAC assays, CB391 is ranked in the profile qualified by the profile D7 with, among others, ML45, CB396, ML44, and CB398. These compounds showed a slightly better activity in ALP assay than that of CB391. As the CN residue did, halogen atoms induced an ABTS•– activity decrease relative to that of the compounds which bear methoxy, thioether, and alkane chains. However, halogen atom substitution on the benzene does not affect the antioxidant activity of azaaurone core compared to that of CB391 (with a CN group). Indeed, the azaaurone derivative ML45 has similar antioxidant properties as ML44 (4’-Cl), CB396 (3’-F), and CB398 (2’-Cl,6’-F). They are not susceptible to readily transfer an electron. The remaining active compound is CB393. It belongs to the antioxidant profile E3 since it failed to scavenge ABTS•– radical, unlike the compounds of D7. By considering CB393, ML44, CB398, and CB400, the structural difference implies that the position and the number of the chlorine has an effect on the antioxidant ability toward ABTS•– and peroxyl radical even though the compounds of D7 are poor ABTS•– scavengers. When placed in 4’ (ML44) or in 2’ with the influence of a fluorine atom in 6’ (CB398), the compound may reduce the free radical, but it fails when a chlorine atom is the only substituent in 2’. Furthermore, an azaaurone derivative with two chlorine atoms in position 2’ and 6’ (CB400) does not possess any antioxidant activity in the studied conditions. From an overall point of view, the azaaurone derivatives are more susceptible to be antioxidant compounds than aurone derivatives. For instance, the azaaurone derivative ML44 showed weak antioxidant activity toward peroxyl

189

Chapter 5 and ABTS•– radical when the aurone CB285 does not. The same is true of ML44 compared with CB304. Therefore, the nitrogen atom is required to improve the electron transfer from the naked core. But the substituents have a key role. Hydroxyl group in the 4-position induces the greatest antioxidant activity improvement since SO-I-39 is the best ABTS•– scavenger and among the best peroxyl radical detoxifying agent even though it is an aurone-derivative. The benzene cycle substituents, which enhance the electron transfer, are hydroxyl, tertiary amine, thioether, methoxy groups, and alkyl chains. Nitrile and halogen atoms do not have any influence on the antioxidant capacities of the azaaurone core, except when the chlorine is in the position 2’.

190

Chapter 5

5-3.3 N-derivatives of (3S,4S)- and (3R,4R)-pyrrolidine-3,4- diols (15 antioxidants among the 27 compounds)

HO 4 5

HO 3 N 2 1 R

Stereochemistry R Profile

VRL-31 3S,4S –(CH2)2-NH2 D6

VRL-40 3S,4S –(CH2)3-NH2 D6

VRL-43 3S,4S –(CH2)2-NH-Et C6

VRL-45 3S,4S –(CH2)2-N(Me)2 C6

VRL-58 3R,4R –(CH2)2-NH2 D6

VRL-69 3R,4R –(CH2)3-NH2 D6

VRL-71 3R,4R –(CH2)3-NH-Et C6

VRL-75 3S,4S –(CH2)2-NH-CH2-Ph C6

VRL-76 3S,4S –(CH2)2-NH-CH2-Ph-p-Ph C6

VRL-77 3S,4S –(CH2)2-NH-CH2-Ph-p-Cl C6

VRL-78 3S,4S –(CH2)3-NH-CH2-Ph C6

VRL-79 3R,4R –(CH2)3-NH-CH2-Ph C6

VRL-80 3R,4R –(CH2)2-NH-CH2-Ph C6

VRL-81 3R,4R –(CH2)2-NH-CH2-Ph-p-Ph C6

VRL-82 3R,4R –(CH2)2-NH-CH2-Ph-p-Cl C6

Figure 5-25: Structures of ABTS•– scavenging N-derivatives of (3S,4S)- and (3R,4)- pyrrolidine-3,4-diols.

Some N-derivatives of pyrrolidine-3,4-diols showed an ABTS•– reducing ability. A primary, secondary or tertiary amine on the side chain confers the capacity for interacting with the radical with an ER50 over a range from 0.21 to 0.65 (annex II-D). As regards the secondary and tertiary amine, the second group bounded to the nitrogen on the side chain does not affect the ABTS•– trapping ability. Indeed, VRL-43 and VRL-71 (with a secondary ethylamine) as well as

191

Chapter 5

VRL-45 (with a tertiary dimethylamine) have a similar ER50 parameter as VRL- 75, VRL-76, VRL-77, VRL-78, VRL-79, VRL-80, VRL-81, and VRL-82. These compounds bear a N-benzyl group either non-substituted or substituted by a phenyl or a halogen atom (Figure 5-25). They have a glutathione-like antioxidant behavior that means that they are capable of scavenging the ABTS•– radical with a potent or intermediate activity (0.21 < ER50 < 0.31). Therefore, they are defined by the profile C6. However, the primary amines (VRL-31, VRL-40, VRL-58, and VRL-69) are less potent scavengers than the secondary and tertiary amines

(0.5 < ER50 < 0.65). They are then described by the profile D6. The number of carbon atoms (2 or 3) which form the side chain between the two nitrogen atoms does not have any influence on the ABTS•– scavenging ability. Like the reference compounds, the reducing ability drastically decreases in acidic conditions. These results imply that they interact with the free radical through electron transfer from the amine residue. Indeed, in a protic media, the nitrogen atom would be protonated, hence the inability of the lone pair to transfer an electron. The involvement of the primary, secondary, and tertiary amines in the antioxidant activity toward ABTS•– radical is confirmed as it was already reported14. The relatively small structure may also have a key role; it enables the interaction with the free radical. Adducts might be among the by-products. The naked pyrrolidine-3,4-diol or bearing a N-linked alkane or benzyl residues does not show any antioxidant properties.

192

Chapter 5

5-3.4 Racemic conduramine B-1 analogs (4 antioxidants among the 22 compounds)

HO OH

HO NH R

R Profile

ROBI-26 –CH2-Ph-p-OH C5

ROBI-27 –CH2-Ph-o-OH C5

ROBI-33 –CH2-Ph-o-Br-m-OH-p-OMe E4

ROBI-163 –CH2-indole D7

Figure 5-26: Structures of racemic conduramine B-1 analogs.

N-benzyl conduramine B-1 derivatives has an interest in the inhibition of β- glucosidase94,95, β-xylolase, β-galactosidase, and α-mannosidase96 Five compounds of this family showed an antioxidant activity, even weak, toward either ABTS•– radical or peroxyl radicals or both (annex II-E). The most potent are ROBI-26 and ROBI-27 ranked in the profile C5. Their phenol-like antioxidant properties are consistent with their structure since both compounds bear a phenol moiety. They differ from each other in the hydroxyl position on the benzyl group. The ortho position (ROBI-27) possesses more potent reducing

•– activity in the ABTS (ER50 = 0.15), ORAC (pEC50 = 5.43), and ALP (pEC50 =

5.51) methods than that of the para position (ROBI-26) (ER50 = 0.44, pEC50 =

5.16, pEC50 = 5.32, respectively). The conduramine B-1 moiety does not affect the antioxidant capacities when the OH is located in ortho position since ROBI-27 and phenol have similar peroxyl-radical-reducing ability. It even enhances the

•– activity toward ABTS radical; ER50 parameter of ROBI-27 belongs to the class 3 when that of phenol is ranked in the class 2 range. However, when the two substituents are in para position (ROBI-26), the antioxidant abilities of the compound tend to decrease with respect to that of phenol. A protic media inhibits

193

Chapter 5 the ABTS•– scavenging activity for both compounds that suggests that a SPLET mechanism occurs to trap the free radical. ROBI-26 is a potent inhibitor of β- glucosidase from almonds95 and from Saccharomyces cerevisiae as well as a potent inhibitor of α-mannosidase from jack bean and from almonds96. The third conduramine B-1 analog which bears a phenol moiety is ROBI-33. It belongs to the profile E4 referring to aniline. The benzyl moiety is also substituted with a methoxy group and a bromine atom. Because of their negative inductive effect, these two electron-withdrawing groups reduce the electron transfer ability from the hydroxyl group; hence the ability to scavenge the ABTS•– radical. However, the peroxyl radical detoxifying activity is similar to that of ROBI-26 (pEC50 = 5.36 and pEC50 = 5.21 in the ALP and ORAC methods, respectively). Only the phenol residue interacts with the reactive radicals. The indole nucleus does not induce an antioxidant activity as potent as the phenol moiety, but it contributes to a weak reducing activity. Indeed, the

•– ER50 value of ROBI-163 equals 0.90 in ABTS assay and the pEC50 in ALP and ORAC assays is 5.07 and 4.74, respectively. It is among the compounds characterized by the profile D7. The remaining active conduramine B-1 derivative, ROBI-31, solely has a weak antioxidant activity toward ABTS•– radical. This is attributed to the tertiary amine which might transfer an electron and then be stabilized thanks to the positive inductive effect from the alkane chains. It is included in the profile D6, with the compounds which owe their ABTS•– scavenging capacity to the amine residue. The naked conduramine core (ROBI-16, ROBI-18) is not an antioxidant property providing structure. Furthermore, none of N-linked benzyl rings either naked (ROBI-22) or with for instance methoxy (ROBI-23, ROBI-34) or halogen atom (ROBI-25, ROBI-32) as substituents induces an antioxidant activity.

194

Chapter 5

5-3.5 Secondary amines (None antioxidant among the 3 compounds)

The three compounds of this family did not show any antioxidant properties. The steric hindrance of their structure relative to the small structure of the N- derivatives of pyrrolidine-3,4-diols, which are ABTS•– scavengers, likely blocks the interaction with the radical (annex II-F).

5-3.6 Esters of (3S, 4S)-pyrrolidine-3,4-diols (None antioxidant among the 2 compounds)

Despite an electron rich scaffold, any residue is not capable of initiating an electron or a hydrogen transfer (annex II-G).

5-3.7 Racemic conduramine F-1 analogs (3 antioxidants among the 9 compounds)

HO OH

HO NH R

R Profile

ROBI-329 –CH2-(N-Ac)indole E3

ROBI-333 –CH2-Ph-p-OH E4

ROBI-335 -CH2-furan-Ph-m-CF3-o-Cl E3

Figure 5-27: Structures of racemic conduramine F-1 analogs.

ROBI-333 is a configurational stereoisomer of ROBI-26. Their activity toward peroxyl radicals are similar, as well as their impotency toward DPPH•

195

Chapter 5 radical. However, the conduramine F-1 analogs did not turn to be ABTS•– scavengers while ROBI-26 did. Although the trapping ability has been shown to be due to the phenol moiety, the three-dimensional structure induced by the S carbon atom configuration inhibits the electron transfer. Since such an activity loss is not observed in the ALP and ORAC assays, the steric hindrance of ROBI- 333 may be the main factor of the ABTS•– scavenging inability (annex II-H).

The results of ROBI-335 in the ALP (pEC50 = 5.11) and ORAC (pEC50 < 4.5) assays lead to the fact that the compound directly interacts with the protein on peroxyl radical-attacked sites. The structure differs from the other conduramine F-1 analogs in the presence of a furan moiety substituted by a para- chlorobenzotrifluoride ring. This scaffold is not susceptible to make hydrogen bonds; therefore the interaction might be the result from a three-dimensional structure effect. ROBI-329 has a weak ALP protecting activity against the peroxyl radical- induced oxidation. The N-acetylindole is an electron rich scaffold which could lead to the detoxification of peroxyl radicals. The activity loss of this compound relative to ROBI-163 is caused either by the molecular configuration or by the N- acetyl substitution or both. The electron-withdrawing group decreases the electron and hydrogen loss ability from the indole moiety. This substitution appears to be the main reason for which ROBI-329 is drastically less potent than ROBI-163.

196

Chapter 5

5-3.8 Racemic conduramine F-1 epoxides (3 antioxidants among 11 compounds)

HO OH HO OH

HO NH HO NH R R aO bO

Stereochemistry R Profile

ROBI-470 a -CH2-Ph-m,p-Cl-o-OH E4

ROBI-475 a -CH2-Ph-p-OH D7

ROBI-486 b -CH2-Ph-p-OH C7

Figure 5-28: Structures of racemic conduramine F-1 epoxides.

The results obtained for ROBI-470, ROBI-475, and ROBI-486 provide the low double bond effect in the conduramine analogs. Indeed, these compounds solely showed a slightly weaker activity than that of ROBI-333 toward ROO• radicals (annex II-I). The significant drop of the activity of ROBI-475 is attributed to a steric hindrance induced by the epoxide group which is on the same side as the secondary amine and the two hydroxyl groups. The benzene substitution seems not to affect the antioxidant activity since ROBI-470 did not show a relevant activity difference. However, the ABTS•– scavenging ability of these compounds put them inside three profiles. Indeed, ROBI-486 is the most potent (ER50 = 0.38), that leads it to the uric acid profile C7. The epoxide, on the opposite side of the amine, allowed the interaction with the radical, while the double bond did not (ROBI- 333). Therefore, the activity of ROBI-486 assessed by ABTS•– method is similar to that of the conduramine B-1 analog ROBI-26. The presence of the epoxide on the same side as the amine residue and the two hydroxyl groups decreases the capacity of ROBI-475 for scavenging ABTS•– radical. Its weak activity makes it belong to the cluster D7.

197

Chapter 5

Besides the steric feature of ROBI-475, ROBI-470 has two chlorine atoms in ortho and meta of the hydroxyl group. These electron-withdrawing residues inhibit the ABTS•– scavenging ability and put ROBI-470 among the aniline-like compounds (E4).

5-3.9 Compounds extracted from Jacaranda caucana (7 antioxidants among the 7 compounds)

HO OR3

O R1O O O O O

HO OH O

HO HO

OH OR2

R1 R2 R3 Profile JCA33 –H –H –H A2 JCA42 –H –Me –H B1

JCA44 –H –H -CO-CH2-cyclohexane-1,4-diol A2

JCA47 –H –H -CO-CH2-(4-OH)cyclohexanone B1 JCA54 –Me –Me –H B1

Figure 5-29 : Structures of phenylethanoid glycosides.

198

Chapter 5

O HO OH O

O HO O O HO O OH

OH HO

Figure 5-30 : Structure of JCA37 (profile A2).

Seven compounds, extracted and isolated from plants of the family Bignoniacae from Panama97, have been assessed and revealed antioxidant properties ranked among the best ones (annex II-K). JCA33, JCA37, JCA44, and JCA47 have been ranked among the caffeic acid-like compounds characterized by the profile A2. This classification is due to their great potency in the four assays as well as their rapid DPPH• scavenging activity. Despite their more hindered steric structure relative to that of caffeic acid, their antioxidant properties are greater than those of the reference compound. Besides a cinnamate moiety, they bear a catechol group which provides a second highly reactive antioxidant moiety. This group enhances the antioxidant activity of the ester of caffeic acid relative to this reference compound. The combination of the two active groups makes the DPPH• scavenging ability as great as that of gallic acid (ER50 = 0.04). However, this does not imply that the catechol is the only residue capable of scavenging the radical, but points out that it compensates for the unfavorable steric hindrance effect, which reduces the interaction with DPPH•. The extracted compound behavior in acidic conditions describes a HAT interaction with free radical. Indeed, despite a slightly higher ER50 for JCA44

(ER50 = 0.16 in EtOH/AcOH when ER50 = 0.10 in EtOH) and JCA47 (ER50 = 0.17 in EtOH/AcOH when ER50 = 0.08 in EtOH), the ER50 values in acidic conditions

199

Chapter 5 are closed to those in ethanol. However, the significant decrease in the percentage of DPPH• scavenged in a minute emphasizes a SPLET mechanism which takes place before the hydrogen atom transfer occurs. The radical scavenging interaction is consistent with that of caffeic acid. The substituent on the sugar moiety, which makes them different, does not affect the activity of the compounds. The three remaining compounds extracted from Jacaranda Caucana, namely JCA42, JCA54, and JCA7, have been ranked in the profile B1 related to the antioxidant capacities of chlorogenic acid, gallic acid, and resveratrol. Their antioxidant abilities differ from those of the extracts of the profile A2 in their structure. JCA42 is a caffeic acid ester with a methoxy-substituted catechol (Figure 5-29), thus it is a metoxylated analog of A33. The main antioxidant activity difference between these two compounds lies in the DPPH• scavenging rate (40% of DPPH• scavenged in a minute by JCA42 toward 91% by JCA33) and in the ORAC assay (pEC50 = 5.03 for JCA42 whereas pEC50 = 5.75 for

JCA33). The higher ER50 value in acidic conditions (ER50 = 0.55) implies that JCA42 mainly scavenge DPPH• through a SPLET mechanism, unlike JCA33 which mainly acts through a HAT. Therefore, the methoxylated catechol induces

• a DPPH scavenging activity decrease (ER50 = 0.18 for JCA42 when ER50 = 0.08 for JCA33) as well as a change in the main free radical interacting mechanism. The hydrogen atom transfer is minimized in favor of a SPLET mechanism. By considering the chlorogenic acid antioxidant behavior, which is another ester of caffeic acid, it appears that JCA42 also differs from it in the DPPH• scavenging mechanism. A hydrogen-bond-induced folded structure might occur in this compound that would reduce the hydrogen atom transfer. As regards the activity toward peroxyl radicals, it appears that JCA42 is involved in an interaction with the protein. Indeed, the activity recorded in the ORAC method (pEC50 = 5.03) is consistent with the structure since JCA42, bearing a substituted catechol moiety, should be less potent than JCA33, with a catechol moiety. Furthermore, it has been assumed that the antioxidant capacities could be reduced due to a folded scaffold of JCA42. However, the potency value obtained from the ALP assay was much higher (pEC50 = 5.96) than that obtained from the ORAC assay. The

200

Chapter 5 protein-protective activity would be enhanced by a direct interaction of JCA42 with the protein. The structure bears several hydrogen acceptor sites as well as several hydrogen donor sites able to make hydrogen bonds with amino acids from the protein. The protein-binding scaffold might hinder the attack from peroxyl radicals on the protein. JCA54 is a methoxylated analog of JCA42 (Figure 5-29). It is substituted by a methoxy on the catechol group of the cinnamate moiety. This structure change induces a weaker activity in the DPPH• and ALP assays and a slightly better activity in the ORAC assay than that of JCA42. The reducing capacity toward ABTS•– is similar. The decrease in DPPH• assay resides in the lack of catechol moiety compared to JCA42. The free radical is scavenged by a SPLET mechanism as pointed out by the strong drop in acidic conditions. The methoxy residue in JCA54 instead of a OH group removes a protein-binding site. As a result, the activity in ALP assay is slightly poorer than that of JCA42. It must be mentioned that JCA54 had a better activity toward ABTS•– in acidic condition than it did in ethanol (ER50 =0.11 and ER50 = 0.20, respectively). Proton atoms from the media might be involved in the three-dimensional conformation, yielding a structure more susceptible to transfer an electron. The protocatechuic acid (JCA7) is a well-known antioxidant compound, which has been widely under investigation in structure-activity relationship studies9,52,98-101. It is lacking a hydroxyl group relative to the structure of gallic acid. This difference makes the 3,4-dihydroxylbenzoic acid less potent toward

• •– DPPH (ER50 = 0.23) and ABTS (ER50 = 0.14) than gallic acid is. The third OH group on gallic acid contributes to a better stabilization of the produced radical. Due to the lack of this hydroxyl group and therefore to the weaker stabilization of the radical, JCA7 is less susceptible to lose a proton and an electron. The DPPH• trapping mechanism also differs since JCA7 mainly acts through a SPLET, unlike gallic acid. Indeed, the ER50 value decreases from 0.23 in ethanol to 0.50 in the EtOH/AcOH mixture. However, protocatechuic acid showed a slightly better peroxyl radical detoxifying activity than that of gallic acid but not strong enough to be relevant. The unchanging activity toward ABTS•– radical in both media is consistent with a radical reduction through an electron transfer.

201

Chapter 5

5-3.10 1-substituted and 2-substituted pyrazolopyrrolizine derivatives (7 antioxidants among the 21 compounds)

8 8 4 X 7 4 X 5 11 5 7 11 1 1 N N 10 N N 3 6 10 R N N 3 6 2 9 2 9

Profile Substitution X R E3 MC2372 N2 (±) CH-OH p-Cl-Ph E3 MC2385 N2 CH2 p-Cl-Ph E3 MC2386 N1 (±) CH-OH p-Cl-Ph (±) CH-O-CO- E3 N2 p-Cl-Ph MC2388 CH3 E4 MC2405 N2 cis-C=N-OH p-Cl-Ph E3 MC2406 N2 trans-C=N-OH p-Cl-Ph trans-C=CH-CO- E3 N2 p-Cl-Ph MC2416 OEt Figure 5-31: Structures of 1-substituted (left) and 2-substituted (right) pyrazolopyrrolizine-derivatives.

The active compounds are ranked in the profile E3, among the compounds solely capable of protecting the protein ALP. The lower activity in ORAC assay for mostly active compounds may be attributed to the weaker sensitivity of this assay with respect to ALP method. It has been noticed that the pEC50 values are lower in the method using fluorescein. However, some compounds from this set showed a relevant difference of activity between the ALP and the ORAC assays (annex II-L, II-M). A protein-binding interaction obviously takes place with these scaffolds. The active compounds differ from the inactive compounds from the same set in the three-ring structure as well as in the substituent and the position of the benzene ring (Figure 5-31). The comparison of activity of MC2385 with MC2387 emphasizes the benzene ring position. Indeed, the N1-linked 4- chlorobenzene derivative (MC2387) exhibited no activity unlike MC2385 (N2- 202

Chapter 5 subtituted). The activity difference between MC2372 and MC2386 is not as obvious but tends to the same position effect since MC2372 has a somewhat more potent activity in the ALP and ORAC methods (pEC50 = 5.08, pEC50 = 4.67, respectively) than that of MC2386 (pEC50 = 4.83, pEC50 = 4.46, respectively). Thus, the configuration plays a role providing an interaction with peroxyl radicals. The electron rich three-ring scaffold is obviously involved in the protection against peroxyl radical-induced oxidative damage but its potency depends on the substituent at the position 8. Only an oxime group enhances the ALP protective ability but does not induce such an effect in ORAC assay.

MC2406 and MC2405 activity assessed in ALP method (pEC50 = 5.73 and 5.63, respectively) tends to provide an interaction with the protein which yields a protective effect. The OH moiety has a key role since the methoxylated structure (MC2410 and MC2411) did not show any activity. The ability to make a hydrogen bond from the OH is obviously the main protein-binding site but not the only one. Indeed, the nitrogen atom is also involved probably thanks to its lone pair, which is a hydrogen acceptor moiety, because MC2372, bearing only an OH group is much less potent. An acetate group (MC2388) slightly decreases the antioxidant capacity compared with that of the naked derivative while a double- bonded ethyl acetate moiety (MC2416) induces a stronger drop. Although the latter enhances the electron delocalization, the former is more potent. A carbonyl (MC2117) and a thioketone (MC2415) inhibit the peroxyl radical detoxifying capacity. The compounds of this series seem to act as ALP protective agents through a protein-binding effect. Indeed, they have acceptor and donor hydrogen bond residues on their structure. Although each pyrazolopyrrolizine-derivative bears a chlorine atom, the involvement of the halogen on the benzene ring is not significant. The same is true of the substituent as well as the substitution position on the benzene ring. The inactivity of the compounds MC2363, MC2412, MC2417, MC2418, MC2123, MC2147, MC2170, and MC2171 is obviously due to the 8-carbonyl group. Pyrrole and the 2-methylquinoline-3-carboxylate derivatives have no antioxidant active group.

203

Chapter 5

5-3.11 Xanthone-derivatives (15 antioxidants among the 24 compounds)

Xanthone-derivatives represent a large amount of the polyphenol present in

102-106 • •–107 plants and are capable of detoxifying radicals such as OH and O2 . Some of them have been reported as inhibitors of the growth of human cancer lines108. Mangiferin (2-β-D-glucopyranosyl-1,3,6,7-tetrahydroxy-9H-xanthen-9- one), bellidifolin, desmethylbellidifolin, and norswetianin are among the most known xanthones. Many studies reported the pharmacological effects of mangiferin such as the inhibition of lipid peroxidation in rat liver microsomes induced by both an enzymatic system and by a Fenton system109, the free radical scavenging activity109, the iron chelating ability110. Among other therapeutical properties, desmethylbellidifolin is an inhibitor of Cu2+-induced LDL oxidation and is capable of decreasing the nitric oxide level111. Bellidifolin, desmethybellidifolin, and norswetianin are MAO-A, MAO-B112,113, and AChE inhibitors113. Norswertianin also showed the ability to protect the butyrylcholinesterase against peroxyl radical-induced damages114. It has been reported that not only the presence of hydroxyl groups on the xanthone scaffold contributes to antioxidant properties, but also the position where the OH group is located. The xanthones OH-substituted at the C-1 and at the C-3 are not susceptible to act as an antioxidant compound56,115 (Figure 5-32).

O 8 1 7 2

6 O 3 5 4

Figure 5-32: Atom numbering of xanthone.

The 15-active-xanthone-derivative set it split between seven antioxidant profiles, namely A1, A2, B1, B2, C5, E3, and E4. Two compounds, X23 and X14, were not taken into account for the cluster analysis as the product amount of X23

204

Chapter 5 was not enough, and X14 was fluorescent in the ALP assay. 7 xanthone derivatives showed no antioxidant activity in any methods (annex II-O).

R10 OH O O HO HO R9 HO O OH

Profile R9 R10 B2 APL58 p-OH-benzoate –OH B2 APL70 p-OH-cinnamate –OH B1 APL98 cinnamate –OH B2 APL101 –OH p-OH-cinnamate A2 APLL3 m,p-OH-cinnamate –OH B1 7418 benzoate –OH

Figure 5-33: Structures of mangiferin-derivatives.

Mangiferin-derivatives (annex II-P) and xanthone scaffolds bearing a hydroquinone moiety belong to the most potent xanthone derivatives. APLL3 is the only compound from this set which is characterized by the antioxidant profile A2. It is capable of both interacting toward the three radicals in the four assays with the greatest potency and interacting with DPPH• and ABTS•– quickly. The APLL3 scaffold is both a mangiferin derivative and an ester of caffeic acid. Hence, its rapid and potent antioxidant activity is due to the catechol moiety in position 6,7 on the mangiferin residue and to the caffeic acid structure, whose the great antioxidant potency has been described above. Besides, APLL3 resides into the profile A2, referring to caffeic acid. A mangiferin- and quercetin-like antioxidant profile describes the antioxidant properties of APL58, APL70, and APL101. They are as potent compounds as APLL3 but scavenge DPPH• more slowly than APLL3 does. This lower trapping rate is likely due to the lack of a catechol moiety on the second

205

Chapter 5 group linked to the sugar moiety. These three compounds owe their potency to the catechol of the mangiferin core and to the phenol group from the 4- hydroxycinnamate (APL70 and APL101), and from the 4-hydroxybenzoate (APL58) residues. The double bond of the 4-hydroxycinnamate moiety does not seem to play a crucial role in scavenging radicals since the pEC50 obtained from the ALP and ORAC assays, and the ER50 as well as the percentage of free radical scavenged in a minute in the DPPH• and ABTS•– assays are similar for APL58, APL70, and APL101. The same is true of the position of the 4-hydroxycinnamate group on the sugar moiety when the compounds APL70 and APL101 are considered. Although their DPPH• scavenging rate is similar to that of mangiferin and of other compounds from B2, which is slow and very slow respectively, the remaining mangiferin derivatives, namely APL98 and 7418, belong to the profile B1. Indeed their potency is not great toward all radicals. APL98 showed an intermediate capacity for trapping the DPPH• radical and the antioxidant potency of 7418 was high only in the ALP assay and intermediate in the other ones. These lower activities are likely due to the lack of hydroxyl group on the cinnamate and benzoate moiety, hence only the catechol group is involved in the antioxidant power. That is also the case for mangiferin but the cinnamate and the benzoate residues may induce a steric hindrance, which decreases the interaction between the compounds and the radicals. The double bond of the cinnamate moiety seems to be involved in the antioxidant ability since the pEC50 of APL98 in the ORAC assay is greater than that of 7418 (5.59 and 5.14,

• respectively), as well as the ER50 parameter in both the DPPH (0.23 and 0.35, respectively) and ABTS•– (0.18 and 0.29, respectively) assays. Except the compound 7418, the DPPH• and ABTS•– scavenging capacity has been assessed in acidic media (EtOH/AcOH (1%, 10mM)). The ER50 in both assays as well as the percentage of ABTS•– radical scavenge in a minute at the screening ratio are similar for all compounds. However, the DPPH• scavenging rate was slower for each mangiferin-derivative. The compounds of the kinetic profile B were capable of scavenging less than 10% of DPPH• in acidic conditions when they scavenged around 30% in ethanolic solution. The most important decrease concerns APLL3

206

Chapter 5 which scavenged only 10% of DPPH• in protic media when it scavenged 89% in ethanol. The potency results emphasize that a HAT is susceptible to take place alone; the DPPH• scavenging is therefore due to the abstraction of a hydrogen atom from the mangiferin derivatives over 90 minutes. However, a SPLET mechanism may occur at first in ethanolic solution inducing a faster DPPH• trapping.

R8 O R1

R7 R2

R6 O R3

R5 R4

Profile R1 R2 R3 R4 R5 R6 R7 R8 - X25 –OH –H –H –H –H –H –H –H C5 X13 –H –OH –H –H –H –H –H –H C5 X24 –H –H –H –OH –H –H –H –H E4 X8 –H –H –OH –OMe –H –H –H –H C5 X9 –H –H –OMe –OH –H –H –H –H C5 ID22 –OH –H –OMe –H –OH –H –H –H E3 ID23 –OH –H –OMe –H –OMe –H –H –H A1 ID39 –OH –H –OH –H –OH –H –H –OH A1 ID40 –OH –H –OMe –H –OH –H –H –OH B1 ID42 –OH –H –OH –H –H –H –OH –OH

Figure 5-34: Structures of xanthone-derivatives.

Desmethylbellidifolin (ID39) and bellidifolin (ID40) belong to the profile A1 due to their potent and intermediate activity toward the three radicals and their rapidity to scavenge the DPPH• radical. The close parameters values in each assay for ID39 and ID40 emphasize that the metoxy moiety at C-3 on the bellidifolin structure does not affect its antioxidant activity. Both compounds owe

207

Chapter 5 their good potency to the para-hydroquinone moiety, known to readily undergoes oxidation116-118. The free radical scavenging capacity of these two compounds is slower and less potent in acidic media. The ER50 value is almost 2-fold higher in the acidic DPPH• assay and the percentage of radical scavenged in a minute is less than 10% for each compound. The same but less marked tendency appears in the ABTS•– assay. Hence a SPLET takes place concerted with the hydrogen transfer in the ethanol solution. ID42 is slightly more potent against the peroxyl radicals than ID39 and ID40 are, its pEC50 value in ALP assay is 5.58 when that of ID39 and ID40 is 5.37 and 5.29, respectively. Moreover, the pEC50 values are 5.42, 5.20, and 5.24 for ID42, ID39, and ID40, respectively, in the ORAC method. However, the ortho-hydroquinone compound is slower than both para-hydroquinone compounds in the DPPH• scavenging activity. It scavenged 45% of DPPH• when ID39 and ID40 scavenged 79% and 78%, respectively, of radical in 1 minute. That slower rate makes ID42 be described by the antioxidant profile B1. The less potent activity of norswertianin (ID42) in DPPH• assay with acidic conditions implies that it detoxifies the free radical through a SPLET, as ID39 and ID40 do. A better antioxidant activity of mangiferin- derivatives over the compounds of the cluster A1 and ID42 was expected. Besides the catechol moiety known to react more easily than the para-hydroquinone5,119,120, the catechol at C-7 and C-8 (ID42) is less available for an hydrogen transfer since the hydroxyl group in position 8 is involved a hydrogen bond with the keto group in position 9. As a result, ID42 is less efficient than the mangiferin-derivatives but 7418. It is also slightly less potent than APL98 in the DPPH• and ABTS•– assays. These close or lower activity of ID42 with respect to 7418 and APL98, despite the catechol in position 6,7, may be due to the steric hindrance of the benzoate and the cinnamate moieties, which decreases the interaction with both radicals. The same is true of ID39 and ID40 toward the DPPH• and ABTS•– radicals; para-hydroquinone structure induces the same activity as a scaffold which bears a catechol moiety and a hindered residue. The four xanthone derivatives (X13, X24, X9, and ID22), characterized by the antioxidant profile C5, bear at least a hydroxyl group. They showed a potent

208

Chapter 5 or an intermediate activity toward peroxyl radicals (in both the ALP and ORAC assays) and ABTS•– radicals. However, they are unable to scavenge the DPPH• radical. The stability of the neutral radical species produced after a hydrogen atom abstraction has been reported for some xanthone-derivatives. The ΔBDE parameter, which describe the ability to donate a hydrogen atom, is equal to 82.81, 82.16, 78.89 kcal/mol for X13, X24, and X9, respectively, when that of mangiferin is 80.41 kcal/mol56. Furthermore, the ability of these compounds to donate an electron, described by the ΔIP parameter, equals 172.05, 174.79, and 171.3 kcal/mol, respectively, when that of mangiferin is 172.32 kcal/mol56. Besides quantum chemical descriptors, the electrochemical behavior of the four xanthone-derivatives as well as that of mangiferin was studied by cyclic voltammetry at pH = 7.4 (phosphate buffer (0.07 M, pH = 7.4) / ethanol (1/1)).

Their oxidation potential Epa equals 0.63 V for X13, 0.63 V for X24, 0.44 for X9, 0.45 for ID22, and 032 V (vs Ag/AgCl) for mangiferin at a scan rate of 100 mV.s-1 115. According to these results, a HAT seems to be the main antioxidant mechanism for the four derivatives. But these compounds did not turn out to be DPPH• scavengers despite the hydrogen abstraction energy close to that of 1,4-dihydroxyxanthone (78.35 kcal/mol) and 1,2-dihydroxyxanthone (80.62 kcal/mol) which are a part of the structure for instance of ID39 and ID42, respectively. The capacity of donating a hydrogen atom from X13, X24, X9, and ID22 might not be strong enough to interact with the highly stable DPPH• radical. That emphasizes the importance of an ortho- or para-dihydroxyl moiety on the structure for the DPPH• trapping. The 2-OH group appears to contribute to a greater ABTS•–scavenging activity than the 4-OH group does. Indeed, X13, which bears a hydroxyl group in position C-2, has an ER50 value equal to 0.13 when that of X24 is 0.34 and the one of ID22 is 0.27. The inductive effect of the

•– metoxy group enhances the activity toward the ABTS radical; the ER50 value of X9 is 0.20. The OH group at C-1 is not involved in the antioxidant ability since ID22 does not present a better activity despite the presence of two hydroxyl groups. These results are in agreement with studies reported before56,115. As regards the ABTS•– method in acidic media, the significant activity decrease of the four xanthone-derivatives points out a SPLET involvement in the radical

209

Chapter 5 detoxification. The protic media does not allow the interaction between the solvent (ethanol) and the hydroxyl groups, which would lead to a weaker OH bond. Therefore, an electron transfer does not occur to reduce the radical. Finally, the position of hydroxyl groups does not drastically affect the reductive activity of peroxyl radicals. The pEC50 value of the four compounds of the antioxidant profile C5 is of the same order of magnitude as ID39, ID40, and ID42 in the ALP and ORAC assays. Only mangiferin in ORAC assay showed a slightly better activity. The two remaining active xanthone derivatives are not DPPH• or ABTS•– scavengers but differ from each other in their activity in ORAC assay. Indeed, X8

(E5) and ID23 (E4) have a similar ALP protecting capacity (pEC50 = 5.42 and 5.41, respectively) but, X8 is capable of protecting fluorescein against peroxyl- radical attacks (pEC50 value equals 5.19) when ID23 cannot (pEC50 equals 4.84). The ΔBDE (84.46 kcal/mol) and ΔIP (173.01 kcal/mol)56 of X8 are slightly higher than that of other xanthone derivatives but allow an interaction with the reactive peroxyl radicals. The 4-methoxy group enhances the reactivity of the 3-OH group which is not located in an active position. As regards ID23, the lack of activity was expected since it bears only one OH group in an inactive position. But its ALP protective ability is interesting. It is similar to the other xanthone derivatives (except that of mangiferin and its derivatives). The compound might make a link with the protein at the peroxyl-radical-attacked site, which could hinder the oxidation-induced damages. Except ID23, the peroxyl radical-detoxifying xanthone-derivatives are not involved in any interaction with ALP protein. Indeed, their pEC50 values are almost the same in the ALP and ORAC assays. The ineffectiveness was expected for the remaining xanthone-derivatives. X14 and X25 bear a hydroxyl group but the position is not favorable to a hydrogen atom or electron transfer56. The naked xanthone and the methoxylated xanthones did not show any activity. That is not surprising and confirms the key role of the phenol group in the antioxidant activity.

210

Chapter 5

5-3.12 Quercetin-derivatives extracted from Flavo Psidium Cattleianium (7 antioxidants among the 7 compounds)

R2 R4

HO O R5

R1 OH O

R1 R2 R3 R4 R5 Profile Morin –OH –OH –H –OH –H A2 RHC1 Myricetin –OH –H –OH –OH –OH B2 RHC2 Luteolin –H –H –OH –OH –H A2 RHC3 Kaempferol –OH –H –H –OH –H A2 RHC4 O-2-tetrahydro-2H-pyran- RHB4 –H –OH –OH –H F9 3,4,5-triol O-2-(3R,4S,5S)-tetrahydro- RHB5 –H –OH –OH –H F9 2H-pyran-3,4,5-triol O-2-(2S,3R,4R,5S)- RHB7 tetrahydro-6-methyl-2H- –H –OH –OH –H F9 pyran-3,4,5-triol

Figure 5-35 : Structures of quercetin-derivatives.

The antioxidant properties of morin, myricetin, luteolin, and kaempferol are well-known65,66,99,121. They are among the seven compounds extracted from Psidium cattleianum and were not taken into account for the cluster analyses. Thanks to their rapid and intermediate DPPH• scavenging rate as well as their 211

Chapter 5 high potency in scavenging the three radicals in the four methods, morin, luteolin, and kaempferol have a caffeic acid-like antioxidant profile (annex II-Q).

Their pEC50 is slightly better than that of caffeic acid in the ALP and ORAC assays and similar to that of quercetin as expected. However their ER50 value in the DPPH• and ABTS•– assays is much closer to that of caffeic acid rather than that of quercetin. Only lutolein is a DPPH• scavenger as potent as quercetin as it was already described122 but reacts more rapidly with DPPH• than the reference compound do (86% of DPPH• scavenged in a minute by luteolin when quercetin scavenges only 22%). The faster scavenging rate of these three compounds and myricetin relative to that of quercetin has been reported before123. Myricetin is a quercetin-like antioxidant compound. Its structure possesses a further hydroxyl group on the B-ring which solely affects the peroxyl radical detoxifying activity. Indeed, the pEC50 values (5.62 and 5.61 in the ALP and ORAC assays) are slightly lower than those of quercetin. Therefore, the additional active antioxidant residue does not enhance the antioxidant properties. The four quercetin-derivatives predominantly scavenge the DPPH• radical through a hydrogen atom transfer since they are as potent in ethanol as they are in acidic conditions, despite a SPLET mechanism-concerted toward the DPPH• scavenging activity of luteolin. The four compounds owe their classification among the most potent antioxidant to their structure. As it has been established for quercetin, the catechol moiety, the 4’-OH, the 3-OH, and the 2,3 double bond induce a stabilization to the hydrogen-abstraction-resulting radical. Moreover, the oxidation by-products exhibit antioxidant properties63,64. A catechol moiety appears to be required to forcefully scavenge the DPPH• radical. Indeed, myricetin and luteolin are more potent than morin and kaempferol. Besides, the 3-OH group seems to be involved in the predominance of the hydrogen atom transfer since a SPLET mechanism contributes to the scavenging ability of luteolin, which lacks this hydroxyl group. The hydroxyl residue also plays a role in the ABTS•– reducing activity since lutolein is the less potent compound. The ortho-quinone/para-quinoid tautomerism induced after the hydrogen abstraction cannot take place without the 3-OH residue, which inhibits the solvent addition

212

Chapter 5 and therefore less radical molecules are scavenged63,64. However, kaempferol has the poorest activity in ABTS•– acidic conditions, which means that the lack of a catechol moiety has a more important impact than the lack of 3-OH group on the highly likely electron transfer in acidic conditions. As regards the peroxyl radicals, the 3-OH-lacking compound, namely lutolein, exhibited the greatest activity in the assays, even higher than that of quercetin in ORAC method. But the gap between the parameters of the compounds is not relevant enough to accurately target the structural part involving in the peroxyl-radical-detoxifying activity. The three remaining extracted quercetin-derivatives did not exhibit such antioxidant properties (annex II-Q). They are ether of quercetin bearing a hydroxyl-substituted tetrahydropyrane. They are able to scavenge the DPPH• radical with potency; the ER50 value of RHB5 belongs to the class 3 when that of RHB4 and RHB7 is ranked into the class 2. The capacity for protecting the protein and fluorescein follows the same tendency but RHB7 in ALP method since it has been revealed as an ALP inhibitor. The acidic conditions in DPPH• assay highlighted a predominant SPLET mechanism, which leads to strongly rise the ER50 value. This confirms the involvement of the 3-OH group in the hydrogen atom transfer. The most important antioxidant behavior difference concerns the ABTS•– reducing ability. Indeed, despite a resorcinol and catechol moieties, the ether derivatives of quercetin exhibited a weak capacity toward this radical with an ER50 value around 0.55. Protic media enhanced the capacity of RHB5 toward ABTS•– but dropped that of RHB4 and even inhibited that of RHB7. Besides the lack of the 3-OH group, the tatrahydropyrane-induced steric hindrance may reduce the antioxidant capacities. It has been proven that a sugar on the C3 induces a torsion of the structure124 whereas the 3-OH group induces a planar structure. Therefore, the radical stabilization and the electron delocalization are favored4,123. Studies on rutin, a quercetin-3-rutinoside, has been performed toward ABTS•– radical and have provided its potent activity, even greater than that of kaempferol and lutolein99,125. The tetrahydropyrane may induce a stronger torsion than the rutinose (rutin) does, which involves a severe decrease in the ABTS•– scavenging activity. The ethers of quercetin highlighted two distinct

213

Chapter 5 antioxidant behaviors toward DPPH• and peroxyl radicals on one hand and toward ABTS•– radicals on the other hand. Besides the dissimilar mechanisms (HAT and ET), a adduct formation could be involved in the altered ABTS•– scavenging activity. Studies21,126 have reported that the resorcinol moiety interacts with ABTS•– before the catechol does. Adducts produced from the linkage on the position 8 of catechin with ABTS•– have been isolated and identify using reverse-phase high performance liquid chromatography (HPLC) and electrospray ionization mass spectrometry (ESI-MS). The structure of catechin does not bear any 2,3 double bond thus it may be compared to the RHB- derivatives since the electron delocalization property of the double bond is inhibited by the tetrahydropyrane. Besides the inhibition of the electron delocalization, the tetrahydropyrane moiety may induce a steric hindrance and obstruct the addition of the radical. Therefore, the quercetin-derivatives, which bear a tetrahydropyrane, are capable of scavenging DPPH• and peroxyl radicals thanks to a SPLET mechanism from the catechol moiety, but, they are almost inactive toward ABTS•– radical since the tetrahydropyrane-induced torsion inhibits the electron delocalization over the structure and the steric hindrance suppresses an adduct formation. The antioxidant behavior of the three RHB compounds does not allow the ranking of these compounds among the defined antioxidant profiles since the potent and intermediate DPPH• scavengers from the set under investigation showed the same order of magnitude of activity toward ABTS•–. A new profile needs to be formed, labeled B9. The properties of this profile are a potent and intermediate activity toward both DPPH• and peroxyl radicals and a poor ability to scavenge the ABTS•– radical as well as a slow scavenging rate. The procedure used to reach the antioxidant profile has therefore been modified to lead to the profile B9 (Figure 5-36).

214

Chapter 5

E3 or E3 or ALP assay Class 0 inactive inactive compound compound Mannitol-like Mannitol-like Class 3, 2, 1 3, assay Class 1 Class ORAC E4 E3 D6 assay •– D7 assay Class 1 ORAC Class 1 Cluster C, D, E D, C, ABTS C5 D5 CD E assay ORAC ALP ALP assay C6, C7 C6, C7 C6 • B8 DPPH •– screening say procedure to reach the antioxidant profiles. the antioxidant profiles. to reach say procedure scavengers • assay 2 1 ABTS B1, B9 Class Class Slow or very slow B DPPH Class 2 assay ORAC B9 B1 2 Class •– 2 1 Class Class B2 ALP assay assay Class 3 Class Class 3 B1, B2 ABTS B1, B2, B9 B1, B2, A8 A1 A Class 2 assay ORAC Class 2 Class 2 •– Figure 5-36: Modification of the well-ordered as well-ordered Modification of the Figure 5-36: A2 ALP assay Class 3 Class assay Class 3 A1, A2 A1, A2 ABTS

215

Chapter 5

5-3.13 Trolox®-derivatives (7 antioxidants among the 7 compounds)

HO O

R2N

R1 R2 Profile

COL-1 -H -(CH2)2-indole B1 COL-22 -H quinoline B8 COL-24 piperazine-trolox® B8 COL-30 -H -Ph-m-OH B1

COL-50 -H -(CH2)2-Ph-m,p-OMe B8 COL-52 -H -Et B8

COL-62 -H -(CH2)2-SO3H B8

Figure 5-37: Structures of trolox®-derivatives.

Trolox® is a well-known antioxidant compound with a water-soluble ability. However, under certain conditions, it may be less potent. In this investigation, it showed a poor activity toward the ROO• radicals and an intermediate DPPH• and ABTS•– scavenging capacity. Trolox®-derivatives are synthesized to enhance the trolox® antioxidant properties. COL-1127 and COL-50128 are known as inhibitors of iron-induced lipid peroxidation. COL-1 and COL-30 have been characterized by a chlorogenic acid-like antioxidant profile and the five other derivatives are cysteine-like compounds (see chapter 3). None of them possess trolox®-like antioxidant properties. COL-1 and COL-3 differ from trolox® in the peroxyl radical detoxifying capacity (annex

® II-R). Their pEC50 is ranked in a potency class above than that of trolox , and even two classes above as concerns the activity in the ALP assay of COL-30

® (pEC50 = 5.70 (class 3) when pEC50 = 4.85 (class 1) for trolox ). The R residue is

216

Chapter 5 an indole (COL-1) and a phenol (COL-30) moiety. The addition of these electron rich structures enhances the electron or hydrogen atom donating ability of the scaffold. This emphasizes the higher potency in the ROO• detoxifying activity. The DPPH• scavenging ability is slightly greater while that toward ABTS•– radicals is almost similar to that of trolox®. However, the DPPH• scavenging rate is lower than that of the reference compound. Therefore, the R residue obviously plays a role in the HAT mechanism. The steric hindrance slows down the scavenging rate but the indole and phenol moieties, through their hydrogen atom donating ability, enhance the activity, especially toward the ROO• radicals. The phenol residue (COL-30) acts through a SPLET mechanism since the scavenging capacity decreases in the mixture EtOH/AcOH. The other compounds mainly differ from trolox® in their slow DPPH• scavenging rate. The steric hindrance is obviously the main reason since the percentage of DPPH• scavenged in a minute drastically dropped when COL-24 was assessed (10% of scavenged DPPH• when 59% with trolox®). However, the presence of two trolox® residues enhances the scavenging activity toward this radical. Indeed the COL-24 scaffold bears two trolox® moieties linked by a

® piperazine cycle; the ER50 value of this compound is 0.13. Unlike trolox , this activity decreases in acidic conditions, which emphasizes a SPLET-concerted HAT mechanism in ethanol. COL-24 also showed a higher potency toward ROO• radicals but the efficiency in protecting the protein was not over 70%, even at high concentration of compound. A weak efficiency was also revealed despite a great potency for the compound COL-22. The pEC50 is 5.50 but the efficiency of the compound was not over 30% of protection. In that case, the results from the ORAC method reflect a more specific behavior toward the ROO• radicals. The quinoline residue does not affect the DPPH• scavenging activity relative to that of trolox® but it hinders the rearrangement in the acidic conditions and induces a steric hindrance which decreases the ABTS•– scavenging activity. The residues on the structures of the compounds COL-50, COL-52, and COL-62 do not affect the trolox® antioxidant properties but they do not allow the rearrangement in the EtOH/AcOH mixture.

217

Chapter 5

The DPPH• scavenging activity of any trolox®-derivatives did not enhance in the protic media, therefore they did not undergo the rearrangement that trolox did.

5-3.14 SAR overview regarding the antioxidant profiles

The results obtained from the screening and potency assessing methods pointed out the difference between the stability of the three radicals. Generally, all DPPH• scavengers are able to reduce ABTS•– and peroxyl radicals. The same is true of ABTS•– scavengers capable of interacting with peroxyl radicals. That emphasizes the higher stability of DPPH• relative to that of ABTS•–, the latter is more stable than the peroxyl radicals. This stability ranking was expected since the peroxyl radicals are known to be highly reactive74. The phenol compounds included in the kinetic cluster A are characterized by the ability to scavenge through both mechanisms HAT and ET. The hydrogen atom transfer may be concerted with a SPLET in ionization-promoting solvents. The latter is the main mechanism of action of the compounds with an antioxidant profile A1. The mechanism feature of the antioxidant profile A8 is a pH- dependant HAT mechanism and a pH-independent ET. The compounds that constitute the cluster B are also able to transfer electrons and hydrogen atoms. The hydrogen transfer occurs through both HAT and SPLET. The compounds characterized by the antioxidant profiles A1, A2, B1, and B2 are all polyphenols. Except melatonin, all compounds characterized by the antioxidant profile C5 bear a phenol residue on their structure. They reduce the radicals through an electron transfer, melatonin included. The phenol ionization contributes to the potency of these compounds. The entity, which forms the antioxidant profile D5 on its own, has phenol moieties as well and follows the same mechanism as the phenol-like compounds. The antioxidant profile C6 describes the compounds which solely reduce the ABTS•– radical through an electron transfer. They bear either a secondary or a tertiary amine or a thiol group. The primary amines constitute the profile D6

218

Chapter 5 since they are less potent at transferring an electron to the ABTS•– radical than the compounds of the profile C6. It is interesting to note that both profiles C6 and D6 are constituted by non-aromatic compounds. The compounds belonging to the potency cluster 7 and split into the profiles C7 and D7 mainly react through electron transfer. The scaffolds are richer in electrons than those of the profiles C6 and D6. Therefore, uric acid-like compounds and those characterized by the antioxidant profile D7 are capable of interacting with peroxyl radicals in addition to ABTS•– radicals. Unlike the compounds characterized by the antioxidant profiles C6 and D6, the peroxyl-radical scavengers included in the profile E4 are aromatic compounds. Their electron-rich structure allows the ROO• detoxification. However, some features including the electron withdrawing effect of residues avoid the DPPH• and ABTS•– scavenging activity. The ALP protective compounds (E3) are likely involved in protein binding through hydrogen bonds. Both hydrogen bond acceptors and donors are present on their structure allowing the protein binding.

5-4 Conclusion

The chemically diversified scaffolds provide interesting data concerning the reduction mechanism and the structure-activity relationships. First of all, the ABTS•– radical is obviously reduced by an electron transfer. The similar activity in ethanol and in acidic media as well as some cores (primary, secondary, tertiary amines, indole moiety, and azaaurone-derivative) capable of scavenging this radical support this mechanism. The ability to undergo further reactions after the oxidation plays a key role for high antioxidant potency, and an electron rich structure is required for the detoxification of peroxyl radicals. Antioxidant properties toward peroxyl and ABTS•– radicals have been proven for some aurone- and azaaurone-derivatives, conduramine B-1 and F-1 analogs, N- derivatives of pyrrolidine-3,4-diols, and conduramine F-1 epoxide. In addition to scavenging ABTS•– and peroxyl radicals, the DPPH• radical scavenging ability

219

Chapter 5 has been proven for the extracts from Jacaranda caucana, quercetin- and trolox®- derivatives. Finally, some 1-substituted and 2-substituted pyrazolopyrrolizine- derivatives exhibited an ALP protein protecting activity. Furthermore, the mechanisms which the compounds act through, namely HAT, SPLET, and ET, have been detected. The aza-aurone-derivatives, two aurone-derivatives, three substituted pyrazolopyrrolizine-derivatives, and three xanthone-derivatives have been detected as multifunctional compounds. A new antioxidant profile has been formed thanks to three quercetin- derivatives. The antioxidant properties are a potent capacity for the DPPH• scavenging through a slow rate, a potent ability to detoxify the ROO• radicals, and a poor activity toward ABTS•–.

220

Chapter 5

Bibliographic references

1. Metodiewa, D.; Jaiswal, A. K.; Cenas, N.; Dickancaite, E.; Segura-Aguilar, J. Quercetin may act as a cytotoxic prooxidant after its metabolic activation to semiquinone and quinoidal product. Free Rad. Biol. Med. 1999, 26, 107- 116. 2. Boots, A. W.; Kubben, N.; Haenen, G. R. M. M.; Bast, A. Oxidized quercetin reacts with thiols rather than with ascorbate: implication for quercetin supplementation. Biochem. Biophys. Res. Commun. 2003, 308, 560-565. 3. Galati, G.; Moridani, M. Y.; Chan, T. S.; O'Brien, P. J. Peroxidative metabolism of apigenin and naringenin versus luteolin and quercetin: Glutathione oxidation and conjugation. Free Rad. Biol. Med. 2001, 30, 370-382. 4. Leopoldini, M.; Marino, T.; Russo, N.; Toscano, M. Antioxidant properties of phenolic compounds: H-atom versus electron transfer mechanism. J. Phys. Chem. A 2004, 108, 4916-4922. 5. Wright, J. S.; Johnson, E. R.; DiLabio, G. A. Predicting the activity of phenolic antioxidants: Theoretical method, analysis of substituent effects, and application to major families of antioxidants. J. Am. Oil Chem. Soc. 2001, 123, 1173-1183. 6. Litwinienko, G.; Ingold, K. U. Abnormal solvent effects on hydrogen atom abstraction. 2. Resolution of the curcumin antioxidant controversy. The role of sequential proton loss electron transfer. J. Org. Chem. 2004, 69, 5888-5896. 7. Litwinienko, G.; Ingold, K. U. Abnormal solvent effects on hydrogen atom abstractions. 1. The reactions of phenols with 2,2-diphenyl-1- picrylhydrazyl (dpph(center dot)) in alcohols. J. Org. Chem. 2003, 68, 3433-3438. 8. Krishnamachari, V.; Levine, L. H.; Paré, P. W. Flavonoid oxidation by the radical generator AIBN: A unified mechanism for quercetin radical scavenging. J. Agric. Food Chem. 2002, 50, 4357-4363. 9. Amorati, R.; Pedulli, G. F.; Cabrini, L.; Zambonin, L.; Landi, L. Solvent and pH effects on the antioxidant activity of caffeic and other phenolic acids. J. Agric. Food Chem. 2006, 54, 2932-2937. 10. Ozgen, M.; Reese, R. N.; Tulio, A. Z.; Scheerens, J. C.; Miller, A. R. Modified 2,2- azino-bis-3-ethylbenzothiazoline-6-sulfonic acid (ABTS) method to measure antioxidant capacity of selected small fruits and comparison to ferric reducing antioxidant power (FRAP) and 2,2 '-diphenyl-1- picrylhydrazyl (DPPH) methods. J. Agric. Food Chem. 2006, 54, 1151- 1157. 11. Campos, A. M.; Lissi, E. A. Kinetics of the reaction between 2,2'-azinobis(3- ethylbenzothiazoline-6-sulfonic acid) (ABTS) derived radical cations and phenols. Int. J. Chem. Kinetics 1997, 29, 219-224. 12. Labrinea, E. P.; Georgiou, C. A. Stopped-flow method for assessment of pH and timing effect on the ABTS total antioxidant capacity assay. Anal. Chim. Acta 2004, 526, 63-68. 13. Tan, D. X.; Hardeland, R.; Manchester, L. C.; Poeggeler, B.; Lopez-Burillo, S.; Mayo, J. C.; Sainz, R. M.; Reiter, R. J. Mechanistic and comparative studies of melatonin and classic antioxidants in terms of their interactions with the ABTS cation radical. J. Pineal Res. 2003, 34, 249-259. 14. Thaipong, K.; Boonprakob, U.; Crosby, K.; Cisneros-Zevallos, L.; Hawkins Byrne, D. Comparison of ABTS, DPPH, FRAP, and ORAC assays for estimating

221

Chapter 5

antioxidant activity from guava fruit extracts. J. Food Comp. Anal. 2009, 19, 669-675. 15. Chegaev, K.; Lazzarato, L.; Rolando, B.; Marini, E.; Tosco, P.; Cena, C.; Fruttero, R.; Bertolini, F.; Reist, M.; Carrupt, P. A.; Lucini, V.; Fraschini, F.; Gasco, A. No-donor melatonin derivatives: synthesis and in vitro pharmacological characterization. J. Pineal Res. 2007, 42, 371-385. 16. Marjasvaara, A.; Janis, J.; Vainiotalo, P. Oxidation of a laccase mediator ABTS as studied by ESI-FTICR mass spectrometry. J. Mass. Spectrom. 2008, 43, 470-477. 17. Rosen, J.; Than, N. N.; Koch, D.; Poeggeler, B.; Laatsch, H.; Hardeland, R. Interactions of melatonin and its metabolites with the ABTS cation radical: extension of the radical scavenger cascade and formation of a novel class of oxidation products, C2-substituted 3-indolinones. J. Pineal Res. 2006, 41, 374-381. 18. Sekher Pannala, A.; Chan, T. S.; O'Brien, P. J.; Rice-Evans, C. A. Flavonoid B- ring chemistry and antioxidant activity: fast reaction kinetics. Biochem. Biophys. Res. Commun. 2001, 282, 1161-1168. 19. Perez-Jimenez, J.; Saura-Calixto, F. Anti-oxidant capacity of dietary polyphenols determined by ABTS assay: a kinetic expression of the results. Int. J. Food Sci. Technol. 2008, 43, 185-191. 20. Mordido, S. C.; Rebelo, M. J. F. Electrochemistry of caffeic acid in acetate- ethanolic solutions. Portugal. Electrochim. Acta 2006, 24, 313-322. 21. Nenadis, N.; Wang, L. F.; Tsimidou, M.; Zhang, H. Y. Estimation of scavenging activity of phenolic compounds using the ABTS+ assay. J. Agric. Food Chem. 2004, 52, 4669-4674. 22. Giacomelli, C.; Ckless, K.; Galato, D.; Miranda, F. S.; Spinelli, A. Electrochemistry of caffeic acid aqueous solutions with pH 2.0 to 8.5. J. Braz. Chem. Soc. 2002, 13, 332-338. 23. Hotta, H.; Ueda, M.; Nagano, S.; Tsujino, Y.; Koyama, J.; Osakai, T. Mechanistic study of the oxidation of caffeic acid by digital simulation of cyclic voltammograms. Anal. Biochem. 2002, 303, 66-72. 24. Beltran, J. L.; Sanli, N.; Fonrodona, G.; Barron, D.; Özkan, G.; Barbosa, J. Spectrophotometric, potentiometric and chromatographic pKa values of polyphenolic acids in water and acetonitrile-water media. Anal. Chim. Acta 2003, 484, 253-264. 25. Kallel Trabelsi, S.; Belhadj Tahar, N.; Abdelhedi, R. Electrochemical behavior of caffeic acid. Electrochim. Acta 2004, 49, 1647-1654. 26. Soobrattee, M. A.; Neergheen, V. S.; Luximon-Ramma, A.; Aruoma, O. I.; Bahorun, T. Phenolics as potential antioxidant therapeutic agents: mechanism and actions. Mutation Res. 2005, 579, 200-213. 27. Roginsky, V. Chain-breaking antioxidant activity of natural polyphenols as determined during the chain oxidation of methyl linoleate in Triton X-100 micelles. Arch. Biochem. Biophys. 2003, 414, 261-270. 28. Soares, D. G.; Andreazza, A. C.; Salvador, M. Sequestering ability of butylated hydroxytoluene, propyl gallate, resveratrol, and vitamins C and E against ABTS, DPPH, and hydroxyl free radicals in chemical and biological systems. J. Agric. Food Chem. 2003, 51, 1077-1080. 29. Regoli, F.; Winston, G. W. Quantification of total oxidant scavenging capacity of antioxidants for peroxynitrite, peroxyl radicals, and hydroxyl radicals. Toxicol. Appl. Pharmacol. 1999, 156, 96-105.

222

Chapter 5

30. Wang, S.; Milne, G. W. A.; Yan, X.; Posey, I. J.; Nicklaus, M. C.; Graham, L.; Rice, W. G. Discovery of novel, non-peptide HIV-1 protease inhibitors by pharmacophore searching. J. Med. Chem. 1996, 39, 2047-2054. 31. Sawai, Y.; Moon, J. H. NMR snalytical spproach to vlarify the molecular mechanisms of the sntioxidative and tadical-dcavenging sctivities of sntioxidants in tea using 1,1-Diphenyl-2-picrylhydrazyl. J. Agric. Food Chem. 2000, 48, 6247-6253. 32. Kilmartin, P. A.; Zou, H.; Waterhouse, A. L. A cyclic voltammetry method suitable for characterizing antioxidant properties of wine and wine phenolics. J. Agric. Food Chem. 2001, 49, 1957-1965. 33. Dar, A.; Faizi, S.; Naqvi, S.; Roome, T.; Zikr-ur-Rehman, S.; Ali, M.; Firdous, S.; Moin, S. T. Analgesic and antioxidant activity of mangiferin and its derivatives: the structure activity relationship. Biol. Pharm. Bull. 2005, 28, 596-600. 34. Lo Scalzo, R. Organic acids influence on DPPH. scavenging by ascorbic acid. Food Chem. 2008, 107, 40-43. 35. Fujinami, Y.; Tai, A.; Yamamoto, I. Radical scavenging activity against 1,1- diphenyl-2-picrylhydrazyl of ascorbic acid 2-glucoside (AA-2G) and 6-acyl- AA-2G. Chem. Pharm. Bull. 2001, 49, 642-644. 36. Shukla, M. K.; Mishra, P. C. Electronic structures and spectra of two antioxidants, uric acid and ascorbic acid. J. Mol. Struct. 1996, 377, 247- 259. 37. Arts, M. J. T. J.; Haenen, G. R. M. M.; Voss, H. P.; Bast, A. Antioxidant capacity of reaction products limits the applicability of the Trolox Equivalent Antioxidant Capacity (TEAC) assay. Food Chem. Toxicol. 2004, 42, 45-49. 38. Friaa, O.; Brault, D. Kinetics of the reaction between the antioxidant Trolox (R) and the free radical DPPH center dot in semi-aqueous solution. Org. Biomol. Chem. 2006, 4, 2417-2423. 39. Yamauchi, R.; Kato, K.; Ueno, Y. Formation of trimers of á-tocopherol and its model compound, 2,2,5,7,8-pentamethylchroman-6-ol, in autoxidizing methyl linoleate. Lipids 1988, 23, 779-783. 40. Karbarz, M.; Malyszko, J. Voltammetric behavior of trolox in methanol and ethanol solutions. Electroanalysis 2008, 20, 1884-1890. 41. Huang, S. W.; Frankel, E. N.; Schwarz, K.; German, J. B. Effect of pH on antioxidant activity of alpha-tocopherol and Trolox in oil-in-water emulsions. J. Agric. Food Chem. 1996, 44, 2496-2502. 42. Malyszko, J.; Karbarz, M. Electrochemical oxidation of trolox and [alpha]- tocopherol in acetic acid: A comparative study. J. Electroanal. Chem. 2006, 595, 136-144. 43. Capitani, C. D.; Carvalho, A. C. L.; Rivelli, D. P.; Barros, S. B. M.; Castro, I. A. Evaluation of natural and synthetic compounds according to their antioxidant activity using a multivariate approach. Eur. J. Lipid. Sci. Technol. 2009, 111, 1090-1099. 44. Born, M.; Carrupt, P. A.; Zini, R.; Brée, F.; Tillement, J. P.; Hostettmann, K.; Testa, B. Electrochemical behaviour and antioxidant activity of some natural polyphenols. Helv. Chim. Acta 1996, 79, 1147-1158. 45. Kweon, M. H.; Hwang, H. J.; Sung, H. C. Identification and antioxidant activity of novel chlorogenic acid derivatives from bamboo (Phyllostachys edulis). J. Agric. Food Chem. 2001, 49, 4646-4655. 46. Espin, J. C.; Wichers, H. J. Study of the oxidation of resveratrol catalyzed by polyphenol oxidase. Effect of polyphenol oxidase, laccase and peroxidase on the antiradical capacity of resveratrol. J. Food Biochem. 2000, 24, 225-250.

223

Chapter 5

47. Walker, R. B.; Everette, J. D. Comparative reaction rates of various antioxidants with ABTS radical cation. J. Agric. Food Chem. 2009, 57, 1156-1161. 48. Namazian, M.; Zare, H. R. Electrochemistry of chlorogenic acid: experimental and theoretical studies. Electrochim. Acta 2005, 50, 4350-4355. 49. Lu, Z.; Nie, G.; Belton, P. S.; Tang, H.; Zhao, B. Structure-activity relationship analysis of antioxidant ability and neuroprotective effect of gallic acid derivatives. Neurochem. Int. 2006, 48, 263-274. 50. Gunckel, S.; Santander, P.; Cordano, G.; Ferreira, J.; Munoz, S.; Nunez-Vergara, L. J.; Squella, J. A. Antioxidant activity of gallates: an electrochemical study in aqueous media. Chem. Biol. Interactions 1998, 114, 45-59. 51. Kasture, V. S.; Katti, S. A.; Mahajan, D.; Wagh, R.; Mohan, M.; Kasture, S. B. Antioxidant and antiparkinson activity of gallic acid derivatives. Pharmacologyonline 2009, 1, 385-395. 52. Kawabata, J.; Okamoto, Y.; Kodama, A.; Makimoto, T.; Kasai, T. Oxidative dimers produced from protocatechuic and gallic esters in the DPPH radical scavenging reaction. J. Agric. Food Chem. 2002, 50, 5468-5471. 53. Wang, M.; Jin, Y.; Ho, C. T. Evaluation of resveratrol derivatives as potential antioxidants and identification of a reaction product of resveratrol and 2,2- Diphenyl-1-picryhydrazyl radical. J. Agric. Food Chem. 1999, 47, 3974- 3977. 54. Corduneanu, O.; Janeiro, P.; Brett, A. M. O. On the electrochemical oxidation of resveratrol. Electroanalysis 2006, 18, 757-762. 55. Zhang, H.; Xu, L.; Zheng, J. Anodic voltammetric behavior of resveratrol and its electroanalytical determination in pharmaceutical dosage form and urine. Talanta 2007, 71, 19-24. 56. Migliavacca, E. Quantum Chemical Descriptors In Quantitative Structure-Activity Relationship; PhD Thesis: University of Geneva, 1998. 57. Gomez-Zaleta, B.; Ramirez-Silva, M. T.; Gutierrez, A.; Gonzalez-Vergara, E.; G³izado-Rodriguez, M.; Rojas-Hernandez, A. UV/vis, 1H, and 13C NMR spectroscopic studies to determine mangiferin pKa values. Spectrochim. Acta Part A 2006, 64, 1002-1009. 58. Mishra, B.; Priyadarsini, K. I.; Sudheerkumar, M.; Unnikrishhnan, M. K.; Mohan, H. Pulse radiolysis studies of mangiferin: s C-glycosyl xanthone isolated from Mangifera indica. Radiat. Phys. Chem. 2006, 75, 70-77. 59. Belguendouz, L.; Fremont, L.; Linard, A. Resveratrol inhibits metal ion- dependent and independent peroxidation of porcine low-density lipoproteins. Biochem. Pharmacol. 1997, 53, 1347-1355. 60. Makris, D. P.; Rossiter, J. T. Comparison of quercetin and a non-orthohydroxy flavonolas antioxidants by competing in vitro oxidation reactions. J. Agric. Food Chem. 2001, 49, 3370-3377. 61. Firuzi, O.; Lacanna, A.; Petrucci, R.; Marrosu, G.; Saso, L. Evaluation of the antioxidant activity of flavonoids by "ferric reducing antioxidant power" assay and cyclic voltammetry. Biochim. Biophys. Acta 2005, 1721, 174- 184. 62. Leopoldini, M.; Russo, N.; Chiodo, S.; Toscano, M. Iron chelation by the powerful antioxidant flavonoid quercetin. J. Agric. Food Chem. 2006, 54, 6343-6351. 63. Dangles, O.; Fargeix, G.; Dufour, C. One-electron oxidation of quercetin and quercetin derivatives in protic and non protic media. J. Chem. Soc. , Perkin Trans. 2 1999, 1387-1395. 64. Alluis, B.; Dangles, O. Quercetin (=2-(3,4-dihydroxyphenyl)-3,5,7-trihydroxy-4H- 1-benzopyran-4-one) glycosides and sulfates: Chemical synthesis,

224

Chapter 5

complexation, and antioxidant properties. Helv. Chim. Acta 2001, 84, 1133-1156. 65. Tsimogiannis, D. I.; Oreopoulou, V. The contribution of flavonoid C-ring on the DPPH free radical scavenging efficiency. A kinetic approach for the 3',4'- hydroxy substituted members. Innov. Food Sci. Emerg. Technol. 2006, 7, 140-146. 66. Cao, G.; Sofic, E.; Prior, R. L. Antioxidant and prooxidant behavior of flavonoids: structure-activity relationships. Free Rad. Biol. Med. 1997, 22, 749-760. 67. Blois, M. S. Antioxidant determinations by the use of a stable free radical. Nature 1958, 181, 1199-1200. 68. Aliaga, C.; Lissi, E. A. Reactions of the radical cation derived from 2,2 '- azinobis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS(.+)) with amino acids. Kinetics and mechanism. Canad. J. Chem. 2000, 78, 1052-1059. 69. Covington, A. D. Unhairing. In Tanning Chemistry: The Science of Leather; Covington, T. Ed.; The Royal Society of Chemistry: Cambridge, 2009; pp 112-133. 70. Allegra, M.; Reiter, R. J.; Tan, D. X.; Gentile, C.; Tesoriere, L.; Livrea, M. A. The chemistry of melatonin's interaction with reactive species. J. Pineal Res. 2003, 34, 1-10. 71. Pieri, C.; Marra, M.; Moroni, F.; Recchioni, R.; Marcheselli, F. Melatonin: a peroxyl radical scavenger more effective than vitamin E. Life Sci. 1994, 55, 271-276. 72. Antunes, F.; Barclay, L. R. C.; Ingold, K. U.; King, M.; Norris, J. Q.; Scaiano, J. C.; Xi, F. D. On the antioxidant activity of melatonin. Free Rad. Biol. Med. 1999, 26, 117-128. 73. Migliavacca, E.; Ancerewicz, J.; Carrupt, P. A.; Testa, B. Theoretical parameters to characterize antioxidants. Part 2. The cases of melatonin and carvedilol. Helv. Chim. Acta 1998, 81, 1337-1348. 74. Buettner, G. R. The packing order of free radicals and antioxidants: lipid peroxidation, α-tocopherol, and ascorbate. Arch. Biochem. Biophys. 1993, 300, 535-543. 75. Pisoschi, A. M.; Cheregi, M. C.; Danet, A. F. Total antioxidant capacity of some commercial fruit juices: electrochemical and spectrophotometrical approaches. Molecules 2009, 14, 480-493. 76. Ivekovic, D.; Milardovic, S.; Roboz, M.; Grabaric, B. S. Evaluation of the antioxidant activity by flow injection analysis method with electrochemically generated ABTS radical cation. Analyst 2005, 130, 708- 714. 77. Karoui, H.; Hogg, N.; Frejaville, C.; Tordo, P.; Kalyanaraman, B. Characterization of sulfur-centered radical intermediates formed during the oxidation of thiols and sulfite by peroxynitrite - ESR-SPIN trapping and oxygen uptake studies. J. Biol. Chem. 1996, 271, 6000-6009. 78. Gilbert, H. F. Thiol/disulfide exchange and redox potential of proteins. In Bioelectrochemisty Of Biomacromolecules; Lenaz, G.; Milazzo, G. Eds.; Birkhäuser Verlag: Basel, 1997; pp 256-284. 79. Simic, M. G.; Jovanovic, S. V. Antioxidation mechanisms of uric-scid. J. Am. Chem. Soc. 1989, 111, 5778-5782. 80. Allen, R. N.; Shukla, M. K.; Leszczynski, J. A theoretical study of the structure and properties of uric acid: A potent antioxidant. Int. J. Quant. Chem. 2004, 100, 801-809. 81. Klemperer, F. W. Enzymatic oxidation of uric acid. J. Biol. Chem. 1945, 160, 111- 121.

225

Chapter 5

82. Santos, C. X. C.; Anjos, E. I.; Augusto, O. Uric acid oxidation by peroxynitrite: Multiple reactions, free radical formation, and amplification of lipid oxidation. Arch. Biochem. Biophys. 1999, 372, 285-294. 83. Guskov, E. P.; Kletskii, M. E.; Kornienko, I. V.; Olekhnovich, L. P.; Chistyakov, V. A.; Shkurat, T. P.; Prokofev, V. N.; Zhdanov, Y. A. Allantoin as a free- radical scavenger. Doklady Biochem. Biophysics 2002, 383, 105-107. 84. Kostopanagiotou, G.; Pandazi, A. K.; Andreadou, I.; Markantonis, S. L.; Niokou, D.; Teloudis, A.; Costopanagiotou, C.; Arkadopoulos, N.; Smyrniotis, V. Effects of mannitol in the prevention of lipid peroxidation during liver resection with hepatic vascular exclusion. J. Clin. Anesthesia 2006, 18, 570-574. 85. Shen, B.; Jensen, R. G.; Bohnert, H. J. Mannitol protects against oxidation by hydroxyl radicals. Plant Physiol. 1997, 115, 527-532. 86. Aufmkolk, M.; Koehrle, J.; Hesch, R. D.; Cody, V. Inhibition of rat-liver iodothyronine deiodinase - Interaction of aurones with the iodothyronine ligand-binding site. J. Biol. Chem. 1986, 261, 1623-1630. Devel op ment and Vali dati on of Target- Based Drug Desi gn Tool s: Virt ual 87. Bravo, J. Screeni ng of Monoa mi ne Oxi dase I nhi bit ors ; PhD Thesis: University of Geneva, 2009. 88. Mercader, A. G.; Pomilio, A. B. QSAR study of flavonoids and biflavonoids as influenza H1N1 virus neuraminidase inhibitors. Eur. J. Med. Chem. 2010, 45, 1724-1730. 89. Okombi, S.; Rival, D.; Bonnet, S.; Mariotte, A. M.; Perrier, E.; Boumendjel, A. Discovery of benzylidenebenzofuran-3(2H)-one (aurones) as inhibitors of tyrosinase derived from human melanocytes. J. Med. Chem. 2006, 49, 329- 333. 90. Kayser, O.; Kiderlen, A. F. Leishmanicidal activity of aurones. Tokai J. Exp. Clin. Med. 1999, 23, 423-426. 91. Lee, C. Y.; Chew, E. H.; Go, M. L. Functionalized aurones as inducers of NAD(P)H:quinone oxidoreductase 1 that activate AhR/XRE and Nrf2/ARE signaling pathways: Synthesis, evaluation and SAR. Eur. J. Med. Chem. 2010, 45, 2957-2971. 92. Boumendjel, A.; Beney, C.; Deka, N.; Mariotte, A. M.; Lawson, M. A.; Trompier, D.; Baubichon-Cortay, H.; Di Pietro, A. 4-hydroxy-6-methoxyaurones with high-affinity binding to cytosolic domain of P-glycoprotein. Chem. Pharm. Bull. 2002, 50, 854-856. 93. Gerby, B.; Boumendjel, A.; Blanc, M.; Bringuier, P. P.; Champelovier, P.; Fortuné, A.; Ronot, X.; Boutonnat, J. 2-Arylidenedihydroindole-3-ones: Design, synthesis, and biological activity on bladder carcinoma cell lines. Bioorg. Med. Chem. Lett. 2007, 17, 208-213. 94. Desjardins, M.; Lallemand, M. C.; Freeman, S.; Hudlicky, T.; Abboud, K. A. Synthesis and biological evaluation of conduritol and conduramine analogs. J. Chem. Soc. , Perkin Trans. 1 1999, 621-628. 95. Lysek, R.; Schütz, C.; Vogel, P. Total asymmetric synthesis of (-)-conduramine B-1 and of its enantiomer. N-Benzyl derivatives of conduramine B-1 are [beta]- glucosidase inhibitors. Bioorg. Med. Chem. Lett. 2005, 15, 3071-3075. 96. Lysek, R.; Schutz, C.; Vogel, P. (1S,2S,3R,6R)-6-aminocyclohex-4-ene-1,2,3-triol (=(-)-conduramine B-1) is a selective inhibitor of alpha-mannosidases. Its inhibitory activity is enhanced by N-benzylation. Helv. Chim. Acta 2005, 88, 2788-2811. 97. Martin, F.; Hay, A. E.; Quinteros Condoretty, V. R.; Cressend, D.; Reist, M.; Gupta, M. P.; Carrupt, P. A.; Hostettmann, K. Antioxidant phenylethanoid 226

Chapter 5

glycosides and a neolignan from Jacaranda caucana. J. Nat. Prod. 2009, 72, 852-856. 98. Brand-Williams, W.; Cuvelier, M. E.; Berset, C. Use of a free radical method to evaluate antioxidant activity. Food Sci. Technol. 1995, 28, 25-30. 99. Rice-Evans, C. A.; Miller, N. J.; Paganga, G. Structure-antioxidant activity relationships of flavonoids and phenolic acids. Free Rad. Biol. Med. 1996, 20, 933-956. 100. Huang, D.; Ou, B.; Hampsch-Woodill, M.; Flanagan, J. A.; Prior, R. L. High- Throughput Assay of Oxygen Radical Absorbance Capacity (ORAC) Using a Multichannel Liquid Handling System Coupled with a Microplate Fluorescence Reader in 96-Well Format. J. Agric. Food Chem. 2002, 50, 4437-4444. 101. Villano, D.; Fernandez-Pachon, M. S.; Moya, M. L.; Troncoso, A. M.; Garcia- Parrilla, M. C. Radical scavenging ability of polyphenolic compounds towards DPPH free radical. Talanta 2007, 71, 230-235. 102. Martin, F.; Hay, A. E.; Cressend, D.; Reist, M.; Vivas, L.; Gupta, M. P.; Carrupt, P. A.; Hostettmann, K. Antioxidant C-Glucosylxanthones from the Leaves of Arrabidaea patellifera. J. Nat. Prod. 2008, 71, 1887-1890. 103. Gamiotea-Turro, D.; Cuesta-Rubio, O.; Prieto-Gonzalez, S.; De Simone, F.; Passi, S.; Rastrelli, L. Antioxidative constituents from the leaves of Hypericum styphelioides. J. Nat. Prod. 2004, 67, 869-871. 104. Hay, A. E.; Aumond, M. C.; Mallet, S.; Dumontet, V.; Litaudon, M.; Rondeau, D.; Richomme, P. Antioxidant santhones from garcinia vieillardii. J. Nat. Prod. 2004, 67, 707-709. 105. Jung, H. A.; Su, B. N.; Keller, W. J.; Mehta, R. G.; Kinghorn, A. D. Antioxidant xanthones from the pericarp of garcinia mangostana (Mangosteen). J. Agric. Food Chem. 2006, 54, 2077-2082. 106. Mahabusarakam, W.; Nuangnaowarat, W.; Taylor, W. C. Xanthone derivatives from Cratoxylum cochinchinense roots. Phytochemistry 2006, 67, 470-474. 107. Lee, B. W.; Lee, J. H.; Lee, S. T.; Lee, H. S.; Lee, W. S.; Jeong, T. S.; Park, K. H. Antioxidant and cytotoxic activities of xanthones from Cudrania tricuspidata. Bioorg. Med. Chem. Lett. 2005, 15, 5548-5552. 108. Pedro, M.; Cerqueira, F.; Sousa, M. E.; Nascimento, M. S. J.; Pinto, M. Xanthones as inhibitors of growth of human cancer cell lines and their effects on the proliferation of human lymphocytes in vitro. Bioorg. Med. Chem. 2002, 10, 3725-3730. 109. Sato, T.; Kawamoto, A.; Tamura, A.; Tatsumi, Y.; Fujii, T. Mechanism of antioxidant action of Pueraria glycoside (PG)-1 (an isoflavonoid) and mangiferin (a xanthonoid). Chem. Pharm. Bull. 1992, 40, 721-724. 110. Pardo-Andreu, G. L.; Delgado, R.; Nunez-Sellés, A. J.; Vercesi, A. E. Dual mechanism of mangiferin protection against iron-induced damage to 2- deoxyribose and ascorbate oxidation. Pharmacol. Res. 2006, 53, 253-260. 111. Jiang, D. J.; Jiang, J. L.; Zhu, H. Q.; Tan, G. S.; Liu, S. Q.; Xu, K. P.; Li, Y. J. Demethylbellidifolin preserves endothelial function by reduction of the endogenous nitric oxide synthase inhibitor level. J. Ethnopharmacol. 2004, 93, 295-306. 112. Gnerre, C.; Thull, U.; Gaillard, P.; Carrupt, P. A.; Testa, B.; Fernandes, E.; Silva, F.; Pito, M.; Pinto, M. M. M.; Wolfender, J. L.; Hostettmann, K.; Cruciani, G. Natural and synthetic xanthones as monoamine oxidase inhibitors: biological assay and 3D-QSAR. Helv. Chim. Acta 2001, 84, 552-570. 113. Urbain, A.; Marston, A.; Grilo, L. S.; Bravo, J.; Purev, O.; Purevsuren, B.; Batsuren, D.; Reist, M.; Carrupt, P. A.; Hostettmann, K. Xanthones from

227

Chapter 5

Gentianella amarella ssp. acuta with Acetylcholinesterase and Monoamine Oxidase Inhibitory Activities. J. Nat. Prod. 2008, 71, 895-897. 114. Salvi, A.; Brühlmann, C.; Migliavacca, E.; Carrupt, P. A.; Hostettmann, K.; Testa, B. Protein protection by antioxidants: development of a convenient assay and structure-activity relationships of natural polyphenols. Helv. Chim. Acta 2002, 85, 867-881. 115. Born, M. Redox studies of some natural and synthetic antioxidants. PhD Thesis; University of Lausanne: Lausanne, 1997. 116. Laviron, E. Electrochemical reactions with protonations at equilibrium. Part X. The kinetics of the p-benzoquinone/hydroquinone couple on a platinum electrode. J. Electroanal. Chem. 1984, 164, 213-227. 117. Desbenemonvernay, A.; Lacaze, P. C.; Cherigui, A. Uv-visible spectroelectrochemical study of some para-benzoquinoid and ortho- benzoquinoid compounds - Comparative-evaluation of their electrochromic properties. J. Electroanal. Chem. 1989, 260, 75-90. 118. Bailey, S. I.; Ritchie, I. M. A cyclic voltammetric study of the aqueous electrochemistry of some quinones. Electrochim. Acta 1985, 30, 3-12. 119. Lucarini, M.; Mugnaini, V.; Pedulli, G. F. Bond dissociation enthalpies of polyphenols: the importance of cooperative effects. J. Org. Chem. 2002, 67, 928-931. 120. Lucarini, M.; Peduli, G. F.; Guerra, M. A critical evaluation of the factors determining the effect of intramolecular hydrogen bonding on the O-H bond dissociation einithalpy of catechol and of flavonoid antioxidants. Chem. Eur. J. 2004, 10, 933-939. 121. Silva, M. M.; Santos, M. R.; Caroco, G.; Rocha, R.; Justino, G.; Mira, L. Structure- antioxidant activity relationships of flavonoids: A re-examination. Free Rad. Res. 2002, 36, 1219-1227. 122. Butkoviç, V.; Klasinc, L.; Bors, W. Kinetic study of flavonoid reactions with stable radicals. J. Agric. Food Chem. 2004, 52, 2816-2820. 123. Burda, S.; Oleszek, W. Antioxidant and antiradical activities of flavonoids. J. Agric. Food Chem. 2001, 49, 2774-2779. 124. van Acker, S. A. B. E.; de Groot, M. J.; van den Berg, D. J.; Tromp, M. N. J. L.; Donné-op den Kelder, G. M.; van der Vijgh, W. J. F.; Bast, A. A quantum chemical explanation of the antioxidant activity of flavonoids. Chem. Res. Toxicol. 1996, 9, 1305-1312. 125. Lien, E. J.; Ren, S. J.; Bui, H. Y. H.; Wang, R. B. Quantitative structure-activity relationship analysis of phenolic antioxidants. Free Rad. Biol. Med. 1999, 26, 285-294. 126. Osman, A. M.; Wong, K. K. Y.; Fernyhough, A. ABTS radical-driven oxidation of polyphenols: Isolation and structural elucidation of covalent adducts. Biochem. Biophys. Res. Commun. 2006, 346, 321-329. 127. Vajragupta, O.; Boonchoong, P.; Wongkrajang, Y. Comparative quantitative structure-activity study of radical scavengers. Bioorg. Med. Chem. 2000, 8, 2617-2628. 128. Jacobsen, E. J.; VanDoornik, F. J.; Ayer, D. E.; Belonga, K. L.; Braughler, J. M.; Hall, E. D.; Houser, D. J. 2-(Aminomethyl)chromans that inhibit iron- dependent lipid peroxidation and protect against nervous system trauma and ischemia. J. Med. Chem. 1992, 35, 4464-4472.

228

Conclusions and perspectives

Conclusions

Nowadays, the antioxidant compounds are of great interest due to their ability to prevent and delay oxidative stress-related diseases. The identity of reactive species and the oxidative processes involved in oxidative stress are mostly well-known. However, getting well-described features of antioxidant compounds is complex due to the high number of ROS/RNS/RCS and oxidation- altered biomolecules. This study has provided tools for the characterization process of antioxidant compounds. Parameters necessary for a high throughput screening have been set in four methods using three radicals of different stability, which are peroxyl, DPPH•, and ABTS•– radicals. The assays are suitable for HTS robotics and the screening parameters allow the identification of antioxidant hits from a library of compounds in a short time and using low concentrations. Thanks to the stability difference of the radicals and the involvement of the alkaline phosphatase in one of the assays, the four methods provide complementary data regarding the antioxidant properties of the compounds. These properties are based on the potency and the kinetics in counteracting the radicals. The assessment of antioxidant capacities of a structurally diverse 85-compound set has led to 14 profiles. An additional profile has been found for the antioxidant abilities of some quercetin-derivatives. The 15 profiles characterize the compounds depending on their antioxidant properties in the four assays and some of them are reference antioxidant-related. Furthermore, an antioxidant assay procedure has been established to allow the characterization of new chemical entities. Only two methods, one of them is a 10- minute screening, are necessary to reach some profiles and describe an NCE by the properties provided by the four methods.

229

Conclusions and perspectives

To further characterize the antioxidant properties, another parameter is described. The formal standard potential is an electrochemical parameter which defines the antioxidant power of a compound. However, no method has been well established so far to determine this feature for antioxidant compounds. Therefore, a potential scale has been designed, the steps of which have been set by the formal standard potential of ferrocene and two of its derivatives. The ability of a compound to regenerate the electrochemically oxidized ferrocene (or its derivatives) leads to place its E0’ with respect to that of ferrocene (or its derivatives). This method has allowed an estimation of the formal standard potential of caffeic acid, which is between 2 and 235 mV, and that of melatonin and para-hydroquinone, which is above 313 mV. As regards p-hydroquinone, the electrochemical reaction with aminoferrocene has lead to a limitation of the method. Indeed, by-products have been yielded from the reaction between the species, which has hindered accurate data concerning the position of the formal standard potential of p-HQ relative to that of aminoferrocene. Finally, the outcomes obtained from the methods and the diversity in the compound structures have allowed an antioxidant profile-related structure- activity relationship. The reducing mechanisms through HAT, SPLET, and ET have been detected. Furthermore, new antioxidant compounds toward DPPH•, ABTS•–, and peroxyl radicals have been revealed. Some of them have solely exhibited a capacity for protecting ALP against ROO•-induced oxidative damage. Their protective activity is likely based on a protein-binding ability at relevant sites which hinders the protein oxidation or leads the compounds to undergo the radical-induced oxidation instead of the protein. The aza-aurone-, two aurone-, three substituted pyrazolopyrrolozine-, and three xanthone-derivatives have been revealed as multifunctional compounds.

Perspective

The treatment of neurodegenerative diseases requires drugs which are capable of counteracting the reactive species in the lipid-rich brain, highly susceptible to oxidative damage. Therefore, the ability of the compound set to

230

Conclusions and perspectives prevent and inhibit the lipid peroxidation will be assessed and added, to further characterize the antioxidant profiles. Furthermore, the capacities for chelating metal and reducing iron will provide very useful data in the antioxidant characterization. Then, the addition of properties concerning DNA and carbohydrate oxidative damage and the RNS detoxification will complete the antioxidant profiles. In order to validate the electrochemical system used to estimate the formal standard potential of antioxidant compounds, this parameter will be calculated by simulation software for caffeic acid and melatonin. In addition, some compounds, for which the formal standard potential is known, will be assessed. N,N-dimethylaminoferrocene will be tested as species to regenerate in order to replace aminoferrocene and avoid nucleophilic additions or side reactions. Furthermore, to contract the potential ranges which provide an estimation of the formal standard potential, other well-known electrochemical species will be included in the system. Computational calculations of quantum chemistry will be performed to support the mechanisms detected in the four methods. Furthermore, the assessment of other structurally diverse compounds in the antioxidant assay procedure will release new antioxidant compounds and provide further information regarding the structure-activity relationship. This is of great interest in designing new multifunctional compounds. Regarding the compounds which have solely exhibited a protective activity of ALP against peroxyl radical-induced effectiveness loss, further investigations, such as computational methods, are necessary to figure out how and where they exactly act.

231

Annex I Supporting data

Table I-1: Compounds included into antioxidant profiles. AuD; aurone-derivatives, AzD: azaaurone-derivatives, B1: conduramine B-1 analogs, F1: conduramine F-1 analogs, F1o: conduramine F-1 epoxide, JC: extracts from Jacaranda caucana, MD: mangiferin-derivatives, NP: N-derivatives of pyrrolidine-3,4-diol, 1PD (2PD): 1-substituted (2-substituted) pyrazolopyrrolidizine derivatives, XD: xanthone-derivatives.

Groups Compounds Groups Compounds Groups Compounds A1 ID39 (XD) C5 Melatonin D5 SO-IV-561 (AuD) ID40 (XD) Phenol SO-I-39 (XD) D6 VRL-31 (NP) A2 Caffeic acid ROBI-26 (B1) VRL-40 (NP) JCA33 (JC) ROBI-27 (B1) VRL-58 (NP) JCA37 (JC) X13 (XD) VRL-69 (NP) JCA44 (JC) X24 (XD) ROBI-31 (B1) JCA47 (JC) X9 (XD) APLL3 (MD) ID22 (XD) D7 CB396 (AzD) CB398 (AzD) A8 Ascorbic acid C6 Glutathione ML44 (AzD) Trolox® VRL-43 (NP) ML45 (AzD) VRL-45 (NP) ROBI-163 (B1) B1 Chlorogenic acid VRL-71 (NP) ROBI-475 (F1o) Gallic acid VRL-75 (NP) Resveratrol VRL-76 (NP) E3 SO-II-233 (AuD) JCA7 (JC) VRL-77 (NP) CB393 (AzD) JCA42 (JC) VRL-78 (NP) ROBI-329 (F1) JCA52 (JC) VRL-79 (NP) ROBI-335 (F1) ID42 (XD) VRL-80 (NP) MC2372 (2PD) APL98 (MD) VRL-81 (NP) MC2385 (2PD) 7418 (MD) VRL-82 (NP) MC2386 (1PD) MC2388 (2PD) B2 Mangiferin C7 Uric acid MC2406 (2PD) Quercetin ML48 (AzD) MC2416 (2PD) APL58 (MD) ML40 (AzD) ID23 (XD) APL70 (MD) CB392 (AzD) APL101 (MD) CB395 (AzD) E4 Aniline CB397 (AzD) N,N-diethylaniline B8 Cysteine ROBI-486 (F1) CB290 (AuD) ROBI-33 (B1) ROBI-333 (F1) ROBI-470 (F1o) MC2405 (2PD) X8 (XD)

Annex I

Figure I-1: DPPH• scavenging activity of reference compounds at ratio 0.5 for 90 minutes. (The dotted line represents the end of the screening (10 minutes))

Annex I 1e-005

5e-006

ipa 0 i (A) 5 10 15 ipc Vν -5e-006

-1e-005 Figure I-2 : Plot of the peak currents versus the square root of the scan rate of aminoferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. r² = 0.99 for both oxidation and reduction processes.

1.5×10 -5

1.0×10 -5

5.0×10 -6 ipa 0 i (A) 5 10 15 ipc Vν -5.0×10 -6

-1.0×10 -5

-1.5×10 -5 Figure I-3 : Plot of the peak currents versus the square root of the scan rate of ferrocene (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. r² = 0.99 for both oxidation and reduction processes.

Annex I 2e-006

1e-006 i (A)

-1.0 -0.5 0.5 1.0 1.5 E (V)

-1e-006 Figure I-4 : Voltammogram of caffeic acid (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 0.5 mV.s-1.

2e-006

1e-006 i (A)

-1.0 -0.5 0.5 1.0 1.5 E (V)

-1e-006 Figure I-5 : Voltammogram of caffeic acid (1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 1 mV.s-1.

1.0×10 -5

5.0×10 -6

-1.0 i (A) -0.5 0.5 1.0 1.5 -5.0×10 -6 E (V)

-1.0×10 -5

Figure I-6 : Voltammogram of caffeic acid (1 mM) and ferrocene (0.1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 80 mV.s-1.

Annex I 4.0×10 -6

2.0×10 -6 i (A)

-1.0 -0.5 0.5 1.0 1.5 E (V) -2.0×10 -6 Figure I-7 : Voltammogram of caffeic acid (1 mM) and ferrocenecarboxwlic acid (0.1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 5 mV.s-1.

1.0×10 -5 i (A) -1.0 -0.5 0.5 1.0 1.5 E (V)

-1.0×10 -5

Figure I-8 : Voltammogram of caffeic acid (1 mM) and ferrocenecarboxwlic acid (0.1 mM) in borate buffer (pH = 8.3)/EtOH (50/50). WE: GC vs Ag/AgCl, CE: Pt. Scan rate: 100 mV.s-1.

Annex I 2 2 2.3 2log 2 2.3 RT pH F 0.059 . 0.699 0.059 8.3 . 0.209 Eq. I-1: Calculation of the formal standard potential of para-hydroquinone at pH = 8.3. θ (E ’p-BQ/p-HQ pH=0 = 0.699V)

Annex I Annex II Structures and antioxidant properties of studied compounds.

Cross-references and abbreviations used in each table: a: percentage of remaining ALP activity in 90 minutes at a compound concentration of 10-5M. b: percentage of protected fluorescein in 90 minutes at a compound concentration of 10-5M. c: percentage of DPPH• radical scavenged in 10 minutes at a ratio [compound]/[DPPH•] of 0.5. d: percentage of DPPH• radical scavenged in 1 minutes at a ratio [compound]/[DPPH•] of 0.5. e: percentage of ABTS•– radical scavenged in 10 minutes at a ratio [compound]/[ABTS•–] of 0.2. f: percentage of ABTS•– radical scavenged in 1 minutes at a ratio [compound]/[ABTS•–] of 0.2. NA: not active ND: not determined

Annex II-A Reference compounds.

Annex II-A.a Structures of reference compounds.

OH HO HO O OH NH2 HO O HO O OH O O NH2 HO OH HO HS OH O HO COOH Aniline Ascorbic acid Caffeic acid Chlorogenic acid Cysteine

O HO

HO HS O OH O OH O O O OH OH OH H HO N HOH2C HO HO N OH HO CH OH H 2 HO O OH OH NH2 O OH OH OH Gallic acid Glutathione Mangiferin Mannitol

OH H OH HO O N OH

HN OH HO O O OH O OH Melatonin Phenol Quercetin Resveratrol

Annex II-A CH3 O HO N CH HO O H 3 CH3 CH3 CH3 N HN O O H3C O N O N H HO H CH3 Trolox® Uric acid N,N-diethylaniline α-tocopherol

Annex II-A Annex II-A.b Antioxidant properties of reference compounds.

ALP assay ORAC assay DPPH• assay ABTS•– assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

Name Profile % DPPH• % DPPH• % ABTS•– % ABTS•– pEC50 pEC50 ER50 ER50 ER50 ER50 a b c e Screening Screening Screening scavengedd scavengedd Screening scavengedf scavengedf Class Class Class Class Class Class Class Class Class Class

4.51 ± 0.17 4.27 ± 0.03 0.28 ± 0.08 96 ± 1 0.14 ± 0.01 82 ± 2 0.30 ± 0.03 31 ± 5 0.33 ± 0.02 20 ± 1 Ascorbic acid 16 ± 6 1 ± 0 88 ± 8 29 ± 5 A8 1 1 2 3 3 3 2 1 2 0

6.03 ± 0.03 5.18 ± 0.02 ND 0 ± 1 ND 1 ± 2 ND Aniline 100 ± 7 63 ± 3 0 ± 1 ND ND 1 ± 2 ND E4 3 2 0 0 0 0 0

5.66 ± 0.10 5.65 ± 0.08 0.17 ± 0.01 50 ± 5 0.14 ± 0.03 14 ± 2 0.17 ± 0.03 45 ± 2 0.24 ± 0.03 34 ± 2 Caffeic acid 100 ± 6 89 ± 11 95 ± 1 51 ± 2 A2 3 3 3 2 3 0 3 1 2 1

Chlorogenic 5.68 ± 0.03 5.65 ± 0.08 0.11 ± 0.03 15 ± 2 0.16 ± 0.01 9 ± 1 0.26 ± 0.01 36 ± 2 0.23 ± 0.01 36 ± 3 100 ± 7 89 ± 11 53 ± 6 37 ± 3 B1 acid 3 3 3 0 3 0 2 1 2 1

4.54 ± 0.08 NA 0.76 ± 0.05 19 ± 3 0.42 ± 0.04 16 ± 1 0.07 ± 0.01 66 ± 5 0.42 ± 0.03 Cysteine 19 ± 9 8 ± 2 24 ± 5 75 ± 7 ND B8 1 0 1 0 2 0 3 2 2

5.31 ± 0.03 5.29 ± 0.06 0.04 ± 0.01 44 ± 7 0.05 ± 0.01 27 ± 4 0.06 ± 0.01 71 ± 2 0.05 ± 0.01 93 ± 4 Gallic acid 100 ± 2 85 ± 14 83 ± 1 84 ± 2 B1 2 2 3 1 3 1 3 2 3 3

NA NA NA 0 ± 2 NA 0.17 ± 0.01 28 ± 3 0 ± 1 Glutathione 14 ± 6 3 ± 1 0 ± 1 ND 50 ± 2 2.75 ± 0.55 C6 0 0 0 0 0 3 1 0

5.74 ± 0.17 6.04 ± 0.06 0.12 ± 0.06 15 ± 4 0.09 ± 0.02 7 ± 3 0.11 ± 0.01 64 ± 1 0.10 ± 0.01 84 ± 1 Mangiferin 100 ± 9 89 ± 10 40 ± 3 71 ± 1 B2 3 3 3 0 3 0 3 2 3 3

NA NA NA 1 ± 1 NA NA 0 ± 2 ND Mannitol 2 ± 7 0 ± 1 0 ± 1 ND 1 ± 0 ND - 0 0 0 0 0 0 0 0

5.20 ± 0.12 5.28 ± 0.05 NA 1 ± 1 NA 0.32 ± 0.07 5 ± 2 0.49 ± 0.08 1 ± 2 Melatonine 50 ± 7 51 ± 11 0 ± 1 ND 18 ± 1 C5 2 2 0 0 0 2 0 2 0

Annex II-A 5.66 ± 0.05 5.41 ± 0.23 ND 1 ± 1 ND 0.27 ± 0.01 19 ± 3 Phenol 100 ± 12 61 ± 4 1 ± 1 ND 39 ± 3 C5 3 2 0 0 0 2 0

5.95 ± 0.04 5.78 0.02 0.09 ± 0.01 22 ± 3 0.07 ± 0.01 15 ± 1 0.07 ± 0.01 58 ± 3 0.07 ± 0.01 59 ± 3 Quercetin 91 ± 8 95 ± 3 66 ± 3 82 ± 1 B2 3 3 3 0 3 0 3 2 3 2

5.72 ± 0.03 5.57 ± 0.04 0.46 ± 0.08 16 ± 4 0.70 ± 0.01 2 ± 1 0.12 ± 0.02 38 ± 1 0.12 ± 0.01 30 ± 2 Resveratrol 82 ± 4 90 ± 3 34 ± 1 50 ± 1 B1 3 3 2 0 1 0 3 1 3 1

0.20 ± 0.03 60 ± 6 0.12 ± 0.01 74 ± 1 0.26 ± 0.01 36 ± 1 0.25 ± 0.02 35 ± 4 α-tocopherol NS NS NS NS 94 ± 1 37 ± 2 - 2 2 3 2 2 1 2 1

4.85 ± 0.14 4.74 ± 0.12 0.21 ± 0.02 59 ± 7 0.11 ± 0.01 60 ± 0 0.28 ± 0.04 39 ± 1 0.27 ± 0.01 35 ± 1 Trolox 34 ± 11 15 ± 4 97 ± 1 40 ± 1 A8 1 1 2 2 3 2 2 1 2 1

4.89 ± 0.05 NA NA 2 ± 4 ND 0.25 ± 0.02 10 ± 2 >3 1 ± 8 Uric acid 29 ± 9 1 ± 0 2 ± 0 ND 29 ± 1 C7 1 0 0 0 0 2 0 0 0

N,N- 5.91 ± 0.04 5.94 ± 0.02 0 ± 1 0 ± 1 100 ± 5 81 ± 4 0 ± 1 ND ND ND 0 ± 2 ND ND ND E4 diethylaniline 3 3 0 0

NS: not soluble

Annex II-A Annex II-B Aurone-derivatives

Annex II-B.a Structures of aurone-derivatives.

Br Cl F N

MeO O MeO O MeO O MeO O

O O O O OMe OMe OMe OMe CB284 CB285 CB286 CB290

MeO O MeO O MeO O MeO O

OMe O OMe O OMe O OMe O CB301 CB303 CB304 CB305

Annex II-B MeO O HO MeO O HO O O S O O OH O O O OMe OH O OMe CB306 SO-I-39 SO-II-233 SO-II-356

HO O

O SO-IV-561

Annex II-B Annex II-B.b Antioxidant properties of aurone-derivatives.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

Name % DPPH• % DPPH• % ABTS•– % ABTS•– Profile pEC50 pEC50 ER50 ER50 ER50 ER50 a b c d Screening Screening Screening scavengedd scavengedd Screening scavengedf scavengedf Class Class Class Class Class Class Class Class Class Class

NA NA NA 2 ± 2 7 ± 9 CB284 6 ± 4 1 ± 1 1 ± 2 ND ND 7 ± 8 ND ND ND - 0 0 0 0 0

NA NA NA 2 ± 2 4 ± 3 CB285 11 ± 4 1 ± 1 2 ± 2 ND ND 4 ± 3 ND ND ND - 0 0 0 0 0

NA NA NA 4 ± 1 3 ± 2 CB286 16 ± 7 3 ± 2 6 ± 1 ND ND 3 ± 2 ND ND ND - 0 0 0 0 0

5.62 ± 0.03 5.49 ± 0.04 NA 0 ± 1 NA 6 ± 5 CB290 100 ± 1 76 ± 11 1 ± 2 ND ND 10 ± 4 ND ND E4 3 2 0 0 0 0

NA NA NA 2 ± 2 6 ± 3 CB301 13 ± 3 2 ± 1 12 ± 12 ND ND 6 ± 3 ND ND ND - 0 0 0 0 0

NA NA NA 2 ± 2 6 ± 5 CB303 13 ± 1 1 ± 1 11 ± 10 ND ND 6 ± 6 ND ND ND - 0 0 0 0 0

NA NA NA 2 ± 2 6 ± 7 CB304 9 ± 2 3 ± 3 2 ± 2 ND ND 6 ± 6 ND ND ND - 0 0 0 0 0

NA NA NA 0 ± 1 ND 1 ± 1 CB305 32 ± 7 13 ± 3 0 ± 1 ND 1 ± 2 ND ND ND - 0 0 0 0 0

NA NA NA 2 ± 2 5 ± 7 CB306 24 ± 2 4 ± 2 9 ± 10 ND ND 6 ± 6 ND ND ND - 0 0 0 0 0

5.69 ± 0.20 5.34 ± 0.12 NA 2 ± 3 0.13 ± 0.01 17 ± 1 0.97 ± 0.08 0 ± 5 SO-I-39 100 ± 9 73 ± 1 9 ± 8 ND ND 33 ± 2 C5 3 2 0 0 3 0 1 0

5.61 ± 0.13 NA NA 2 ± 2 3 ± 0 SO-II-233 55 ± 8 14 ± 8 15 ± 15 ND ND 3 ± 0 ND ND ND E3 3 0 0 0 0

Annex II-B NA NA NA 2 ± 2 4 ± 5 SO-II-356 0 ± 5 1 ± 1 11 ± 12 ND ND 4 ± 5 ND ND ND - 0 0 0 0 0

5.80 ± 0.27 5.52 ± 0.02 NA 2 ± 2 0.60 ± 0.02 21 ± 1 0.98 ± 0.22 0 ± 1 SO-IV-561 100 ± 4 75 ± 11 11 ± 11 ND ND 26 ± 2 D5 3 3 0 0 1 1 1 0

Annex II-B Annex II-C Azauaurone-derivatives.

Annex II-C.a Structures of azaaurone-derivatives.

H H H H MeO N MeO N MeO N MeO N OMe Cl

O O O O OMe OMe OMe OMe CB391 CB392 CB393 CB395

F OMe F Cl H H H H MeO N MeO N MeO N MeO N Cl Cl

O O O O OMe OMe OMe OMe CB396 CB397 CB398 CB400

Annex II-C Cl S

H H H H MeO N MeO N MeO N MeO N

O O O O OMe OMe OMe OMe ML40 ML44 ML45 ML48

Annex II-C Annex II-C.b Antioxidant properties of azaaurone-derivatives.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

NA 4.55 ± 0.05 NA 0 ± 4 0.70 ± 0.10 13 ± 0 0.52 ± 0.04 6 ± 1 CB391 26 ± 8 17 ± 2 0 ± 4 ND ND 19 ± 1 D7 0 1 0 0 1 0 1 0

5.00 ± 0.14 4.55 ± 0.09 NA 4 ± 4 0.45 ± 0.06 18 ± 1 0.53 ± 0.02 7 ± 1 CB392 46 ± 3 18 ± 2 2 ± 3 ND ND 25 ± 2 C7 2 1 0 0 2 0 1 0

4.57 ± 0.13 4.56 ± 0.06 NA 1 ± 4 6 ± 1 7 ± 4 CB393 30 ± 2 17 ± 5 5 ± 2 ND ND 10 ± 1 NA ND E3 1 1 0 0 0 0

5.31 ± 0.05 4.80 ± 0.08 NA 2 ± 4 0.32 ± 0.01 21 ± 1 0.31 ± 0.01 20 ± 1 CB395 87 ± 6 24 ± 5 2 ± 3 ND ND 28 ± 2 C7 2 1 0 0 2 0 2 0

4.86 ± 0.04 4.57 ± 0.11 NA 0 ± 9 0.72 ± 0.04 15 ± 1 0.47 ± 0.03 12 ± 1 CB396 36 ± 7 27 ± 4 2 ± 3 ND ND 18 ± 1 D7 1 1 0 0 1 0 2 0

4.87 ± 0.07 4.65 ± 0.13 NA 4 ± 3 0.45 ± 0.01 17 ± 1 0.39 ± 0.01 9 ± 1 CB397 40 ± 2 25 ± 6 2 ± 4 ND ND 22 ± 2 C7 1 1 0 0 2 0 2 0

4.65 ± 0.03 4.78 ± 0.07 NA 6 ± 4 0.81 ± 0.03 7 ± 1 0.62 ± 0.02 7 ± 3 CB398 40 ± 8 41 ± 2 7 ± 1 ND ND 14 ± 2 D7 1 1 0 0 1 0 1 0

NA NA NA 5 ± 4 ND 5 ± 0 CB400 26 ± 4 31 ± 2 2 ± 3 ND 12 ± 2 NA ND ND - 0 0 0 0 0

4.68 ± 0.11 4.79 ± 0.03 NA 5 ± 4 0.40 ± 0.04 18 ± 1 0.34 ± 0.01 15 ± 1 ML40 37 ± 2 23 ± 2 3 ± 3 ND ND 28 ± 2 C7 1 1 0 0 2 0 2 0

4.74 ± 0.19 4.74 ± 0.07 NA 2 ± 4 0.70 ± 0.07 15 ± 2 0.59 ± 0.01 3 ± 1 ML44 51 ± 13 23 ± 4 5 ± 5 ND ND 20 ± 3 D7 1 1 0 0 1 0 1 0

4.83 ± 0.02 4.61 ± 0.07 NA 4 ± 2 0.82 ± 0.13 13 ± 1 0.54 ± 0.03 11 ± 3 ML45 43 ± 6 19 ± 4 3 ± 1 ND ND 17 ± 2 D7 1 1 0 0 1 0 1 0

Annex II-C 5.22 ± 0.04 4.86 ± 0.06 NA 0 ± 2 0.39 ± 0.01 25 ± 2 0.35 ± 0.01 25 ±2 ML48 58 ± 1 35 ± 20 2 ± 2 ND ND 31 ± 2 C7 2 1 0 0 2 0 2 0

Annex II-C Annex II-D N-derivatives of (3S,4S)- and (3R,4R)-pyrrolidine-3,4-diols.

Annex II-D.a Structures of N-derivatives of (3S,4S)- and (3R,4R)-pyrrolidine-3,4-diols.

HO HO HO HO HO HO

N HO N HO HO N HO N N NH HO HO VRL-2 VRL-3 VRL-14 VRL-15 VRL-21 VRL-23

HO N HO N NH2 HO HO N N HN N NH HO HO 2 HO HO VRL-31 VRL-40 VRL-43 VRL-45

NH2 N N NH N N HO HO HO HO HO

HO HO HO HO HO VRL-52 VRL-55 VRL-56 VRL-58 VRL-61

N N N NH2 N N HO HO HO HO H

HO HO HO HO VRL-62 VRL-63 VRL-69 VRL-71

Annex II-D Cl

H H H N N N N HO HO N HO N

HO HO HO VRL-75 VRL-76 VRL-77

H N N N N N N HO H HO H HO

HO HO HO VRL-78 VRL-79 VRL-80

HO HO

N N

HO NH HO NH Cl

VRL-81 VRL-82

Annex II-D Annex II-D.b Antioxidant properties of N-derivatives of (3S,4S)- and (3R,4R)-pyrrolidine-3,4-diols.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

NA NA 0 ± 2 0 ± 1 VRL-2 26 ± 2 2 ± 1 ND 0 ± 1 ND ND 1 ± 1 ND ND ND - 0 0 0 0

NA NA 0 ± 2 0 ± 1 VRL-3 14 ± 5 1 ± 1 ND 0 ± 1 ND ND 0 ± 1 ND ND ND - 0 0 0 0

NA NA 0 ± 3 1 ± 1 VRL-14 18 ± 7 1 ± 1 ND 1 ± 2 ND ND 3 ± 0 ND ND ND - 0 0 0 0

NA NA 0 ± 3 0 ± 3 VRL-15 7 ± 14 1 ± 1 ND 0 ± 2 ND ND 2 ± 2 ND ND ND - 0 0 0 0

NA NA 2 ± 4 1 ± 1 VRL-21 7 ± 14 1 ± 1 ND 1 ± 3 ND ND 1 ± 1 ND ND ND - 0 0 0 0

NA 1 ± 3 1 ± 1 VRL-23 1 ± 3 ND 1 ± 1 ND 2 ± 2 ND ND 2 ± 1 ND ND ND - 0 0 0

NA 1 ± 4 0.65 ± 0.07 5 ± 1 NA 1 ± 1 VRL-31 2 ± 2 ND 3 ± 1 ND 1 ± 1 ND ND 13 ± 3 D6 0 0 1 0 0 0

NA NA NA 1 ± 4 ND 0.56 ± 0.05 1 ± 1 NA 1 ± 1 VRL-40 2 ± 8 3 ± 2 1 ± 3 ND 13 ± 5 D6 0 0 0 0 1 0 0 0

NA NA 3 ± 1 0.27 ± 0.03 7 ± 3 NA 2 ± 2 VRL-43 11 ± 3 2 ± 1 ND 2 ± 2 ND ND 17 ± 2 C6 0 0 0 2 0 0 0

NA 4.74 ± 0.07 NA 3 ± 1 0.25 ± 0.01 24 ± 3 0.72 ± 0.06 1 ± 1 VRL-45 1 ± 1 2 ± 1 2 ± 1 ND ND 31 ± 3 C6 0 1 0 0 2 0 1 0

NA NA 2 ± 1 1 ± 1 VRL-52 5 ± 9 2 ± 1 ND 1 ± 1 ND ND 1 ± 1 ND ND ND - 0 0 0 0

Annex II-D NA NA 3 ± 1 2 ± 1 VRL-55 1 ± 1 2 ± 2 ND 2 ± 1 ND ND 4 ± 1 ND ND ND - 0 0 0 0

NA NA 1 ± 1 ND 2 ± 1 VRL-56 5 ± 5 1 ± 1 ND 1 ± 1 ND ND 2 ± 1 ND ND - 0 0 0 0

NA NA 1 ± 1 0.50 ± 0.05 7 ± 1 NA 1 ±1 VRL-58 4 ± 4 3 ± 1 ND 1 ± 1 ND ND 12 ± 1 D6 0 0 0 1 0 2 0

NA NA 3 ± 1 1 ± 1 VRL-61 3 ± 3 1 ± 1 ND 2 ± 1 ND ND 2 ±12 ND ND ND - 0 0 0 0

NA ND NA 4 ± 1 1 ± 1 VRL-62 2 ± 2 1 ± 1 3 ± 2 ND ND 3 ± 1 ND ND ND - 0 0 0 0

NA NA 2 ± 1 2 ± 1 VRL-63 6 ± 6 1 ± 1 ND 2 ± 1 ND ND 4 ± 1 ND ND ND - 0 0 0 0

NA NA 2 ± 1 0.59 ± 0.03 5 ± 1 NA 1 ± 2 VRL-69 5 ± 5 2 ± 1 ND 1 ± 1 ND ND 18 ± 1 D6 0 0 0 1 0 0 0

NA ND NA 1 ± 1 0.26 ± 0.06 7 ± 1 NA 1 ± 1 VRL-71 1 ± 1 2 ± 1 1 ± 1 ND ND 22 ± 1 C6 0 0 0 2 0 0 0

NA ND NA 1 ± 1 0.27 ± 0.01 20 ± 1 NA 1 ± 2 VRL-75 5 ± 5 3 ± 1 1 ± 1 ND ND 29 ± 1 C6 0 0 0 2 0 0 0

NA ND NA 2 ± 3 0.28 ± 0.01 20 ± 1 NA 1 ± 1 VRL-76 6 ± 7 3 ± 2 1 ± 2 ND ND 28 ± 1 C6 0 0 0 2 0 0 0

NA ND NA 1 ± 3 0.28 ± 0.01 21 ± 1 NA 1 ± 1 VRL-77 4 ± 6 3 ± 1 1 ± 1 ND ND 27 ± 1 C6 0 0 0 2 0 0 0

NA ND NA 2 ± 2 0.23 ± 0.04 15 ± 2 NA 1 ± 2 VRL-78 9 ± 9 2 ± 1 1 ± 1 ND ND 27 ± 1 C6 0 0 0 2 0 0 0

NA ND NA 1 ± 1 0.31 ± 0.02 4 ± 1 NA 1 ± 1 VRL-79 8 ± 6 2 ± 1 1 ± 1 ND ND 15 ± 1 C6 0 0 0 2 0 0 0

NA ND NA 1 ± 1 0.23 ± 0.01 16 ± 1 NA 1 ± 1 VRL-80 5 ± 5 2 ± 1 1 ± 1 ND ND 28 ± 1 C6 0 0 0 2 0 0 0

NA ND NA 3 ± 3 0.21 ± 0.01 16 ± 2 NA 1 ± 1 VRL-81 2 ± 2 2 ± 1 1 ± 1 ND ND 26 ± 1 C6 0 0 0 2 0 0 0

Annex II-D NA ND NA 3 ± 3 0.25 ± 0.01 18 ± 2 0.88 ± 0.08 1 ± 1 VRL-82 5 ± 1 5 ± 2 1 ± 1 ND ND 29 ± 1 C6 0 0 0 2 0 1 0

Annex II-D Annex II-E Racemic conduramine B-1 analogs.

Annex II-E.a Structures of racemic conduramine B-1 analogs.

HO OH HO OH HO OH HO OH HO NH HO NH

HO NH2 HO NH2 O

ROBI-16 ROBI-18 ROBI-22 ROBI-23

HO OH HO OH HO OH Cl OH HO NH HO NH HO NH

ROBI-24 ROBI-25 ROBI-26

HO OH HO OH F OH H HO OH HO N HO NH S HO N HO NH OH HO F ROBI-27 ROBI-28 ROBI-31 ROBI-32

OH OH OH OH HO OH H H HO N HO N O O N OH HO NH H OH Br HO HO O ROBI-33 ROBI-34 ROBI-36 ROBI-51

Annex II-E O OH OH HN OH H H H HO N HO N HO N O

HO HO HO ROBI-56 ROBI-163 ROBI-207

O O O OH OH OH H H H HO N HO N HO N O Br N

HO HO HO ROBI-208 ROBI-214 ROBI-225

O OH H HO N N

N HO ROBI-237

Annex II-E Annex II-E.b Table racemic conduramine B-1 analogs.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

1 ± 1 1 ± 1 ROBI-16 7 ± 7 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

NA 4 ± 2 1 ± 1 ROBI-18 12 ± 7 1 ± 1 ND 2 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0 0

NA 3 ± 1 1 ± 1 ROBI-22 10 ± 7 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0 0

4 ± 2 1 ± 1 ROBI-23 7 ± 7 ND 1 ± 1 ND 2 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

4 ± 1 1 ± 1 ROBI-24 7 ± 6 ND 1 ± 1 ND 3 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

4 ± 1 1 ± 1 ROBI-25 1 ± 3 ND 1 ± 1 ND 2 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

5.32 ± 0.04 5.16 ± 0.01 5 ± 1 0.44 ± 0.01 14 ± 2 NA 3 ± 3 ROBI-26 65 ± 2 53 ± 5 3 ± 2 ND ND ND 28 ± 1 C5 2 2 0 2 0 0 0

5.51 ± 0.18 5.43 ± 0.02 4 ± 1 0.15 ± 0.02 37 ± 1 NA 1 ± 1 ROBI-27 70 ± 5 68 ± 5 4 ± 1 ND ND ND 50 ± 1 C5 3 2 0 3 1 0 0

NA NA 3 ± 1 1 ± 1 ROBI-28 17 ± 7 1 ± 1 ND 1 ± 1 ND ND 1 ± 2 ND ND ND - 0 0 0 0

NA 3 ± 1 0.63 ± 0.06 3 ± 1 NA 1 ± 1 ROBI-31 11 ± 5 ND 4 ± 1 ND 1 ± 1 ND ND 10 ± 1 D6 0 0 1 0 0 0

NA 2 ± 1 1 ± 1 ROBI-32 4 ± 4 ND 1 ± 1 ND 1 ± 1 ND ND 1 ± 1 ND ND ND - 0 0 0

Annex II-E 5.36 ± 0.06 5.21 ± 0.08 NA 5 ± 3 NA 24 ± 1 ROBI-33 93 ± 8 73 ± 3 8 ± 1 ND ND 34 ± 1 ND ND E4 2 2 0 0 0 0

NA NA 1 ± 1 1 ± 1 ROBI-34 14 ± 3 1 ± 1 ND 1 ± 1 ND ND 1 ± 1 ND ND ND - 0 0 0 0

NA NA 4 ± 5 1 ± 1 ROBI-36 15 ± 9 1 ± 1 ND 3 ± 3 ND ND 1 ± 1 ND ND ND - 0 0 0 0

NA 4 ± 4 1 ± 1 ROBI-51 5 ± 5 ND 1 ± 1 ND 2 ± 2 ND ND 1 ± 1 ND ND ND - 0 0 0

ND NA 2 ± 2 1 ± 1 ROBI-56 10 ± 4 ND 2 ± 1 2 ± 3 ND ND 1 ± 1 ND ND ND - 0 0 0

5.07 ± 0.19 4.74 ± 0.05 NA 4 ± 4 0.90 ± 0.09 6 ± 1 NA 1 ± 1 ROBI-163 61 ± 6 20 ± 3 4 ± 4 ND ND 10 ± 1 D7 2 1 0 0 1 0 0 0

NA 3 ± 3 1 ± 1 ROBI-207 7 ± 7 ND 1 ± 1 ND 5 ± 3 ND ND 1 ± 1 ND ND ND - 0 0 0

ND NA 3 ± 1 ND 1 ± 1 ROBI-208 8 ± 2 ND 1 ± 1 3 ± 2 ND ND 1 ± 1 ND ND - 0 0 0

NA ND NA 4 ± 2 ND 1 ± 2 ROBI-214 36 ± 10 1 ± 1 4 ± 1 ND ND 1 ± 1 ND ND - 0 0 0 0

NA ND NA 1 ± 4 ND 1 ± 2 ROBI-225 31 ± 10 2 ± 1 1 ± 2 ND ND 1 ± 1 ND ND - 0 0 0 0

ND NA 1 ± 1 ND 1 ± 1 ROBI-237 6 ± 5 ND 2 ± 1 1 ± 1 ND ND 1 ± 1 ND ND - 0 0 0

Annex II-E Annex II-F Secondary amines.

Annex II-F.a Structures of secondary amines.

O O

N N

H O N H N N H N N H H O ROBI-231 ROBI-232 ROBI-236

Annex II-F Annex II-F.b Antioxidant properties of secondary amines.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

NA 1 ± 1 3 ± 2 ROBI-231 12 ± 7 ND 3 ± 1 ND 1 ± 1 ND ND 4 ± 2 ND ND ND - 0 0 0

NA 1 ± 1 1 ± 1 ROBI-232 17 ± 4 4 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0 0

1 ± 1 1 ± 1 ROBI-236 10 ± 4 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

Annex II-F Annex II-G Ester of (3s,4s)-pyrrolidine-3,4-diols.

Annex II-G.a Structures of ester of (3s,4s)-pyrrolidine-3,4-diols.

O O O O Br Br N N O O

O O HO HO

ROBI-316 ROBI-318

Annex II-G Annex II-G.b Antioxidant properties of ester of (3s,4s)-pyrrolidine-3,4-diols.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

NA 1 ± 1 1 ± 2 ROBI-316 3 ± 3 ND 1 ± 1 ND 1 ± 1 ND ND 1 ± 2 ND ND ND - 0 0 0

1 ± 1 1 ± 1 ROBI-318 2 ± 4 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

Annex II-G Annex II-H Racemic conduramine F-1 analogs.

Annex II-H.a Structures of racemic conduramine F-1 analogs.

OH OH OH N OH AcN H H H HO NH2 HO N HO N HO N O

F HO HO HO HO ROBI-39 ROBI-202 ROBI-290 ROBI-329

F3C

N OH Cl OH OH OH O H H H HO N HO N HO N

HO HO HO ROBI-331 ROBI-333 ROBI-335

OH H OH HO N H HO N O HO

HO O ROBI-280 ROBI-356

Annex II-H Annex II-H.b Antioxidant properties of racemic conduramine F-1 analogs.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

NA 1 ± 1 1 ± 1 ROBI-39 2 ± 2 ND 1 ± 1 ND 1 ± 1 ND ND 1 ± 2 ND ND ND - 0 0 0

2 ± 2 1 ± 1 ROBI-202 1 ± 1 ND 1 ± 1 ND 2 ± 2 ND ND ND 1 ± 1 ND ND ND - 0 0

1 ± 1 1 ± 1 ROBI-290 6 ± 1 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

4.65 ± 0.18 1 ± 1 2 ± 1 ROBI-329 29 ± 4 8 ± 2 ND 1 ± 1 ND ND ND 2 ± 1 ND ND ND E3 1 0 0

1 ± 1 14 ± 1 ROBI-331 5 ± 4 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

5.46 ± 0.11 5.34 ± 0.09 1 ± 1 NA 1 ± 1 1 ± 1 NA ROBI-333 93 ± 12 66 ± 2 1 ± 1 ND ND ND 24 ± 1 E4 2 2 0 0 0 0 0

5.11 ± 0.11 NA 1 ± 1 1 ± 1 ROBI-335 55 ± 7 11 ± 2 1 ± 1 ND ND ND 1 ± 1 ND ND ND E3 2 0 0 0

NA 1 ± 1 1 ± 1 ROBI-280 1 ± 1 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0 0

NA 1 ± 1 1 ± 1 ROBI-356 3 ± 3 6 ± 5 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0 0

Annex II-H Annex II-I Racemic conduramine F-1 epoxides.

Annex II-I.a Structures of racemic conduramine F-1 epoxides.

O Cl OH OH OH H H H HO N HO N HO N O Cl OH HO HO HO O O O ROBI-459 ROBI-466 ROBI-470 F F OH OH OH OH F H H H HO N HO N HO N

HO HO HO O O O ROBI-472 ROBI-475 ROBI-477

O OH OH OH OH H H H HO N HO N HO N O

HO HO HO O O O ROBI-482 ROBI-484 ROBI-486

Annex II-I Br

OH S OH H HO N HO NH2

HO HO O O

ROBI-491 ROBI-494

Annex II-I Annex II-I.b Antioxidant properties of racemic conduramine F-1 epoxides.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

Name • • •– Profile % DPPH % DPPH % ABTS % ABTS•– pEC50 pEC50 ER50 ER50 ER50 ER50 a b c d Screening Screening Screening scavengedd scavengedd Screening scavengedf scavengedf Class Class Class Class Class Class Class Class Class Class

1 ± 1 1 ± 1 ROBI-459 6 ± 5 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 2 ND ND ND - 0 0

1 ± 1 1 ± 1 ROBI-466 6 ± 4 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

5.22 ± 0.10 5.17 ± 0.12 NA 1 ± 1 NA 1 ± 1 NA 1 ± 1 ROBI-470 80 ± 8 81 ± 5 1 ± 1 ND ND 16 ± 1 E4 2 2 0 0 0 0 0 0

1 ± 1 1 ± 1 ROBI-472 6 ± 3 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

5.11 ± 0.13 4.79 ± 0.03 1 ± 1 0.89 ± 0.40 1 ± 1 NA 1 ± 1 ROBI-475 48 ± 11 38 ± 3 1 ± 1 ND ND ND 21 ± 1 D7 2 1 0 1 0 0 0

ND 1 ± 1 1 ± 1 ROBI-477 1 ± 1 ND 1 ± 1 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

NA 1 ± 1 1 ± 1 ROBI-482 14 ± 7 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0 0

NA 1 ± 1 1 ± 1 ROBI-484 15 ± 6 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0 0

5.35 ± 0.11 4.99 ± 0.04 1 ± 1 0.38 ± 0.08 14 ± 1 0.61 ± 0.06 1 ± 1 ROBI-486 70 ± 11 48 ± 5 1 ± 1 ND ND ND 27 ± 1 C7 2 1 0 2 0 1 0

1 ± 1 1 ± 1 ROBI-491 1 ± 1 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

1 ± 1 1 ± 1 ROBI-494 11 ± 5 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

Annex II-I Annex II-J 7-Azabicyclo[2.2.1]hept-2-ene- and indole-derivatives.

Annex II-J.a Structures of 7-Azabicyclo[2.2.1]hept-2-ene- and indole-derivatives.

H N H N H HO N OH

OH HO Cl OH

ROBI-549 ROBI-552 ROBI-559

Annex II-J Annex II-J.b Antioxidant properties of 7-Azabicyclo[2.2.1]hept-2-ene- and indole-derivatives.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

1 ± 1 1 ± 1 ROBI-549 9 ± 7 ND 2 ± 1 ND 2 ± 1 ND ND ND 1 ± 2 ND ND ND - 0 0

1 ± 1 1 ± 1 ROBI-552 8 ± 3 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

1 ± 1 1 ± 1 ROBI-559 1 ± 2 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

Annex II-J Annex II-K Phenylethanoid glycosides and neolignan isolated from Jacaranda caucana.

Annex II-K.a Structures of phenylethanoid glycosides and neolignan isolated from Jacaranda caucana.

HO OH O

O HO OH OOH HO O O

O O O O HO O

HO OH O OH OH O O HO

OH HO HO OH JCA7 OH OH OH Protocatechuic acid JCA33 JCA37

Annex II-K OH

HO HO OH HO O OH O HO O O O OH HO O O O O O O O HO OH O HO OH O HO HO HO OH O OH JCA42 JCA44

Annex II-K O

HO HO OH HO O OH O OH O O O O HO O O O O O O O HO OH HO OH O O

HO HO HO OH O OH JCA47 JCA54 HO

O HO O O HO O OH

O OH HO

OH JCB16

Annex II-K Annex II-K.b Antioxidant properties of phenylethanoid glycosides and neolignan isolated from Jacaranda caucana.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

JCA7 5.60 ± 0.11 5.45 ± 0.03 0.23 ± 0.03 13 ± 3 0.50 ± 0.05 4 ± 3 0.14 ± 0.01 38 ± 2 0.18 ± 0.01 30 ± 1 Protocatech 100 ± 8 96 ± 3 32 ± 5 45 ± 1 B1 3 2 2 0 1 0 3 1 3 1 uic acid

6.00 ± 0.17 5.76 ± 0.02 0.08 ± 0.01 91 ± 1 0.10 ± 0.01 12 ± 1 0.13 ± 0.01 66 ± 2 0.15 ± 0.01 49 ± 3 JCA33 100 ± 14 93 ± 2 93 ± 1 79 ± 1 A2 3 3 3 3 3 0 3 2 3 1

5.94 ± 0.19 5.59 ± 0.04 0.07 ± 0.01 90 ± 1 0.11 ± 0.02 16 ± 5 0.16 ± 0.01 57 ± 6 0.12 ± 0.01 55 ± 3 JCA37 100 ± 15 93 ± 3 91 ± 1 68 ± 1 A2 3 3 3 3 3 0 3 2 3 2

5.96 ± 0.05 5.03 ± 0.08 0.17 ± 0.02 40 ± 3 0.55 ± 0.01 5 ± 1 0.20 ± 0.01 37 ± 2 0.29 ± 0.01 13 ± 1 JCA42 100 ± 16 87 ± 3 55 ± 3 48 ± 1 B1 3 2 3 1 1 0 2 1 2 0

5.90 ± 0.02 5.74 ± 0.09 0.10 ± 0.01 89 ± 1 0.16 ± 0.02 8 ± 1 0.13 ± 0.01 59 ± 5 0.18 ± 0.02 40 ± 2 JCA44 100 ± 18 90 ± 4 93 ± 1 73 ± 1 A2 3 3 3 3 3 0 3 2 3 1

5.90 ± 0.06 5.78 ± 0.03 0.08 ± 0.01 90 ± 1 0.17 ± 0.01 11 ± 4 0.11 ± 0.01 70 ± 6 0.15 ± 0.01 44 ± 1 JCA47 100 ± 12 88 ± 4 92 ± 1 80 ± 2 A2 3 3 3 3 3 0 3 2 3 1

5.64 ± 0.17 5.32 ± 0.03 0.38 ± 0.15 9 ± 5 0.94 ± 0.01 2 ± 2 0.20 ± 0.01 40 ± 4 0.11 ± 0.01 21 ± 6 JCA54 100 ± 15 81 ± 2 34 ± 4 51 ± 1 B1 3 2 2 0 1 0 2 1 3 0

5.05 ± 0.12 NA 4 ± 2 0.57 ± 0.09 2 ± 2 0.79 ± 0.01 19 ± 5 0.41 ± 0.05 5 ± 1 JCB16 51 ± 3 ND ND 13 ± 3 25 ± 2 - 2 0 0 1 0 1 0 2 0

Annex II-K Annex II-L 1-substituted pyrazolopyrrolidizine-derivatives.

Annex II-L.a Structures of 1-substituted pyrazolopyrrolidizine-derivatives.

O O O

N Cl N F N N N N N N N

MC2147 MC2170 MC2171 O OH N Cl N Cl N N N N N N N

MC2386 MC2387 MC2412

Annex II-L Annex II-L.b Antioxidant properties of 1-substituted pyrazolopyrrolidizine-derivatives.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

2 ± 2 1 ± 1 MC2147 4 ± 4 ND 3 ± 1 ND 3 ± 3 ND ND ND 2 ± 2 ND ND ND - 0 0

2 ± 1 2 ± 2 MC2170 2 ± 2 ND 2 ± 1 ND 2 ± 1 ND ND ND 2 ± 1 ND ND ND - 0 0

1 ± 1 1 ± 1 MC2171 4 ± 4 ND 3 ± 1 ND 2 ± 2 ND ND ND 1 ± 1 ND ND ND - 0 0

4.83 ± 0.22 NA 2 ± 2 1 ± 1 MC2386 33 ± 5 35 ± 3 2 ± 2 ND ND ND 1 ± 1 ND ND ND E3 1 0 0 0

NA 6 ± 1 ND ND 2 ± 1 MC2387 14 ± 3 ND 18 ± 3 10 ± 1 ND ND 2 ± 1 ND NA - 0 0 0

2 ± 2 1 ± 1 MC2412 4 ± 4 ND 2 ± 1 ND 2 ± 2 ND ND ND 1 ± 1 ND ND ND - 0 0

Annex II-L Annex II-M 2-substituted pyrazolopyrrolidizine-derivatives.

Annex II-M.a Structures of 2-substituted pyrazolopyrrolidizine-derivatives.

N N N N N N O N O N O O

N Cl N F N N

MC2109 MC2117 MC2123 MC2363

N N N N

N OH N N O N N O OH Cl N Cl N Cl N Cl N

MC2372 MC2385 MC2388 MC2405

N OH N O N N N N N N N N N S O Cl N Cl N Cl N Cl N

MC2406 MC2410 MC2411 MC2415

Annex II-M O Cl N O N N Cl N N O N O

Cl N N N

MC2416 MC2417 MC2418

Annex II-M Annex II-M.b Antioxidant properties of 2-substituted pyrazolopyrrolidizine-derivatives.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

1 ± 1 1 ± 1 MC2109 4 ± 4 ND 4 ± 1 ND 2 ± 2 ND ND ND 2 ± 2 ND ND ND - 0 0

2 ± 2 1 ± 1 MC2117 5 ± 3 ND 3 ± 1 ND 3 ± 3 ND ND ND 2 ± 2 ND ND ND - 0 0

1 ± 1 1 ± 1 MC2123 3 ± 2 ND 3 ± 1 ND 2 ± 2 ND ND ND 1 ± 1 ND ND ND - 0 0

4 ± 1 1 ± 1 MC2363 2 ± 2 ND 2 ± 1 ND 4 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

5.08 ± 0.03 4.67 ± 0.11 3 ± 1 ND ND 1 ± 1 MC2372 29 ± 4 33 ± 2 3 ± 1 ND ND 1 ± 1 ND NA E3 2 1 0 0

5.41 ± 0.13 4.74 ± 0.05 NA 13 ± 3 3 ± 2 MC2385 47 ± 2 40 ± 6 15 ± 7 ND ND 5 ± 1 ND ND ND E3 2 1 0 0 0

5.33 ± 0.09 4.67 ± 0.01 4 ± 1 1 ± 1 MC2388 51 ± 5 31 ± 1 5 ± 1 ND ND ND 1 ± 1 ND ND ND E3 2 1 0 0

5.63 ± 0.20 4.60 ± 0.13 3 ± 3 1 ± 1 MC2405 57 ± 10 42 ± 2 4 ± 4 ND ND ND 3 ± 1 ND ND ND E4 3 1 0 0

5.73 ± 0.10 NA 2 ± 2 3 ± 2 MC2406 46 ± 12 41 ± 2 3 ± 3 ND ND ND 3 ± 1 ND ND ND E3 3 0 0 0

NA 4 ± 1 1 ± 1 MC2410 10 ± 6 ND 11 ± 1 4 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0 0

4 ± 1 1 ± 1 MC2411 3 ± 2 ND 6 ± 2 ND 4 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

Annex II-M NA 1 ± 1 4 ± 1 MC2415 10 ± 2 ND 10 ± 2 ND 11 ± 1 ND ND 8 ± 1 ND ND ND - 0 0 0

5.14 ± 0.05 NA 2 ± 1 3 ± 1 MC2416 22 ± 2 13 ± 4 4 ± 1 ND ND ND 4 ± 1 ND ND ND E3 2 0 0 0

3 ± 1 1 ± 1 MC2417 4 ± 4 ND 2 ± 1 ND 3 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

2 ± 1 1 ± 1 MC2418 10 ± 4 ND 3 ± 1 ND 3 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

Annex II-M Annex II-N Pyrrol- and 2-methylquinoline-3-carboxylate-derivatives.

Annex II-N.a Structures of pyrrol- and 2-methylquinoline-3-carboxylate-derivatives.

O O O O O

N OH O O O N O N N H H

N N MC1234 MC2102 MC2103 O O H H O N O N O O O N N MC2104 MC2105

Annex II-N Annex II-N.b Antioxidant properties of pyrrol- and 2-methylquinoline-3-carboxylate-derivatives.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

NA 1 ± 1 1 ± 1 MC1234 11 ± 7 ND 11 ± 2 2 ± 2 ND ND ND 2 ± 2 ND ND ND - 0 0 0

1 ± 1 1 ± 1 MC2102 8 ± 7 ND 1 ± 1 ND 2 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

1 ± 1 1 ± 1 MC2103 8 ± 1 ND 2 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

NA 1 ± 1 1 ± 1 MC2104 20 ± 10 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0 0

NA 1 ± 1 1 ± 1 MC2105 23 ± 7 1 ± 1 ND 3 ± 1 ND ND ND 1 ± 1 ND ND NA - 0 0 0

Annex II-N Annex II-O Xanthone-derivatives.

Annex II-O.a Structures of xanthone-derivatives.

O O OH O O

O O O O OH Xanthone X25 X13 X14

O O O OCH3 O

OCH3 O

OH O O O OCH3 X24 X26 X10 X11

O O O O

O O OH O OCH3 O OCH3

OCH3 OCH3 OH OCH3 X23 X8 X9 X7

Annex II-O OH O OH O O OH O OH

OCH3

O OH O OCH O OCH 3 O OCH3 3 OH OCH3 OH OCH3 ID39 X6 ID22 ID23 Desmethylbellodifolin

OH O OH OH O OH

O OCH3 O OH OH ID40 ID42 Bellidifolin Norswertianin

Annex II-O Annex II-O.b Antioxidant properties of xanthone-derivatives.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

1 ± 1 1 ± 1 xanthone 1 ± 1 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

NA 1 ± 1 1 ± 1 X25 32 ± 8 10 ± 3 ND 1 ± 1 ND ND ND 2 ± 2 ND ND ND - 0 0 0

5.55 ± 0.06 5.52 ± 0.10 1 ± 1 0.13 ± 0.02 20 ± 1 0.76 ± 0.01 2 ± 2 X13 100 ± 5 86 ± 3 1 ± 1 ND ND ND 34 ± 1 C5 3 3 0 3 0 1 0

1 ± 1 1 ± 1 X14 ND ND 5 ± 2 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

5.42 ± 0.06 5.26 ± 0.03 1 ± 1 0.34 ± 0.13 7 ± 1 NA 1 ± 1 X24 96 ± 8 84 ± 4 1 ± 1 ND ND ND 20 ± 1 C5 2 2 0 2 0 0 0

1 ± 1 1 ± 1 X26 12 ± 6 ND 1 ± 0 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

1 ± 1 1 ± 1 X10 1 ± 1 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

NA 1 ± 1 1 ± 1 X11 12 ± 9 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0 0

NA 1 ± 1 1 ± 1 X23 23 ± 3 ND ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0 0

5.42 ± 0.05 5.19 ± 0.10 1 ± 1 1 ± 1 X8 52 ± 15 58 ± 7 1 ± 1 ND ND ND 1 ± 1 ND ND ND E4 2 2 0 0

5.30 ± 0.14 5.33 ± 0.09 3 ± 1 0.20 ± 0.03 24 ± 1 0.75 ± 0.09 4 ± 2 X9 99 ± 5 80 ± 4 6 ± 1 ND ND ND 31 ± 1 C5 2 2 0 2 0 1 0

Annex II-O 1 ± 1 1 ± 1 X7 2 ± 2 ND 1 ± 1 ND 1 ± 1 ND ND ND 1 ± 1 ND ND ND - 0 0

5 ± 2 6 ± 4 X6 1 ± 1 ND 1 ± 1 ND 7 ± 2 ND ND ND 6 ± 4 ND ND ND - 0 0

5.69 ± 0.04 5.59 ± 0.04 1 ± 1 0.27 ± 0.07 6 ± 1 NA 1 ± 1 ID22 100 ± 9 80 ± 6 1 ± 1 ND ND ND 18 ± 1 C5 3 3 0 2 0 0 0

5.41 ± 0.10 4.84 ± 0.23 1 ± 1 NA 1 ± 1 ID23 96 ± 9 27 ± 1 1 ± 1 ND ND ND 1 ± 1 ND ND E3 2 1 0 0 0

5.37 ± 0.06 5.20 ± 0.07 0.20 ± 0.01 80 ± 3 0.36 ± 0.02 7 ± 1 0.21 ± 0.02 43 ± 1 0.34 ± 0.04 26 ± 2 ID39 100 ± 5 77 ± 2 92 ± 3 41 ± 1 A1 2 2 2 3 2 0 2 1 2 1

5.29 ± 0.15 5.24 ± 0.19 0.20 ± 0.01 82 ± 1 0.46 ± 0.06 5 ± 1 0.24 ± 0.01 43 ± 3 0.42 ± 0.02 19 ± 1 ID40 70 ± 9 71 ± 5 92 ± 1 40 ± 1 A1 2 2 2 3 2 0 2 1 2 0

5.58 ± 0.01 5.42 ± 0.02 0.25 ± 0.03 45 ± 2 0.42 ± 0.05 7 ± 1 0.22 ± 0.01 29 ± 1 0.30 ± 0.01 21 ± 1 ID42 100 ± 4 78 ± 4 66 ± 4 34 ± 1 B1 3 2 2 1 2 0 2 1 2 0

X14 was fluorescent in the ALP assay.

Annex II-O Annex II-P Mangiferin-derivatives extracted from the Leaves of Arrabidaea patellifera.

Annex II-P.a Structures of mangiferin-derivatives.

HO O OH O HO HO OH HO O OH O O OH O OH OH HO HO HO HO O OH O HO HO O HO O OH O HO O OH O O O

OH OH APL58 APL70 APL98

Annex II-P HO O OH O HO OH HO HO HO O OH O HO O OH O OH O HO

HO O O HO O OH O O OH O O OH HO

HO HO HO O OH OH OH APL101 APLL3 7418

Annex II-P Annex II-P.b Antioxidant properties of mangiferin-derivatives.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

6.05 ± 0.05 5.76 ± 0.06 0.11 ± 0.05 33 ± 2 0.25 ± 0.01 7 ± 2 0.14 ± 0.01 62 ± 2 0.13 ± 0.01 40 ± 1 APL58 100 ± 6 95 ± 3 71 ± 1 67 ± 1 B2 3 3 3 1 2 0 3 2 3 1

6.07 ± 0.01 5.70 ± 0.05 0.16 ± 0.01 25 ± 1 0.31 ± 0.05 1 ± 1 0.15 ± 0.02 56 ± 1 0.11 ± 0.01 38 ± 1 APL70 100 ± 8 95 ± 3 61 ± 1 61 ± 1 B2 3 3 3 0 2 0 3 2 3 1

5.84 ± 0.02 5.59 ± 0.02 0.23 ± 0.01 26 ± 1 0.40 ± 0.03 5 ± 1 0.18 ± 0.02 53 ± 2 0.17 ± 0.02 42 ± 3 APL98 100 ± 10 90 ± 3 65 ± 1 58 ± 1 B1 3 3 2 1 2 0 3 2 3 1

6.01 ± 0.01 5.81 ± 0.03 0.16 ± 0.01 30 ± 2 0.32 ± 0.04 1 ± 1 0.13 ± 0.01 61 ± 1 0.14 ± 0.02 40 ± 5 APL101 100 ± 14 91 ± 4 68 ± 1 67 ± 2 B2 3 3 3 1 2 0 3 2 3 1

5.95 ± 0.03 5.68 ± 0.04 0.10 ± 0.01 89 ± 1 0.24 ± 0.01 10 ± 2 0.11 ± 0.01 66 ± 7 0.15 ± 0.01 41 ± 1 APLL3 100 ± 12 90 ± 3 90 ± 1 69 ± 1 A2 3 3 3 3 2 0 3 2 3 1

5.64 ± 0.01 5.14 ± 0.02 0.35 ± 0.06 20 ± 4 0.29 ± 0.11 31 ± 1 7418 100 ± 6 55 ± 3 46 ± 1 ND ND 34 ± 1 ND ND B1 3 2 2 0 2 1

Annex II-P Annex II-Q Quercetin-derivatives extracted from Flavo Psidium Cattleianium.

Annex II-Q.a Structures of quercetin-derivatives.

OH OH

HO OH OH OH OH

HO O HO O HO O HO O OH

OH OH OH

OH O OH O OH O OH O

RHC1 RHC2 RHC3 RHC4 Morin Myricetin Luteolin Kaempferol OH OH OH OH

HO O HO O O OH O O OH O OH OH O OH O OH HO OH RHB4 RHB5

Annex II-Q OH

HO O O

O OH

OH O HO OH

RHB6

Annex II-Q Annex II-Q.b Antioxidant properties of quercetin-derivatives.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

RHC1 5.84 ± 0.06 5.86 ± 0.06 0.16 ± 0.01 95 ± 2 0.13 ± 0.02 74 ± 12 0.12 ± 0.02 47 ± 8 0.15 ± 0.01 42 ± 3 ND 82 ± 2 95 ± 1 47 ± 8 A2 Morin 3 3 3 3 3 2 3 1 3 3

RHC2 5.62 ± 0.10 5.61 ± 0.06 0.07 ± 0.01 40 ± 2 0.06 ± 0.02 14 ± 5 0.06 ± 0.01 63 ± 9 0.04 ± 0.01 73 ± 1 ND 79 ± 2 68 ± 1 85 ± 6 B2 Myricetin 3 3 3 1 3 0 3 2 3 2

RHC3 5.92 ± 0.10 6.09 ± 0.13 0.05 ± 0.01 77 ± 1 0.16 ± 0.05 9 ± 5 0.18 ± 0.02 33 ± 4 0.13 ± 0.01 38 ± 2 ND 80 ± 1 86 ± 1 34 ± 4 A2 Luteolin 3 3 3 3 3 0 3 1 3 1

RHC4 5.76 ± 0.05 5.84 ± 0.05 0.14 ± 0.01 57 ± 3 0.16 ± 0.04 11 ± 7 0.14 ± 0.01 45 ± 11 0.21 ± 0.02 39 ± 1 ND 78 ± 2 93 ± 1 45 ± 12 A2 Kaempferol 3 3 3 2 3 0 3 1 2 1

5.44 ± 0.05 5.37 ± 0.06 0.25 ± 0.03 30 ± 1 0.66 ± 0.01 1 ± 1 0.56 ± 0.12 11 ± 3 0.69 ± 0.02 12 ± 1 RHB4 ND ND 47 ± 3 ND B9 2 2 2 1 1 0 1 0 1 0

5.55 ± 0.01 5.55 ± 0.05 0.15 ± 0.01 49 ± 2 0.30 ± 0.02 7 ± 2 0.52 ± 0.05 12 ± 2 0.37 ± 0.01 14 ± 2 RHB5 ND ND 63 ± 3 ND B9 3 3 3 1 2 0 1 0 2 0

5.47 ± 0.05 0.22 ± 0.03 33 ± 1 0.83 ± 0.01 1 ± 1 0.61 ± 0.01 11 ± 2 NA NA RHB7 Inhibition Inhibition ND 48 ± 2 ND B9 2 2 1 1 0 1 0 0 0

Annex II-Q Annex II-R Trolox®-derivatives.

Annex II-R.a Structures of trolox®-derivatives.

O O N O H N HO H N N HO H COL-1 COL-22

O O

O OH O N N OH

N H HO O HO O COL-24 COL-30

O O O O O O OH N S O O N O N H O H H HO HO HO COL-50 COL-52 COL-62

Annex II-R Annex II-R.b Antioxidant properties of trolox®-derivatives.

ALP assay ORAC assay DPPH assay ABTS assay

EtOH EtOH / AcOH 1% EtOH EtOH / AcOH 1%

5.29 ± 0.11 5.29 ± 0.09 0.12 ± 0.02 46 ± 2 0.15 ± 0.02 43 ± 2 0.24 ± 0.02 36 ± 4 0.24 ± 0.02 38 ± 2 COL1 100 ± 1 72 ± 2 75 ± 1 35 ± 4 B1 2 2 3 1 3 1 2 1 2 1

5.50 ± 0.05g NA 0.18 ± 0.02 18 ± 5 0.21 ± 0.01 24 ± 2 0.34 ± 0.04 26 ± 1 0.36 ± 0.03 22 ± 2 COL22 34 ± 9 24 ± 2 50 ± 2 26 ± 1 B8 2 0 3 0 2 0 2 1 2 0

5.19 ± 0.10g 4.93 ± 0.02 0.13 ± 0.01 24 ± 3 0.20 ± 0.04 10 ± 1 0.25 ± 0.04 30 ± 2 0.27 ± 0.01 24 ± 2 COL24 63 ± 11 28 ± 2 61 ± 1 33 ± 2 B8 2 1 3 0 2 0 2 1 2 0

5.70 ± 0.15 5.43 ± 0.06 0.14 ± 0.02 32 ± 2 0.23 ± 0.01 35 ± 4 0.24 ± 0.06 38 ± 1 0.20 ± 0.01 43 ± 3 COL30 100 ± 9 74 ± 3 65 ± 1 42 ± 1 B1 3 2 3 1 2 1 2 1 2 1

4.87 ± 0.08 4.96 ± 0.22 0.16 ± 0.02 27 ± 4 0.23 ± 0.01 34 ± 2 0.25 ± 0.03 35 ± 1 0.28 ± 0.02 30 ± 5 COL50 73 ± 11 33 ± 3 67 ± 10 34 ± 1 B8 1 1 3 1 2 1 2 1 2 1

4.89 ± 0.06 4.82 ± 0.03 0.15 ± 0.03 29 ± 1 0.28 ± 0.02 29 ± 3 0.23 ± 0.03 37 ± 1 0.51 ± 0.01 19 ± 1 COL52 63 ± 1 22 ± 4 66 ± 1 36 ± 2 B8 1 1 3 1 2 1 2 1 1 0

4.91 ± 0.04 4.78 ± 0.03 0.16 ± 0.03 42 ± 1 0.32 ± 0.01 41 ± 3 0.34 ± 0.04 36 ± 2 0.22 ± 0.01 37 ± 3 COL62 28 ± 10 30 ± 2 76 ± 1 39 ± 1 B8 1 1 3 1 2 1 2 1 2 1

g: the compounds COL-22 and COL-24 did not protect the ALP protein over a percentage of 30% and 70%, respectively

Annex II-R

Annex III Publications

J. Nat. Prod. 2008, 71, 1887–1890 1887

Antioxidant C-Glucosylxanthones from the Leaves of Arrabidaea patellifera

Fre´de´ric Martin,† Anne-Emmanuelle Hay,† Delphine Cressend,‡ Marianne Reist,‡ Livia Vivas,§ Mahabir P. Gupta,⊥ Pierre-Alain Carrupt,‡ and Kurt Hostettmann*,† Laboratory of Pharmacognosy and Phytochemistry, School of Pharmaceutical Sciences, UniVersity of GeneVa, UniVersity of Lausanne, Quai Ernest-Ansermet 30, CH-1211 GeneVa 4, Switzerland, Unit of Pharmacochemistry, School of Pharmaceutical Sciences, UniVersity of GeneVa, UniVersity of Lausanne, Quai Ernest-Ansermet 30, CH-1211 GeneVa 4, Switzerland, London School of Hygiene and Tropical Medicine, Department of Infectious and Tropical Diseases, Keppel Street, London WC1E 7HT, U.K., and Center for Pharmacognostic Research on Panamanian Flora (CIFLORPAN), College of Pharmacy, UniVersity of Panama, Panama

ReceiVed July 7, 2008

Chemical investigation of the methanol extract from the leaves of Arrabidaea patellifera, a Bignoniaceae from Panama, afforded mangiferin, isomangiferin, and six new derivatives (3′-O-p-hydroxybenzoylmangiferin, 3′-O-trans-coumaroylmangiferin, 6′-O-trans-coumaroylmangiferin, 3′-O-trans-cinnamoylmangiferin, 3′-O-trans-caffeoylmangiferin, and 3′- O-benzoylmangiferin). All these compounds had antioxidant and radical-scavenging activities, and four of them were relatively active in Vitro against Plasmodium falciparum. The structures were determined by spectrometric and chemical methods, including 1D and 2D NMR experiments and MS analysis.

As part of our ongoing investigations on Panamanian Bignoni- aceae,1 16 extracts (DCM and MeOH) from six Panamanian plants of the family Bignoniaceae were submitted to a rapid TLC 1,1- diphenyl-2-picrylhydrazine (DPPH) test,2 which revealed radical- scavenging activity, and to an in Vitro test against Plasmodium falciparum. The Bignoniaceae family comprises about 120 genera and 800 species, growing mainly in Africa and Central and South America. Species of the Bignoniaceae are used for many purposes, such as horticulture, timber, dyes, and medicine. The best-known medicinal use of the Bignoniaceae is the application of bark preparations of various species of Tabebuia as cancer cures.3 the stems from A. samydoides by Pauletti et al.5 However, pairs of Members of the family have not been extensively chemically constituents with the same molecular formula were detected, thus investigated.4 The genus Arrabidaea belongs to the tribe Bignon- implying the presence of different metabolites from those found in ieae, a large and morphologically diverse clade of neotropical A. samydoides. In order to isolate some potentially new active lianas.5 The genus contains about 70 species, spread from Mexico compounds, separation, isolation, and structure elucidation were to Argentina. Previous phytochemical studies indicate that this genus carried out. is a source of C-glucosylxanthones, phenylpropanoids, flavonoids, - The extract was first rapidly separated by VLC (vacuum liquid anthocyanidins, allantoins, and triterpenes.6 9 Arrabidaea patel- chromatography) on reversed-phase to afford four fractions. The lifera (Schltdl.) Sandwith is a liana with a distribution from Mexico 80% MeOH-H O fraction contained the major UV-active metabo- to Brazil in tropical dry forest and less frequently in moist or wet 2 lites present in the MeOH extract. This fraction was then purified forest. This species seems to be botanically close to A. samydoides by MPLC to yield mangiferin (1). Its identity was confirmed by (Cham.) Sandwith.10 The MeOH extract from the leaves of A. HRMS and comparison of the 1H and 13C NMR spectra with patellifera was selected due to its good activities and moreover literature data.11 Compounds 2, 3, 4, and 5 were obtained in the was not investigated before. This paper describes the isolation and same manner. Compound 6 was isolated using Lobar chromatog- characterization of mangiferin (1), isomangiferin, and six new raphy and compound 7 by Sephadex LH-20 gel filtration. Iso- mangiferin derivates (2-7) and their associated activities. For pure mangiferin was purified by semipreparative LC and identified in compounds, antioxidant and radical-scavenging activities were the same way.12 tested in solution. Compound 2 was shown to have the molecular formula - 1 13 C26H22O13 ([M - H] , m/z 541.0985 by HRMS). The H and C Results and Discussion NMR spectra (Tables 1 and 2) showed proton and carbon signals An initial phytochemical analysis using UPLC/UV-HRMS-TOF equivalent to those of mangiferin: three aromatic singlets (at δ 6.31, was done to investigate the constituents of the MeOH extract from 6.75, and 7.37) and several hydroxymethine protons, indicating the presence of a sugar moiety. The MS and MS2 data in the APCI the leaves of A. patellifera. The UV spectrum (λmax around 240, 260, 320, and 365 nm, see Experimental Section) suggested the positive mode showed a fragment at m/z 423 corresponding to a presence of mangiferin derivatives. Moreover, HRMS of the mangiferin moiety and another fragment at m/z 405 characteristic 6 constituents were close to those isolated from the EtOH extract of for C-glucosylxanthones. All these elements indicated that compound 2 was mangiferin with an additional C7H4O2 residue. 1 * To whom correspondence should be addressed. Tel: +41 22 379 34 Additional aromatic proton signals could be observed in the H 00. Fax: +41 22 379 33 99. E-mail: [email protected]. NMR spectrum at δ 7.97 (2H, d, J ) 8.9 Hz) and 6.83 (2H, d, J † Laboratory of Pharmacognosy and Phytochemistry, University of ) 8.9 Hz) together with 13C NMR signals showing the presence of Geneva, University of Lausanne. an ester carbonyl carbon at δ 168.4, a phenolic carbon at δ 163.5, ‡ Unit of Pharmacochemistry, University of Geneva, University of Lausanne. and two aromatic carbons at δ 116.2 and 133.2. These elements § London School of Hygiene and Tropical Medicine. indicated that compound 2 was esterified by a p-hydroxybenzoyl ⊥ University of Panama. group. A similar compound has been isolated from Senecio

10.1021/np800406q CCC: $40.75  2008 American Chemical Society and American Society of Pharmacognosy Published on Web 10/25/2008 1888 Journal of Natural Products, 2008, Vol. 71, No. 11 Martin et al.

Table 2. 13C NMR Data of Compounds 1-7 (125 MHz, in CD3OD) 13C mangiferin (1)a 234567 1 161.6 163.5 163.5 163.5 163.6 163.4 163.7 12.0, 2.0 Hz) 7.3 Hz) 7.3 Hz) 9.3 Hz) 7.3 Hz) 7.8 Hz) 7.3 Hz)

10.2 Hz) 2 107.4 107.5 107.5 107.4 107.5 107.4 107.6 ) ) ) ) ) ) )

) 3 163.7 165.3 165.3 165.3 165.3 165.3 165.5 4 93.2 94.9 94.9 95.0 94.9 94.8 94.9 4a 156.1 158.9 158.9 158.9 158.9 158.8 158.8 6.37 s 6.81 s 7.45 s 5.06 ( J 5.31 t ( J 3.79 m 3.56 m 3.90 d ( J 3.77 m 7.60 t ( J 7.49 t ( J 5 102.5 103.6 103.6 103.6 103.6 103.4 103.6 6 156.7 155.4 155.4 155.3 155.4 156.5 156.3 7 143.7 144.9 145.0 144.9 145.0 145.2 145.7 8 107.9 109.3 109.3 109.3 109.3 108.8 109.0 8a 111.5 113.9 113.9 114.0 114.0 113.3 113.7 9 178.9 181.3 181.4 181.4 181.4 181.2 181.4 12.7 Hz) 12.0, 5.6 Hz) 8.3, 1.5 Hz) 8.11 d ( J 9a 101.1 103.3 103.3 103.4 103.3 103.2 103.3 1.0 Hz) 8.11 d ( J 8.3 Hz) 7.60 t ( J 15.6 Hz) 15.6 Hz) 9.3 Hz) 9.8, 8.8 Hz) 4.55 m ) ) ) 9.8 Hz) ) ) ) ) 10a 154.1 153.1 153.2 153.2 153.2 153.4 153.1 ) )

) 1′ 73.0 75.4 75.4 75.6 75.4 75.5 75.5 2′ 70.2 70.9 70.9 72.7 70.8 70.8 70.8 3′ 78.8 81.6 81.4 80.0 81.7 81.5 82.0 6.35 s 6.77 s 7.42 s 5.15 t ( J 4.46 t ( J 3.70 m 3.54 m 3.90 brd ( J 3.77 dd ( J 7.06 d ( J 6.78 d ( J 6.96 dd ( J 4′ 70.5 70.4 70.4 72.0 70.3 70.3 70.3 5′ 81.4 82.7 82.7 80.1 82.7 82.7 82.8 6′ 61.4 62.9 62.9 65.2 62.8 62.8 62.9 1′′ 122.9 127.5 127.3 136.1 127.9 132.1 2′′ 133.2 131.2 131.3 129.3 115.1 130.9 3′′ 116.2 116.9 116.9 130.1 149.5 129.6

12.2, 2.0 Hz) 12.2, 5.4 Hz) ′′ 10.2 Hz 5.02 ( J 16.1 Hz)16.1 Hz) 7.61 d ( J 7.36 d ( J 4 163.5 161.2 161.3 131.5 146.8 134.2 9.3 Hz 9.5 Hz) ) ) ′′ ) ) ) 5 116.2 116.9 116.9 130.1 116.6 129.6 ) ) 6′′ 133.2 131.2 131.3 129.3 123.0 130.9 7′′ 168.4 146.6 146.9 146.3 146.9 168.3 8′′ 115.9 115.1 119.5 115.7 6.34 s 6.79 s 7.42 s 5.18 t ( J 3.73 m 3.54 m 3.91 dd ( J 3.77 dd ( J 7.40 m 9′′ 169.5 169.4 168.8 169.4 a In DMSO-d6.

mikanioides,13 where the p-hydroxybenzoyl group was linked to C-2′. However, in 2, the esteriﬁcation pattern seemed to be different.

12.2, 2.0 Hz) 12.0, 5.6 Hz) 1 -1 10.2 Hz) 5.03 d ( J 8.3 Hz) 7.61 m 8.3 Hz) 7.40 m 8.3 Hz) 7.40 m 8.3 Hz) 7.61 m 15.9 Hz)15.9 Hz) 7.75 d ( J 6.62 d ( J The 2D COSY H H short-range correlation spectrum permitted 9.3, 8.8 Hz) 4.48 t ( J ) ) ) ) ) ) ) ) )

) assignment of the hydroxymethine protons. Starting from the anomeric H-1′ at δ 5.05 (1H, d, J ) 10.1 Hz), the 2D correlation spectrum showed that H-2′ had a chemical shift at δ 4.53 (1H, t, J 6.32 s 6.76 s 7.40 s 3.53 m 4.26 t ( J 3.53 m 3.69 m 4.55 dd ( J 4.37 dd ( J ) 9.6 Hz) and H-3′ at δ 5.27 (1H, t, J ) 9.2 Hz). The long-range 1H-13C HMBC spectrum showed correlation between C-7′′ (δ 168.4) and H-3′. These elements indicated that in 2 the p- hydroxybenzoyl group was linked via an ester bond at C-3′.It OD) 3 explained the deshielded chemical shifts corresponding to position ′ ′

12.2, 2.0 Hz) C-3 (δ 81.6) and H-3 (δ 5.27) compared to those for mangiferin. 9.8 Hz) 4.95 d ( J 8.8 Hz) 7.42 d ( J 8.8 Hz) 6.78 d ( J 8.8 Hz) 6.78 d ( J 8.8 Hz) 7.42 d ( J 16.0 Hz)16.0 Hz) 7.62 d ( J 6.35 d ( J 9.3 Hz) 9.8 Hz)

) Compound 2 is thus a new natural product, 3′-O-p-hydroxyben- ) ) ) ) ) ) ) ) ) zoylmangiferin. The molecular formula C28H24O13 could be assigned by HRMS - - 1 13 6.32 s 6.77 s 7.40 s 5.17 t ( J 4.47 t ( J 3.72 m 3.55 m 3.92 dd ( J 3.79 m 7.68 d ( J 6.43 d ( J to compound 3 ([M H] , m/z 567.1122). The H and C NMR

(500 MHz, in CD spectra (Tables 1 and 2) indicated that this compound had a mangiferin skeleton with an additional C9H6O2 residue. In contrast 1 - 7 to 2, deshielding of the aromatic proton signals at δ 6.80 (2H, d,J ) 8.8 Hz) and 7.46 (2H, d, J ) 8.8 Hz) together with the presence of two trans olefinic protons at δ 7.68 (1H, d, J ) 16.0 Hz) and 12.2, 2.1 Hz) 234567 10.1 Hz) 5.03 d ( J 8.9 Hz) 7.46 d ( J 8.9 Hz) 6.80 d ( J 8.9 Hz) 6.80 d ( J 8.9 Hz) 7.46 d ( J ) 9.2 Hz) 9.6 Hz) 6.43 (1H, d, J 16.0 Hz) could be observed. In the same way, the ) ) ) ) ) ) ) ) presence of two olefinic carbons at δ 146.6 and 115.9 (C7′′ and C8′′) was suggested by comparison with the 13C NMR spectrum of 2. These elements indicated the presence of a p-coumaroyl 6.31 s 6.75 s 7.37 s 5.27 t ( J 4.53 t ( J 3.78 m 3.58 m 3.93 dd ( J 3.79 m 7.97 d ( J 6.83 d ( J 6.83 d ( J 7.97 d ( J moiety. Similarities between the chemical shifts of the sugar moieties in 2 and 3, together with a long-range HMBC correlation observed between C-9′′ and H-3′, led to the identification of 3 as

a 3′-O-trans-coumaroylmangiferin, a new compound. Compound 4 had the same molecular formula (C28H24O13)as3 - 11.7, 5.8 Hz) ([M - H] , m/z 567.1118 by HRMS). This suggested that 4 was a 9.9 Hz) 5.05 d ( J 10.7 Hz) 9.1 Hz)

) 1 13 ) ) regioisomer of 3. Furthermore, the H and C NMR spectra (Tables ) . 6 1 and 2) were similar to those of 3, especially for the xanthone mangiferin ( 1 ) and coumaroyl signals. The only noticeable differences were the H NMR Spectroscopic Data of Compounds 1 4.61 d ( J 3.20 m 4.05 t ( J 3.14 m 3.16 m 3.69 d ( J 3.42 dd ( J downﬁeld chemical shifts of H-6′ (1H, δ 4.37 and 4.55, dd each) and C-6′ (δ 65.2) and the upﬁeld chemical shifts of H-3′ (1H, δ

In DMSO- d ′ H a 3.53, m) and C-3 (δ 80.0). These indicated that the coumaroyl 1 4 6.37 s 5 6.86 s 8 7.38 s 1 ′ 3 ′ 2 ′ 4 ′ 5 ′ 6 ′ 2 ′′ 3 ′′ 4 ′′ 5 ′′ 6 ′′ 7 ′′ 8 ′′

Table 1. moiety was not linked via an ester bond at C-3′, but at C-6′. This C-Glucosylxanthones from Arrabidaea patellifera Journal of Natural Products, 2008, Vol. 71, No. 11 1889

Table 3. Activities of Pure Compounds on P. falciparum 3D7 Table 4. Activities of Compounds in DPPH and ALP Assays 3 ( H-Hypoxanthine Assay) a b compound ER50 (DPPH) EC50 (ALP) (µM) compound IC50 (µM) mangiferin (1) 0.12 ( 0.06 1.26 ( 0.10 mangiferin (1) 23.8 2 0.11 ( 0.05 0.98 ( 0.07 2 26.5 3 0.16 ( 0.01 0.83 ( 0.04 3 18.1 4 0.23 ( 0.01 1.48 ( 0.10 4 >38.2 5 0.16 ( 0.01 0.98 ( 0.05 chloroquine 4.70 × 10-3 6 0.10 ( 0.01 1.18 ( 0.08 7 0.35 ( 0.06 1.87 ( 0.72 quercetin (positive control) 0.092 ( 0.01 1.00 ( 0.07 linkage was corroborated by a long-range HMBC correlation a • ER50 is the ratio of antioxidant concentration to DPPH ′′ ′ b between C-9 and H-6 . Compound 4 is thus a new natural product, concentration producing a 50% decrease in DPPH at steady state. EC50 6′-O-trans-coumaroylmangiferin. It is the only compound isolated is the antioxidant concentration that protects ALP to 50% from peroxyl from A. patellifera esterified at C-6′, which may explain the sign radical-induced activity loss. difference in the optical rotation. It should be noted that 6′-O- - benzoylmangiferin has already been isolated from Senecio mikan- The radical-scavenging effect of compounds 1 6 with DPPH is iodes (Asteraceae)14 and showed the same downfield shifts for C-6′ shown in Table 4. All of the xanthones showed activity, with the 16 and H-6′. same order of magnitude as the positive control, quercetin. The alkaline phosphatase test (ALP) is a simple fluorimetric test17 Direct comparison of 1H NMR data for compounds 5 and 3 to assess the antioxidant capacity of chemical entities to protect showed that the two aromatic doublets at δ 6.80 and 7.46 for the proteins from loss of activity caused by reactive oxygen species coumaroyl moiety were replaced by two multiplets at δ 7.61 and (ROS). All of the tested compounds show activity in this assay 7.40. In contrast, the two olefinic protons in the trans arrangement (Table 4). Compounds 2, 3, and 5 had similar activities to the at δ 7.75 (H-7′′,d,J ) 16.1 Hz) and 6.62 (H-8′′,d,J ) 16.1 Hz) positive control, quercetin. had similar chemical shifts to those observed for 3 and 4 (Tables 1 and 2), suggesting that the coumaroyl moiety in 3 and 4 was Experimental Section replaced in 5 by a cinnamoyl moiety. This was confirmed by HRMS General Experimental Procedures. Optical rotations were measured data: compound 5 was assigned the molecular formula C28H24O12 - on a Perkin-Elmer 241 polarimeter (MeOH, c in g/100 mL). UV were ([M - H] , m/z 551.1185), which differed from 3 and 4 by the recorded on a Perkin-Elmer Lamdba-20 UV-vis spectrophotometer. lack of an oxygen atom. Because the chemical shifts in the 1H NMR UV spectra were recorded in MeOH. 1H and 13C NMR spectra were spectrum of the sugar moiety were close to those observed in 3, recorded on a Varian Unity Inova NMR instrument. 1H and 13C NMR the cinnamoyl group was linked at the C-3′ position. Thus, the new spectra were recorded in CD3OD or DMSO-d6 at 500 and 125 MHz, compound 5 is 3′-O-trans-cinnamoylmangiferin. respectively: chemical shifts are given in ppm as δ relative to TMS. - - HRMS spectra were obtained on a Micromass LCT Premier (Waters) Compound 6 has a molecular formula of C28H24O14 ([M H] , using electrospray as the ion source, negative mode, capillary voltage m/z 583.1083), which differed from 4 by an additional oxygen atom. 2.8 kV, cone voltage 40 V, MCP detector voltage 2650 V, source Furthermore, the 1H and 13C NMR spectra showed virtually no temperature 120 °C, desolvation temperature 250 °C, cone gas flow difference in the signals corresponding to the xanthone and sugar 10 L/h, desolvation gas flow 550 L/h. TLC was performed on silica moieties. Even the olefinic protons and carbons in the trans gel 60 F254 Al sheets (Merck) using EtOAc-HCO2H-HOAc-H2O arrangement had similar chemical shifts. The only notable difference (100:11:11:26). VLC separation was carried out on RP-18 Lichroprep was observed for signals assigned to the aromatic protons and (40-63 µm, 120 g of phase). MPLC was performed with a Bu¨chi 681 pump equipped with a Knauer UV detector using a RP-18 Lichroprep carbons: in 6, a phenolic carbon and three aromatic protons in an - × ) (40 63 µm; 460 50 mm i.d.; Merck). LPLC was carried out on a ABX system could be observed (1H, δ 7.06, d, J 1.0 Hz; 1H δ - × ) ) Lobar RP-18 column (LiChroprep 40 63 µm, 310 25 mm i.d.; Merck 6.78, d, J 8.3 Hz and 1H, δ 6.96, dd, J 8.3, 1.5 Hz). These using a CFG ProMinent Duramatt pump equipped with a Bromma 2238 differences indicated the presence of a caffeoyl moiety, linked at Uvicord SII detector). Semipreparative HPLC was performed with a C-3′ via an ester bond. Compound 6 is thus a new compound, 3′- LC-8 pump equipped with a SPD-10A VP (Shimadzu) detector using O-trans-caffeoylmangiferin. a X-Terra Prep-MS C18 ODB column (5 µm, 19 × 150 mm; Waters), - with detection at 254 nm. HPLC-UV-DAD analysis was carried out The molecular formula of compound 7 (C26H22O12:[M- H] , on an HP 1100 system equipped with a photodiode array detector m/z 525.1024) showed that 7 had an oxygen atom less than 2. × 1 13 (Agilent Technologies) with a Nova-Pak RP-18 column (4 µm; 150 Comparing the H and C NMR spectra, an aromatic proton (H- + + ′′ ) 3.9 mm i.d.; Waters) using a CH3CN 0.05% TFA/H2O 0.05% 4 , δ 7.49, t, J 7.8 Hz) could be observed in the spectrum of 7. TFA gradient (2:98-40:60) in 40 min. The detection was performed The pattern of the aromatic protons and carbons was typical of a at 210, 254, 280, and 360 nm. UPLC was performed on an Acquity benzoyl instead of a p-hydroxybenzoyl moiety. The chemical shifts UPLC system (Waters) with a AcquityBEH C18 UPLC column (1.7 of the sugar protons and carbons were similar to those in 2. µm; 150 × 2.1 mm i.d.; Waters). Furthermore, the long-range HMBC correlation between C-7′′ and Plant Material. The leaves of A. patellifera were collected in April H-3′ corroborated the substitution pattern. Compound 7 is thus 3′- 2003 in Valle de Antoń, los Llanitos, Panama, and identified by Prof. O-benzoylmangiferin, a new natural product. Mireya Correa, director of the Herbarium of the University of Panama. Vouchers are deposited at the University of Panama (FLORPAN 6659) Because of the good activities observed for the crude extract of and at the Laboratory of Phamarmacognosy and Phytochemistry, A. patellifera against P. falciparum and the DPPH radical, the Geneva, Switzerland (No. 2005005). isolated compounds were all evaluated for their antioxidant and Extraction and Isolation. The air-dried, powdered leaves of A. antiplasmodial activities (Tables 3 and 4). patellifera (500 g) were first extracted at room temperature with CH2Cl2, As shown in Table 3, mangiferin (1) and compounds 2, 3, and then with MeOH to afford, respectively, 22.2 and 61.5 g of extracts. - 4 were all active in Vitro against the P. falciparum 3D7 clone, which The MeOH extract (30.0 g) was separated by VLC using a MeOH H2O is chloroquine-sensitive. Compounds 5, 6, and 7 were not tested, step gradient. This afforded four fractions: 2.4 g from 25% MeOH, 2.8 g from 50% MeOH, 16 g from 80% MeOH, and 5 g from 100% due to lack of material. However they were less active than the MeOH. A portion of the 80% fraction (10.0 g) was separated by MPLC positive control, chloroquine. Reports of xanthones as antiplasmo- with a MeOH-H2O step gradient (5:95 to 60:40 in 5% steps) to afford dial agents are numerous, and even the mode of action has been 52 fractions. This separation yielded 477 mg of mangiferin (1, fraction proposed.15 However this is the first description of the antiplas- 15), 800 mg of 2 (fraction 28), 90 mg of 3 (fraction 34), 88 mg of 5 modial activity of C-glucosylxanthones. (fraction 44), and 98 mg of 4 (fraction 45). 1890 Journal of Natural Products, 2008, Vol. 71, No. 11 Martin et al.

′ R 25 Fraction 29 was purified by low-pressure liquid chromatography 3 -O-p-Benzoylmangiferin (7): yellow, amorphous solid; [ ]D 18 (LPLC) with a MeOH-H2O step gradient, yielding 10 mg of 6. Fraction (MeOH, c 1.0); UV (MeOH) λmax (log ) 204 sh (4.48), 239 (4.37), 37 was purified on Sephadex LH-20 eluted with MeOH to give 258 (4.37) 316 (4.10) 364 (3.99) nm; 1H and 13C NMR see Tables 1 - compound 7 (4 mg). Semipreparative LC with the eluent H2O-MeCN and 2; HRESIMS m/z 525.1024 (C26H21O12:[M- H] , requires 7% in the isocratic mode on fraction 13 afforded isomangiferin. 525.1033). Radical-Scavenging Activity (DPPH•) TLC Assay.10 A TLC autographic assay of radical-scavenging activity using the stable DPPH Acknowledgment. The authors would like to thank the Swiss radical was applied for extract screening. After application of 100 µg National Science Foundation for financial support of this work (grant of the samples on silica gel 60 F254 Al plates (Merck), development no. 200020-107775/1 to K.H.). was with n-hexane-EtOAc (1:1) for the CH2Cl2 extracts or - - CH2Cl2 MeOH H2O (13:7:1) for the MeOH extracts. Plates were References and Notes thoroughly dried for complete removal of solvents. A solution of 2,2- diphenyl-1-picrylhydrazyl radical (DPPH, 2 mg/mL in MeOH) was then (1) Martin, F.; Hay, A.-E.; Corno, L.; Gupta, M. P.; Hostettmann, K. sprayed. Inhibitors appeared as yellow spots against a purple background. Phytochemistry 2007, 68, 1307–1311. HelV. Chim. ALP and DPPH Microplate Assay. The microplate alkaline (2) Cuendet, M.; Hostettmann, K.; Potterat, O.; Dyatmiko, W. Acta 1997, 80, 1144–1152. phosphatase (ALP) oxidation protection assay was used to deter- 16 (3) Gentry, A. H. Ann. Mo. Bot. Gard. 1992, 79, 53–64. mine the pEC50 of the pure compounds as described before. In the (4) Von Poser, G. L.; Schripsema, J.; Henriques, A.; Jensen, S. R. Biochem. same approach the determination of the ER50 of pure compounds for Syst. Ecol. 2000, 28, 351–366. the radical-scavenging activity of the stable DPPH radical was done in (5) Lohmann, L. G. Am. J. Bot. 2006, 93, 304–318. a microplate assay, based on the technique described by Ancerewicz (6) Pauletti, P. M.; Castro-Gamboa, I.; Silva, D. H. S.; Young, M. C. M.; et al.17 Tomazela, D. M.; Eberlin, M. N.; Bolzani, V. d. S. J. Nat. Prod. 2003, Antiplasmodial Assay. Antiplasmodial activity was determined 66, 1384–1387. using the 3D7 and K1 strains of P. falciparum as previously described.18 (7) Lima, C. S. D. A.; Cavalcanti de Amorim, E. L.; Xiato da Fonseca, R 25 19 K.; de Sena, R.; Chiappeta, A. d. A.; Nunes, X. P.; Agra, M. d. F.; Mangiferin (1): yellow, amorphous solid; [ ]D see Ajdanga; UV 20 1 13 Leitao da-Cunha, E. V.; Sobral da Silva, M.; Barbosa-Filho, J. M. see Shahat et al.; H and C NMR see Tables 1 and 2; HRESIMS ReV. Bras. Cienc. Farm. 2003, 39, 77–81. - - m/z 421.0747 (C19H17O11:[M H] , requires 421.0771). (8) Gonzalez, B.; Suarez-Roca, H.; Bravo, A.; Salas-Auvert, R.; Avila, 3′-O-p-Hydroxybenzoylmangiferin (2): yellow, amorphous solid; D. Pharm. Biol. 2000, 38, 287–290. R 25 [ ]D 65 (MeOH, c 1.0); UV (MeOH) λmax (log ) 203 (4.64), 243 (4.48), (9) Leite, J. P. V.; Oliveira, A. B.; Lombardi, J. A.; Filho, J. D. S.; Chiari, 258 (4.58), 316 (4.14), 365 (4.06) nm; 1H and 13C NMR see Tables 1 E. Biol. Pharm. Bull. 2006, 29, 2307–2309. - (10) Woodson, J.; Robert, E.; Schery, R. W.; Gentry, A. H. Ann. Mo. Bot. and 2; HRESIMS m/z 541.0985 (C26H21O13:[M- H] , requires 541.0982). Gard. 1973, 60, 781–977. J. Chromatogr. A 2006 1104 3′-O-trans-Coumaroylmangiferin (3): yellow-orange, amorphous (11) Sun, Q.; Sun, A.; Liu, R. , , 69–74. (12) Faizi, S.; Zikr-ur-Rehman, S.; Ali, M.; Naz, A. Magn. Reson. Chem. R 25 solid; [ ]D 61 (MeOH, c 1.0); UV (MeOH) λmax (log ) 203 (4.52), 2006, 44, 838–844. 1 231 (4.41), 258 (4.43), 237 (4.23), 316 (4.41), 361 (4.01) nm; H and (13) Catalano, S.; Luschi, S.; Flamini, G.; Cioni, P. L.; Nieri, E. M.; Morelli, 13 C NMR see Tables 1 and 2; HRESIMS m/z 567.1122 (C28H23O13: I. Phytochemistry 1996, 42, 1605–1607. [M - H]-, requires 567.1139). (14) Markham, K. R.; Wallace, J. W. Phytochemistry 1980, 19, 415–20. 6′-O-trans-Coumaroylmangiferin (4): orange, amorphous solid; (15) Ignatushchenko, M. V.; Winter, R. W.; Bachinger, H. P.; Hinrichs, R 25 - D. J.; Riscoe, M. K. FEBS Lett. 1997, 409, 67–73. [ ]D 12 (MeOH, c 1.0); UV (MeOH) λmax (log ) 203 (4.58), 239 (4.50), 258 (4.51), 315 (4.46), 363 (4.06) nm; 1H and 13C NMR see (16) Bertolini, F.; Novaroli, L.; Carrupt, P. A.; Reist, M. J. Pharm. Sci. Tables 1 and 2; HRESIMS m/z 567.1118 (C H O :[M- H]-, 2007, 96, 2931–2944. 28 23 13 (17) Ancerewicz, J.; Migliavacca, E.; Carrupt, P. A.; Testa, B.; Bree, F.; requires 567.1139). ′ Zini, R.; Tillement, J. P.; Labidalle, S.; Guyot, D.; Chauvet-Monges, 3 -O-trans-Cinnamoylmangiferin (5): orange, amorphous solid; A. M.; Crevat, A.; Le Ridant, A. Free Radical Biol. Med. 1998, 25, R 25 [ ]D 21 (MeOH, c 1.0); UV (MeOH) λmax (log ) 203 sh (4.56), 222 113–120. (4.39), 241 (4.38), 258 (4.46), 314 (4.18), 361 (3.97) nm; 1H and 13C (18) Vivas, L.; Easton, A.; Kendrick, H.; Cameron, A.; Lavandera, J.-L.; NMR see Tables 1 and 2; HRESIMS 551.1185 m/z (C28H23O12:[M- Barros, D.; de las Heras, F. G.; Brady, R. L.; Croft, S. L. Exp. H]-, requires 551.1190). Parasitol. 2005, 111, 105–114. ′ R 25 (19) Adjangba, S. Bull. Soc. Chim. Fr. 1964, 376–380. 3 -O-trans-Caffeoylmangiferin (6): yellow, amorphous solid; [ ]D 20 (MeOH, c 1.0); UV (MeOH) λ (log ) 204 sh (4.52), 219 (4.42), (20) Shahat, A. A.; Hassan, R. A.; Nazif, N. M.; Van Miert, S.; Pieters, max L.; Hammuda, F. M.; Vlietinck, A. J. Planta Med. 2003, 69, 241 (4.46), 258 (4.48), 316 (4.33), 363 (4.08) nm; 1H and 13C NMR - 1068–1070. see Tables 1 and 2; HRESIMS m/z 583.1083 (C28H23O14:[M- H] , requires 583.1088). NP800406Q 852 J. Nat. Prod. 2009, 72, 852–856

Antioxidant Phenylethanoid Glycosides and a Neolignan from Jacaranda caucana

Fre´de´ric Martin,† Anne-Emmanuelle Hay,† Valentin R. Quinteros Condoretty,† Delphine Cressend,‡ Marianne Reist,‡ Mahabir P. Gupta,§ Pierre-Alain Carrupt,‡ and Kurt Hostettmann*,† Laboratory of Pharmacognosy and Phytochemistry, School of Pharmaceutical Sciences, UniVersity of GeneVa, UniVersity of Lausanne, Quai Ernest-Ansermet 30, CH-1211 GeneVa 4, Switzerland, Unit of Pharmacochemistry, School of Pharmaceutical Sciences, UniVersity of GeneVa, UniVersity of Lausanne, Quai Ernest-Ansermet 30, CH-1211 GeneVa 4, Switzerland, and Center for Pharmacognostic Research on Panamanian Flora (CIFLORPAN), College of Pharmacy, UniVersity of Panama, Panama

ReceiVed January 23, 2009

Extracts from several plants of the family Bignoniaceae from Panama were submitted to a rapid DPPH TLC test for the detection of radical-scavenging activity. The MeOH extract of the stems of Jacaranda caucana, a tree that grows from Costa Rica to Colombia, was selected due to its interesting activity and the lack of phytochemical studies on the polar extract. This extract was partitioned between ethyl acetate, butanol, and water. The EtOAc fraction afforded two new phenylethanoid glycosides (1, 2), along with protocatechuic acid, acteoside, and jionoside D. Further puriﬁcations yielded isoacteoside and martynoside. The BuOH fraction afforded a new rhamnosyl derivative of sisymbrifolin (8), a neolignan. The structures were determined by means of spectrometric methods, including 1D and 2D NMR experiments and MS analysis.

Several plant extracts were screened as part of our continuing This paper describes the isolation and characterization of two investigations on Panamanian Bignoniaceae species,1,2 with the aim new phenylethanoid glycosides (1, 2), a new glycoside of the of discovering potential antioxidant drug candidates. The Bignoni- neolignan sisymbrifolin (8), and five known compounds, as well aceae family comprises about 120 genera and 800 species, growing as their antioxidant and radical-scavenging activities. mainly in Africa and Central and South America.3 Species of this family are used for many purposes, such as horticulture, ornamen- Results and Discussion 4 tals, timber, food, handicrafts, dyes, and medicine. The best-known A liquid-liquid extraction of the MeOH extract using H2O, medicinal use comes from the inner bark preparations from various EtOAc, and H2O-saturated BuOH afforded three fractions. They species of Tabebuia, mainly T. impetigosa, called pau d’arco in were tested against DPPH. The two organic fractions were active. Brazil, as an analgesic, anti-inflammatory, antineoplasic, and The EtOAc fraction was then chromatographed on RP-18 by MPLC. diuretic.5 Members of the family have not been extensively This afforded directly several pure compounds: the new phenyle- chemically investigated.6 The plants from the genus Jacaranda are thanoid glycosides 1 and 2, protocatechuic acid 3,16 acteoside 4,17 mainly trees, used as ornamentals around the world because of their and jionoside D 5.18 Further separation by semipreparative LC spectacular flowers. Previous phytochemical studies indicate that yielded isoacteoside 617 and martynoside 7.19 This is the first report the genus is a source of various secondary metabolites, such as of protocatechuic acid and martynoside in the genus Jacaranda. phenylethanoid glycosides, flavonoids, quinones, phytosterols, and The new neolignan 8 was isolated from the BuOH fraction by - anthocyanidins.7 9 reversed-phase flash chromatography. In this work, 18 extracts (CH2Cl2 and MeOH) from seven Panamanian plants of the family Bignoniaceae were submitted to a rapid TLC 1,1-diphenyl-2-picrylhydrazyl (DPPH) test.10 The MeOH extract from the stems of Jacaranda caucana Pittier subsp. sandwithiana A.H.Gentry was selected due to its good radical- scavenging activity. J. caucana was first described in 1917 by Pittier. It was found in the Cauca Department of Colombia, hence its name. Further studies by Gentry divided the species into four subspecies: calycina, caucana, glabrata, and sandwithiana.11 In Panama, only the former has been described.12 The only phytochemical studies were done by Ogura et al.,13,14 who reported cytotoxic activities of J. caucana due to the presence of jacaranone, a quinonid compound. They also isolated and characterized, from a total MeOH extract of the twigs and leaves, several triterpenes including betulinic and jacarandic acids. The subspecies were not mentioned. A recent study showed a moderate activity of the MeOH extract of the leaves against Plasmodium falciparum and reported the use of this species for the treatment of leishmaniasis by the local population in an area of Southwestern Colombia.15

* To whom correspondence should be addressed. Tel: +41 22 379 34 00. Fax: +41 22 379 33 99. E-mail: [email protected]. † Laboratory of Pharmacognosy and Phytochemistry, University of Geneva, University of Lausanne. ‡ Unit of Pharmacochemistry, University of Geneva, University of Compound 1 had a UV spectrum quite similar to that of acteoside Lausanne. (λmax of acteoside: 218, 290, and 330 nm, λmax of 1 in Experimental § University of Panama. Section). Moreover, the 1H and 13C NMR spectra (Table 1) were

10.1021/np900038j CCC: $40.75  2009 American Chemical Society and American Society of Pharmacognosy Published on Web 04/10/2009 Antioxidant Phenylethanoid Glycosides from Jacaranda Journal of Natural Products, 2009, Vol. 72, No. 5 853

1 13 Table 1. H and C NMR Data of Compounds 1 and 2 (500 and 125 MHz, in CD3OD) 12 1H 13C 1H 13C acetylcyclohexyl 1 70.4 70.0 2-6 1.78 (2H, m) 35.9 2.62 (2H, m)a 37.5b 1.51 (2H, m) 2.16 (2H, m)a 37.6b 2.06 (2H, m)a 3-5 1.67 (4H, m) 31.4 1.93 (2H, m)a 4 3.50 (1H, m) 70.6 214.3 7 2.48 (2H, s) 48.3 2.68 (2H, s) 47.4 8 172.6 172.3 glycosyl 1′ 4.42 (1H, d, J ) 8.3 Hz) 104.4 4.40 (1H, d, J ) 8.3 Hz) 104.4 2′ 3.41 (1H, m) 76.3 3.39 (1H, m) 76.3 3′ 3.84 (1H, t, J ) 9.0 Hz) 81.5 3.84 (1H, t, J ) 9.0 Hz) 81.5 4′ 5.01 (1H, t, J ) 9.5 Hz) 73.2 5.01 (1H, t, J ) 9.5 Hz) 72.6 5′ 3.76 (1H, m) 73.9 3.78 (1H, m) 73.2 6′ 4.22 (1H, dd, J ) 11.7, 2.7 Hz) 63.9 4.23 (1H, dd, J ) 12.0, 2.7 Hz) 63.9 4.12 (1H, dd, J ) 11.7, 5.1 Hz) 4.14 (1H, dd, J ) 12.0, 5.1 Hz) aglycone 1′′ 131.5 131.5 2′′ 6.70 (1H, d, J ) 2.4 Hz) 117.2 6.69 (1H, d, J ) 2.4 Hz) 117.3 3′′ 146.3 147 4′′ 144.9 144.9 5′′ 6.69 (1H, d,J) 8.3 Hz) 116.5 6.67 (1H, d,J) 8.3 Hz) 116.5 6′′ 6.58 (1H, dd, J ) 8.3, 2.0 Hz) 121.4 6.56 (1H, dd, J ) 8.3, 1.9 Hz) 121.4 R(8′′) 3.98 (1H, m) 72.6 3.98 (1H, m) 72.3 3.72 (1H, m) 3.72 (1H, m) (7′′) 2.81 (2H, m) 36.9 2.78 (2H, m) 36.8 caffeoyl 1′′′ 127.8 127.8 2′′′ 7.07 (1H, d, J ) 1.9 Hz) 115.4 7.05 (1H, d, J ) 1.9 Hz) 115.4 3′′′ 147.0 146.3 4′′′ 150.0 150.0 5′′′ 6.80 (1H, d, J ) 8.3 Hz) 116.7 6.80 (1H, d, J ) 8.3 Hz) 116.5 6′′′ 6.97 (1H, dd, J ) 8.3, 1.9 Hz) 123.4 6.97 (1H, dd, J ) 8.3, 1.9 Hz) 123.4 R(8′′′) 6.29 (1H, d, J ) 16.1 Hz) 114.7 6.27 (1H, d, J ) 16.1 Hz) 114.7 (7′′′) 7.60 (1H, d, J ) 15.6 Hz) 148.3 7.59 (1H, d, J ) 15.6 Hz) 148.4 CO 168.2 168.2 rhamnosyl 1′′′′ 5.20 (1H, d, J ) 1.5 Hz) 103.2 5.19 (1H, d, J ) 1.5 Hz) 103.2 2′′′′ 3.93 (1H, m) 72.5 3.92 (1H, m) 72.5 3′′′′ 3.57 (1H, m) 70.6 3.54 (1H, m) 70.4 4′′′′ 3.31 (1H, m) 73.9 3.28 (1H, m) 73.9 5′′′′ 3.60 (1H, d, J ) 2.9 Hz) 72.2 3.57 (1H, m) 72.6 6′′′′ 1.11 (3H, d, J ) 6.4 Hz) 18.6 1.09 (3H, d, J ) 6.4 Hz) 18.6 a Interchangeable signals. b Interchangeable signals. very similar, particularly for the caffeoyl, hydroxytyrosyl, rham- and 1.67 (H-3, H-5, 4H). The 2D COSY 1H-1H short-range nosyl, and glucosyl moieties. Notable differences were observed correlations spectrum helped to establish the connectivity between between 2 and 3 ppm, where there was no signal in the 1H NMR the protons. Finally, the proton at δ 3.50, characteristic of a methine spectrum of 4, but signals were present in the spectrum of 1. This linked to an OH group, had a 2D-COSY 1H-1H correlation with suggested that 1 could be acteoside with an additional substituent. H-3 and H-5. These elements pointed us to the presence of a Compound 1 was shown to have the molecular formula C37H48O18 saturated symmetric C-6 ring, with OH groups in positions 1 and ([M - H]-: m/z 779.2759 by HRMS). Thus, the new substituent 4 and an acetyl group in position 1. Zhang et al.20 isolated the could have the formula C8H12O3. Downfield chemical shifts of the 1,4-dihydroxycyclohexanacetic acid, with chemical shifts equivalent glucosyl signals H-6′ (δ 3.62 and 3.52 in 4, vs 4.22 and 4.12 in 1) to those observed for this substituent. No information was given and C-6′ (δ 62.5 in 4, vs 63.9 in 1) suggested that this position concerning the relative or absolute configuration. Nevertheless, was esterified. Furthermore, in the long-range 1H-13C HMBC Endo et al.21 isolated a natural product with a closely related spectrum, a correlation between the H-6 protons of the glucosyl structure, rengyol (cis isomer), and synthesized isorengyol (trans group and a new carbon (δ 172.6) confirmed this hypothesis. Singlet isomer). They confirmed both positions of the OH group in the protons (H-7, δ 2.48, 2H) also correlated with this carbon, two molecules. The tertiary OH group was in an axial position in suggesting a substituted acetyl group. These protons in the HMBC the two molecules, and the secondary OH group of rengyol adopted spectrum had correlations with two more carbons, one at δ 70.4, an equatorial orientation, and an axial orientation in isorengyol. typical of a hydroxy group, and one at δ 35.9. The carbon with a Both structures are displayed in Figure 1. A comparison of the 13C signal at δ 70.4 was quaternary, as it did not show any correlation NMR shifts is presented in Table 2. The chemical shifts of the in the short-range 1H-13C HSQC spectrum. According to HSQC cyclohexanacetyl substituent in 1 were much closer to those of and 1H spectra, carbons with signals at δ 35.9 and 31.4 were rengyol. Moreover, a NOE correlation was observed between H-4 respectively linked to protons with δ 1.78, 1.51 (H-2, H-6, 2H each) and H-2/6 (for the signal at δ 1.51), indicating that these protons 854 Journal of Natural Products, 2009, Vol. 72, No. 5 Martin et al.

Table 3. 1H and 13C NMR Data of Compound 8 (500 and 125 MHz, in CD3OD) 1H 13C 1 138.8 2 7.03 (1H, d, J ) 2.0 Hz) 111.4 3 152.2 4 146.6 5 7.08 (1H, d,J) 8.3 Hz) 119.7 6 6.91 (1H, dd, J ) 8.3, 2.0) 119.2 Figure 1. Structures of Rengyol, Isorengyol, and the Substituent 7 5.60 (1H, d, J ) 5.9 Hz) 88.9 in 1. 8 3.46 (1H, m) 55.7 9 a 3.86 (1H, m) 65.1 Table 2. Comparison of the 13C NMR Shifts for cis/trans b 3.77 (1H, m) 3-OMe 3.80 (3H, s) 56.6 Rengyol (both in CDCl3) and the Acetylcyclohexanyl ′ Substituent in 1 (in CD OD) 1 137.2 3 2′ 6.95 (1H, br s) 112.8 trans-rengyol cis-rengyol subs. in 1 3′ 145.4 4′ 149.0 1 71.2 69.9 70.4 ′ 2 34.4 36.1 35.9 5 129.6 6′ 6.90 (1H, br s) 116.8 3 30.9 31.6 31.4 ′ ) 4 67.4 69.8 70.6 7 4.57 (1H, d, J 5.9 Hz) 75.7 8′ 3.68 (1H, m) 77.6 5 30.9 31.6 31.4 ′ ) 6 34.4 36.1 35.9 9 a 3.53 (1H, dd, J 11.5, 4.2 Hz) 64.4 b 3.39 (1H, dd, J ) 11.2, 6.4 Hz) 7 42.9 45.1 48.3 ′ 8 58.9 58.7 172.6 3 -OMe 3.89 (3H. s) 56.9 1′′ 5.34 (1H, d, J ) 1.5 Hz) 101.5 2′′ 4.06 (1H, dd, J ) 3.2, 1.7 Hz) 72.2 ′′ were in an axial orientation, and thus the OH group in position 4 3 3.86 (1H, m) 72.3 4′′ 3.46 (1H, m) 74.0 was in an equatorial position. These elements demonstrate the cis ′′ ′ 5 3.77 (1H, m) 70.9 orientation. Compound 1 is thus 6 -O-(cis-1,4-dihydroxycyclohex- 6′′ 1.22 (3H, d, J ) 6.4 Hz) 18.1 anacetyl)acteoside, a new structure. The UV spectrum of 2 was very similar to that of 1, and the Another long-range correlation between C-8 and an aromatic proton - - molecular formula of 2 was C37H46O18 ([M H] : m/z 777.2575 (δ 6.90, 1H, br s, H-6′) showed the presence of a second aromatic by HRMS), which differed from 1 by the lack of two hydrogens. ring. Another aromatic proton (δ 6.95, 1H, br s, H-2′) was observed Compound 2 could thus be the oxidized form of 1. The 1H and 13C on this ring, as well as two oxygenated aromatic carbons (δ 145.4, NMR spectra confirmed this. The signals for the acteoside moiety C-3′ and 149.0, C-4′) and a quaternary nonoxygenated carbon C-1′ were not modified, but those for the cyclohexanacetyl substituent (δ 137.2). Furthermore, the HMBC data showed a correlation showed major changes. Indeed, in 2, one hydroxylated carbon was between a methoxy group (δ 3.89, 3H, s) and C-3′. The TOCSY missing, whereas one carbon with a chemical shift typical of a spectrum showed another methine-methine-methylene system: δ ketone (δ 214.3) was observed. The substituent was thus a 4.57 (1H, d, J ) 5.9 Hz, H-7′) 3.68 (1H, m, H-8′) 3.53 (1H, dd, J (1-hydroxy-4-oxo)cyclohexanacetyl group. Comparison of NMR ) 11.5, 4.2 Hz, H-9′a), and 3.39 (1H, dd, J ) 11.2, 6.4 Hz, H-9′b), 22 data with the literature confirmed this hypothesis. In this case, which was connected at C-1′, according to the HMBC spectrum. quercetin-3-O-rutinoside substituted with the same cyclohexanacetyl The chemical shifts of the protons and carbons of this side chain group was isolated, and the NMR shifts were in accordance with pointed to a trihydroxypropanoyl group. All these elements revealed those of 2. Compound 2 is thus 6′-O-(1-hydroxy-4-oxo-cyclohex- a8-5′ lignan, often called neolignan.23 The aglycone, dihydroxy- anacetyl)acteoside, also an original structure. dehydrodiconiferyl alcohol or sisymbrifolin, has already been Compound 8 showed a [M - H]- ion peak at m/z 537.2037 in isolated from Eucommia ulmoides Oliv., Eucommiaceae.24 The the HRMS, indicating the molecular formula C26H34O12. As shown relative configurations of H-7 and H-8 were trans. Garcıá-Munõz in the Experimental Section, the UV spectrum of 8 differed from et al.25 concluded in their work on the synthesis of dihydrodehy- those of 1 and 2, indicating that the structure did not belong to the drodiconiferyl alcohol and derivatives that the coupling constant same class of compounds. The 1H NMR spectrum (Table 3) showed of H-7 could not be used to determine the relative configuration. the presence of an aromatic ABX system: δ 7.03 (1H, d, J ) 2.0 However, a NOE correlation was observed between H-7 and H-9 Hz, H-2), 7.08 (1H, d,J) 8.3 Hz, H-5), and 6.91 (1H, dd, J ) a/b and another one between H-8 and both H-2 and H-6. These 8.3, 2.0 Hz, H-6). The long-range 1H-13C HMBC spectrum helped correlations indicated that the relative configuration of H-7 and H-8 to identify the groups that occupied positions 3 and 4 of the benzene was trans. Moreover, the CD spectrum of 8 aided the determination ring. The latter appeared to be substituted by an O-rhamnosyl group. of the absolute configuration of C-7 and C-8. Indeed, it showed a The anomeric proton (δ 5.34, 1H, d, J ) 1.5 Hz, H-1′′) correlated negative Cotton effect at 222 nm and a positive one at 292 nm. with an oxygenated aromatic carbon (δ 146.6, C-4), and the rest Nakanishi et al.26 observed a negative Cotton effect around 220 of the sugar signals were in accordance with those of O-rhamnosyl nm for junipercomnoside A, also a trans neolignan. They concluded in 1 and 2, for example. Moreover, a correlation between a methoxy that the absolute configuration was 7S,8R. In the same way, Antus group (δ 3.80, 3H, s) and another oxygenated aromatic carbon at et al.27 observed a positive Cotton effect around 290 nm for some δ 152.2 (C-3) was observed. Furthermore, in support of the position compounds of a set of trans neolignans and assigned a 7S,8R of the OCH3 group, a NOE was observed between the OCH3 protons absolute configuration. Thus, the same absolute configuration could and H-2. Finally, position 1 of this ring was a quaternary carbon be deduced for C-7 and C-8 in 8. While the absolute configuration (δ 138.8), as it did not show any correlation in the short-range of C-7′ and C-8′ could not be established from the available data, 1H-13C HSQC spectrum. The 2D 1H-1H TOCSY spectrum the relative configuration was determined from the coupling constant suggested a sequence of methine-methine-methylene successively of H-7′. Indeed, Deyama et al.24 isolated both erythro- and threo- coupled in this order: δ 5.60 (1H, d, J ) 5.9 Hz, H-7), 3.46 (1H, dihydroxydehydrodiconiferyl alcohol. They observed a coupling m, H-8), 3.86 (1H, m, H-9a), and 3.77 (1H, m, H-9b). Moreover, constant of 5.7 Hz for the erythro isomer and 7.5 Hz for the threo the HMBC spectrum revealed a correlation between H-8 and C-1. one. As in 8, the coupling constant was 5.9 Hz, and it could be Antioxidant Phenylethanoid Glycosides from Jacaranda Journal of Natural Products, 2009, Vol. 72, No. 5 855

Table 4. Activities of Compounds in DPPH and ALP Assays detector (Agilent Technologies) with a Nova-Pak RP-18 column (5 µm; 150 × 3.9 mm i.d.; Waters) using a CH CN + 0.05% TFA/H O + compound ER (DPPH)a EC (ALP) (µM)b 3 2 50 50 0.05% TFA gradient (2:98 to 40:60 in 40 min). The detection was isoacteoside (6) 0.070 ( 0.004 1.20 ( 0.60 performed at 210, 254, 280, and 360 nm. UPLC was performed prior ( ( acteoside (4) 0.078 0.002 1.04 0.39 to HRMS on an Acquity UPLC System (Waters) with an Acquity BEH ( ( 2 0.079 0.001 1.27 0.17 C column (1.7 µm; 50 × 2.1 mm i.d.; Waters) using a CH CN + 1 0.098 ( 0.001 1.27 ( 0.07 18 3 0.1% FA/H O + 0.1% FA gradient (5:95 to 98:2 in 3 min). jionoside D (5) 0.17 ( 0.02 1.10 ( 0.13 2 protocatechuic acid (3) 0.23 ( 0.03 2.56 ( 0.67 Plant Material. The stems of J. caucana subsp. sandwithiana were martynoside (7) 0.38 ( 0.15 2.42 ( 0.98 collected in December 2005 in the Parque Nacional Soberania, Panama, 8 1.84 ( 0.70 9.07 ( 2.18 and identified by Prof. Mireya Correa, director of the Herbarium of quercetin (positive control) 0.090 ( 0.01 1.00 ( 0.07 the University of Panama. Vouchers are deposited at the University of a • Panama (FLORPAN 6840) and at the Laboratory of Phamacognosy ER50 is the ratio of antioxidant concentration to DPPH concentration producing a 50% decrease in DPPH• at steady state. and Phytochemistry, Geneva, Switzerland (No. 2005007). b Extraction and Isolation. The air-dried powdered stems of J. EC50 is the antioxidant concentration that protects ALP to 50% from peroxyl radical-induced activity loss. caucana subsp. sandwithiana were first extracted at room temperature with CH2Cl2, then with MeOH, affording respectively 2.2 and 14.7 g reasonably concluded that the relative configuration of the trihy- of extracts. The MeOH extract (12.0 g) was partitioned by LLE between droxypropanoyl chain was erythro. Compound 8 is thus 4-O- EtOAc and H2O (500 mL of each). The aqueous fraction was then rhamnosyl-7S,8R-7′,8′-erythro-sisymbrifolin, a new compound. partitioned with H2O-saturated n-BuOH (500 mL). This yielded 2.1 g The radical-scavenging effects for compounds 1-8, measured of EtOAc, 3.4 g of n-BuOH, and 7.0 g of H2O phases. The EtOAc with DPPH, as well as the ALP test are shown in Table 4. The phase was separated by medium-pressure liquid chromatography (MPLC) with a MeCN/H O step gradient (2:98 to 35:65 in 5% steps) phenylethanoid glycosides 1, 2, acteoside, and isoacteoside pre- 2 to afford 85 fractions. This separation yielded 27 mg of protocatechuic sented good activity, with the same order of magnitude as the 28 acid (fraction 7, 3), 33 mg of acteoside (fraction 33, 4),3mgof positive control, quercetin. Both jionoside D (5), and martynoside jionoside D (fraction 34, 5), 29 mg of 1 (fraction 44), and 8 mg of 2 (7), with respectively one and no remaining catechol functions, (fraction 47). presented weaker activities. Compound 3, with one catechol Fractions 37 and 54 were purified by semipreparative LC with the function, showed activity similar to that of jionoside D. Compound eluent MeCN/H2O, affording respectively 3 mg of isoacteoside (6) and 8, with no catechol function, presented poor activity. 3 mg of martynoside (7). The alkaline phosphatase test (ALP) is a simple fluorimetric test29 The BuOH phase (1.5 g) was separated by flash chromatography to assess the antioxidant capacity of chemical entities to protect with a MeCN/H2O gradient (2:98 to 30:60) to afford 49 fractions. This proteins from loss of activity caused by peroxyl radicals. As in the separation yielded 10 mg of 8 (fraction 16). DPPH test, phenylethanoid glycosides 1, 2, acteoside, and isoac- The separations were monitored using HPLC. teoside presented interesting activities, similar to quercetine. Radical-Scavenging Activity (DPPH) TLC Assays. A TLC Jionoside D, in the ALP test, showed comparable activity, differing autographic assay of radical-scavenging activity using the stable DPPH from the results obtained with the DPPH test. The activity in the radical was applied for extract screening. After application of 100 µg DPPH test seemed to be linked with the presence of catechol of the samples on silica gel 60 F254 Al plates (Merck), development function(s). The ALP test, on the other hand, was associated with was with hexane/EtOAc (1:1) for the CH2Cl2 extracts or CH2Cl2/MeOH/ H O (13:7:1) for the MeOH extracts. Plates were thoroughly dried for phenol function(s). In the case of jionoside D, since lacking one 2 complete removal of solvents. A solution of 2,2-diphenyl-1-picrylhy- phenol, a slightly weaker activity to acteoside and isoacteoside was drazyl radical (DPPH · , 2 mg/mL in MeOH) was then sprayed. Inhibitors expected. It was not the case, but the loss was only one on four appeared as yellow spots against a purple background.10 phenol moieties. Furthermore, protocatechuic acid and martynoside, ALP and DPPH Microplate Assay. The microplate ALP oxidation both with two phenol functions, showed the same activity, only protection assay was used to determine the EC50 (the antioxidant slightly lower than that of the compounds discussed above. concentration that protects ALP by 50% from peroxyl radical-induced Compound 8 displayed poor activity in the ALP test in accordance activity loss) of the pure compounds as described before.28 In the same with the DPPH test. approach, the determination of the ER50 (the ratio of antioxidant concentration to DPPH• concentration producing a 50% decrease in Experimental Section DPPH• at steady state) of pure compounds for the radical-scavenging General Experimental Procedures. Specific rotations were mea- activity of the stable DPPH• radical was done in a microplate assay, sured on a Perkin-Elmer-241 polarimeter (MeOH, c in g/100 mL). UV based on the technique described by Ancerewicz et al.29 spectra were recorded in MeOH on a Perkin-Elmer-Lambda-20 UV-vis 6′-O-(cis-1,4-Dihydroxycyclohexanacetyl)acteoside (1): yellow, 25 spectrophotometer. UV spectra were recorded in MeOH. The circular amorphous solid; [R] D -39 (MeOH, c 1.0); UV (MeOH) λmax (log ε) dichroism (CD) spectrum was recorded with a Jasco J-810 spectrometer. 218 (4.25), 289 (3.94), 332 (4.04) nm; 1H and 13C NMR, see Table 1; - The MeOH solution was thermostated at 20.0 °C using a Jasco PFD- HRESIMS m/z 779.2759 (C37H47O18 [M - H] , requires 779.2762). 425S system. 1H and 13C NMR spectra were recorded on a Varian Unity 6′-O-(1-Hydroxy-4-oxo-cyclohexanacetyl)acteoside (2): yellow, 1 13 25 Inova NMR instrument. H and C NMR spectra were recorded in amorphous solid; [R] D -45 (MeOH, c 1.0); UV (MeOH) λmax (log ε) 1 13 CD3OD or DMSO-d6 at 500 and 125 MHz, respectively. HRMS spectra 220 (4.29), 289 (4.02), 333 (4.14) nm; H and C NMR, see Table 1; - were obtained on a Micromass LCT Premier (Waters) using electrospray HRESIMS m/z 777.2575 (C37H45O18 [M - H] , requires 777.2506). as the ion source, negative mode, capillary voltage 2.8 kV, cone voltage 4-O-Rhamnosyl-7S,8R-7′,8′-erythro-sisymbrifolin (8): yellow- ° 25 40 V, MCP detector voltage 2650 V, source temperature 120 C, orange, amorphous solid; [R] D -45 (MeOH, c 1.0); UV (MeOH) λmax desolvation temperature 250 °C, cone gas flow 10 L/h, desolvation gas (log ε) 229 (4.03), 281 (3.60), 328 (2.83) nm; CD (MeOH, c 0.05) flow 550 L/h. TLC was performed on silica gel 60 F Al sheets 1 13 254 λmax (∆ε) 222 (-3.23), 244 (1.11), 292 (1.29); H and C NMR, see (Merck). MPLC was performed with a Bu¨chi 681 pump equipped with Table 3; HRESIMS m/z 537.2037 (C H O [M - H]-, requires - × 26 33 12 a Knauer UV detector using a RP-18 Lichroprep (40 63 µm; 230 537.1972). 50 mm i.d.; Merck) column. Flash chromatography was performed on a Spot System (ARMEN) with a precolumn SVF D26 (RP-18, 40-63 Acknowledgment. The authors would like to thank the Swiss µm, Merck) and a column of RP-18 Lichroprep (15-25 µm; 400 × 30 mm i.d.). Semipreparative HPLC was performed with a LC-8 pump National Science Foundation for financial support of this work (grant equipped with a SPD-10A VP (Shimadzu) detector using a XTerra Prep- no. 200020-107775/1 to K.H.). CD spectra were recorded in the MS C18 ODB column (5 µm, 19 × 150 mm; Waters), with detection Laboratory of Crystallography of Prof. H. Stoeckli-Evans, Institute of at 254 nm using CH3CN/H2O gradients. HPLC-UV-DAD analyses were Chemistry, University of Neuchaˆtel, Switzerland, by Dr. Laurette carried out on a HP1100 system equipped with a photodiode array Schmitt. 856 Journal of Natural Products, 2009, Vol. 72, No. 5 Martin et al.

Supporting Information Available: 1H and 13C NMR spectra of (15) Weniger, B.; Robledo, S.; Arango, G. J.; Deharo, E.; Aragon, R.; compounds 1, 2, and 8. This material is available free of charge via Munoz, V.; Callapa, J.; Lobstein, A.; Anton, R. J. Ethnopharmacol. the Internet at http://pubs.acs.org. 2001, 78, 193–200. (16) Zhang, H. L.; Nagatsu, A.; Okuyama, H.; Mizukami, H.; Sakakibara, References and Notes J. Phytochemistry 1998, 48, 665–668. (17) Wu, J.; Huang, J.; Xiao, Q.; Zhang, S.; Xiao, Z.; Li, Q.; Long, L.; (1) Martin, F.; Hay, A.-E.; Corno, L.; Gupta, M. P.; Hostettmann, K. Huang, L. Magn. Reson. Chem. 2004, 42, 659–662. Phytochemistry 2007, 68, 1307–1311. (18) Sasaki, H.; Nishimura, H.; Chin, M.; Chen, Z.; Mitsuhashi, H. (2) Martin, F.; Hay, A.-E.; Cressend, D.; Reist, M.; Vivas, L.; Gupta, Phytochemistry 1989, 28, 875–879. M. P.; Carrupt, P.-A.; Hostettmann, K. J. Nat. Prod. 2008, 71, 1887– (19) Calis, I.; Lahloub, M. F.; Rogenmoser, E.; Sticher, O. Phytochemistry 1890. 1984, 23, 2313–2315. (3) Spichiger, R.-E.; Perret, M.; Figeat, M.; Jeanmonod, D. Systematic (20) Zhang, G.-G.; Song, S.-J.; Ren, J.; Xu, S.-X. J. Herb. Pharmacother. Botany of Flowering Plants: a New Phylogenetic Approach to 2002, 2, 35–40. Angiosperms of the Temperate and Tropical Regions; Science Publ.: (21) Endo, K.; Seya, K.; Hikino, H. Tetrahedron 1987, 43, 2681–2688. Enfield, 2004; p 330. (22) Hafez, S.; Jakupovic, J.; Bohlmann, F.; Sarg, T. M.; Omar, A. A. (4) Gentry, A. H. Ann. Mo. Bot. Gard. 1992, 79, 53–64. Phytochemistry 1989, 28, 843–847. (5) de Miranda, F.; Vilar, J.; Alves, I.; Cavalcanti, S.; Antoniolli, A. BMC (23) Davin, L.; Lewis, N. Phytochem. ReV. 2003, 2, 257–288. Pharmacol. 2001, 1,6. (24) Deyama, T.; Ikawa, T.; Kitagawa, S.; Nishibe, S. Chem. Pharm. Bull. (6) Von Poser, G. L.; Schripsema, J.; Henriques, A.; Jensen, S. R. Biochem. 1987, 35, 1785–1789. Syst. Ecol. 2000, 28, 351–366. (25) Garcıá-Munõz, S.; A´ lvarez-Corral, M.; Jimeńez-Gonza´lez, L.; Lo´pez- (7) Blatt, C. T. T.; Dos Santos, M. D.; Salatino, A. Plant Syst. EVol. 1998, Sańchez, C.; Rosales, A.; Munõz-Dorado, M.; Rodrı´guez-Garcıá, I. 210 (3/4), 289–292. Tetrahedron 2006, 62, 12182–12190. (8) Moharram, F. A.; Marzouk, M. S. A. Z. Naturforsch., B: Chem. Sci. (26) Nakanishi, T.; Iida, N.; Inatomi, Y.; Murata, H.; Inada, A.; Murata, 2007, 62, 1213–1220. J.; Lang, F. A.; Linuma, M.; Tanaka, T. Phytochemistry 2004, 65, (9) Fernandez, M. E. An. Asoc. Quim. Argent. 1968, 56, 135–137. 207–213. (10) Cuendet, M.; Hostettmann, K.; Potterat, O.; Dyatmiko, W. HelV. Chim. (27) Antus, S.; Kurtan, T.; Juhasz, L.; Kiss, L.; Hollosi, M.; Majer, Z. Acta 1997, 80, 1144–1152. Chirality 2001, 13, 493–506. (11) Gentry, A. H. Bignoniaceae, part II (tribe Tecomeae); New York (28) Bertolini, F.; Novaroli, L.; Carrupt, P. A.; Reist, M. J. Pharm. Sci. Botanical Garden: New York, 1992. 2007, 96, 2931–2944. (12) D’Arcy, W. G. Flora of Panama: Checklist and Index; Missouri (29) Ancerewicz, J.; Migliavacca, E.; Carrupt, P. A.; Testa, B.; Bree, F.; Botanical Garden: St. Louis, MO, 1987. Zini, R.; Tillement, J. P.; Labidalle, S.; Guyot, D.; Chauvet-Monges, (13) Ogura, M.; Cordell, G. A.; Farnsworth, N. R. Lloydia 1976, 39, 255– A. M.; Crevat, A.; Le Ridant, A. Free Radical Biol. Med. 1998, 25, 257. 113–120. (14) Ogura, M.; Cordell, G. A.; Farnsworth, N. R. Lloydia 1977, 40, 157– 168. NP900038J

Page 1 of 16 Natural Product Research

1 2 3 4 Antioxidant potential and radical scavenging effects of flavonoids 5 from the leaves of Psidium cattleianum grown in French Polynesia 6 7 R. Hoa, A. Violettea, D. Cressendb, P. Raharivelomananac, P.A. Carruptb, 8 a 9 K. Hostettmann 10 11 aLaboratory of Pharmacognosy and Phytochemistry, School of Pharmaceutical 12 Sciences, University of Geneva, University of Lausanne, Quai Ernest-Ansermet 30, 13 b 14 CH-1211 Geneva 4, Switzerland; Unit of Pharmaceutical Sciences, University of 15 Geneva, University of Lausanne, Quai Ernest-Ansermet 30, CH-1211 Geneva 4, 16 Switzerland;For cBiodiversité Peer Terrestre Reviewet Marine, University Onlyof French Polynesia, BP 17 6570, 98702 Faa’a, Tahiti, French Polynesia. 18 19 20 Corresponding author: Prof. Kurt Hostettmann, Chalet Gentianella 1938 Champex, 21 Valais, Switzerland, E-mail : [email protected] 22 23 (Received XX Month Year; final version received XX Month Year) 24 25 Psidium cattleianum J. Sabine (Myrtaceae) is a traditional medicinal plant in French 26 Polynesia. The leaves and the roots are used for astringent, analgesic and 27 hepatoprotective properties. These effects may be correlated with the presence of 28 antioxidant compounds. Extracts in organic solvents from the leaves of P. cattleianum 29 have been screened by a rapid DPPH TLC test. The MeOH extract exhibited significant 30 antioxidant activity. Seven flavonoids (1-7) along with a benzoic acid were isolated by 31 bio-guided fractionation. The compounds 1-7 indicated strong antioxidant and radical- 32 scavenging activities for ALP, DPPH•, ABTS•- and ORAC assays. This study 33 demonstrates that the leaves of P. cattleianum contain main compounds with interesting 34 antioxidant and radical-scavenging activities clarified by four biological assays. These 35 findings may justify the use of the leaves in traditional medicine in French Polynesia. 36 Among the total eight known compounds, Reynoutrin and Luteolin were isolated for the 37 first time from the genus Psidium. 38 39 Keywords: ALP assay; DPPH• assay; ABTS•- assay; ORAC assay; Psidium cattleianum; French 40 Polynesia. 41 42 43 44 1.1. Introduction 45 Free radicals are produced in normal cell metabolism and oxidation is essential to 46 47 many living organisms for the production of energy to fuel biological processes. 48 49 50 However, the uncontrolled production of oxygen derived free radicals is found to be 51 52 related to illness, cell damage, cell death or gene mutation and is the cause of many 53 54 55 diseases such as cancer, rheumatoid arthritis, atherosclerosis, heart failure, 56 57 neurodegenerative disorders, diabetes mellitus and degenerative processes associated 58 59 with aging (Gutteridge & Halliwell, 2000). More recently, reactive oxygen species 60 have been of interest in free radical research due to their effects on human health

URL: http://mc.manuscriptcentral.com/gnpl Natural Product Research Page 2 of 16

1 2 3 (Matés & Sanchez-Jiménez, 2000). Many reports have shown the relationship 4 5 6 between oxidative stress and some major chronic diseases. Specifically, the role of 7 8 oxidative stress has been demonstrated for cancer (Kruk & Aboul-Enein, 2007), liver 9 10 11 disease (Cederbaum et al., 2009), Alzheimer’s disease (Cai & Yan, 2007) and 12 13 inflammation (Kao et al., 2009). As a result, antioxidants have been used to prevent 14 15 oxidative damage in these diseases. In this context, natural antioxidants are receiving 16 For Peer Review Only 17 18 increased attention due to their potential for therapeutic use and prevention of these 19 20 diseases. They could be an alternative to the use of synthetic compounds in food and 21 22 pharmaceutical technology. Therefore, the development and use of more effective 23 24 25 natural antioxidants is desirable. 26 27 Traditional medicine is being reevaluated by extensive research on different 28 29 plant species and their therapeutic principles. Since plants produce antioxidants to 30 31 32 control the oxidative stress caused by oxygen, they could represent a source of 33 34 compounds with antioxidant activity. 35 36 37 Psidium cattleianum J. Sabine (Myrtaceae), known as tuava tinito in French 38 39 Polynesia is a shrub or tree up to 8 m high, native to the Brazilian lowlands. This 40 41 plant is also widely distributed in tropical and subtropical areas and considered as an 42 43 44 invasive plant although it has a curative effect. P. cattleianum has been used in 45 46 folkloric medicine by the local population as a source of anti-ulcer, analgesic and anti- 47 48 inflammatory drugs (Crivelaro de Menezes, Delbem, Brighenti, Okamoto, Gaetti- 49 50 51 Jardim, 2010). In French Polynesia, its leaves have been used for many years for their 52 53 astringent, antidiarrheal, analgesic and hepatoprotective properties (Petard, 1986). A 54 55 literature survey showed some studies on the acidic polysaccharides of the fruits 56 57 58 (Vriesmann, Petkowicz, Carneiro, Costa, & Beleski-Carneiro, 2009), on the chemical 59 60 composition of essential oils (Pino, Bello, Urquiola, Marbot, & Martí, 2004; Marques

URL: http://mc.manuscriptcentral.com/gnpl Page 3 of 16 Natural Product Research

1 2 3 et al., 2008) and on the antimicrobial properties on Streptococcus mutans (Crivelaro 4 5 6 de Menezes et al., 2010) of the leaves of P. cattleianum but no phytochemical study 7 8 has yet been carried out on the antioxidant and radical scavenging activity of the 9 10 11 leaves of this plant. 12 13 The aim of this work was to determine the active compounds from the crude 14 15 methanolic extract of the leaves of P. cattleianum and to assess their antioxidant 16 For Peer Review Only 17 • •- 18 properties for ALP, DPPH , ABTS and ORAC assays. To our knowledge, the present 19 20 investigation is the first phytochemical and biological report on the leaves of Psidium 21 22 cattleianum from French Polynesia. 23 24 25 26 1.2. Results and discussion 27 28 1.2.1. Results 29 Bioguided investigation of the methanolic extract resulted in the isolation and 30 31 32 characterization of seven flavonoids, Reynoutrin (1), Guajavarin (2), Quercitrin (3), 33 34 Morin (4), Myricetin (5), Luteolin (6) and Kaempferol (7) along with a benzoic acid, 35 36 protocatechuic acid. The structures of compounds 1-7 are given in Figure 1. 37 38 39 40 41 1.2.1.1. Antioxidant activity 42 43 The antioxidant and radicals-scavenging effects for the flavonoids, measured with 44 •- • 45 ALP, ORAC, ABTS and DPPH assays are shown in Table 1. 46 47 The DPPH• scavenging activity of Quercetin derivatives have been assessed 48 49 50 since they turned out to be active after the screening. 5 and 6 have exhibited an 51 52 efficient activity, close to that of Quercetin (ER50 value equals 0.07, 0.05 and 0.09 53 54 • respectively). The activity toward DPPH radical is slightly less efficient for 7 (ER50 55 56 57 value of 0.14), 2 (ER50 value of 0.15) and 4 (ER50 value of 0.16). 3 and 1 have 58 59 showed the poorer activity with an ER50 value of 0.22 and 0.25, respectively. 60 Nevertheless, this activity is a good DPPH• scavenging activity. Myricetin (5) has also

URL: http://mc.manuscriptcentral.com/gnpl Natural Product Research Page 4 of 16

1 2 3 exhibited the best free radical scavenging capacity toward ABTS•– radical (ER value 4 50 5 6 of 0.06) which is similar to that of Quercetin (ER50 value of 0.07). 4, 7 and 6 have 7 8 turned out to be slightly less efficient in scavenging the ABTS•– radical. Indeed, their 9 10 11 ER50 value equals 0.12, 0.14 and 0.18, respectively. However, the activity of 12 13 glycosyled-quercetin derivatives is poor, the ER50 value equals 0.52 for 2, 0.56 for 1 14 15 and 0.61 for 3. Luteolin (6), Morin (4) and Kaempferol (7) are the most efficient 16 For Peer Review Only 17 18 compounds in protecting fluorescein and ALP protein against peroxyl radical-induced 19 20 damages. Their EC50 equals 6.09, 5.86 and 5.84 in ORAC assay respectively and 5.92, 21 22 5.84 and 5.76 in ALP assay respectively. Their activity is slightly better than that of 23 24 25 Quercetin in ORAC assay (EC50 of 5.78) and slightly poorer than that of Quercetin in 26 27 ALP assay (EC50 of 5.95). Myricetin (5), which has the best free radical-scavenging 28 29 ability toward both DPPH• and ABTS•– radicals, has showed a less efficient activity 30 31 32 toward peroxyl radical with a EC50 of 5.61 and 5.62 in ORAC and ALP assay 33 34 respectively. The gylcosyled-quercetin derivatives have an activity slightly less 35 36 37 efficient than 5 in ORAC and ALP assay. The EC50 value of 3 in ALP assay has not 38 39 been determinate because it turned out to inhibit 53% of the ALP activity at a 40 41 concentration of 10-5M and showed any protecting effect at lower concentration. 42 43 44 45 46 1.2.2. Discussion 47 Each compound is capable of scavenging DPPH• radical by transferring hydrogen 48 49 50 atom. They owe their great activity to the catechol moiety on the B-ring (5, 6, 1, 2 and 51 52 3) and to the OH groups at C3 and C4’ involving a conjugation over both B- and C- 53 54 ring. Moreover, products formed after the H-abstraction from the compounds are 55 56 • 57 themselves able to reduce DPPH radical (Dangles, Fargeix, & Dufour, 1999; Alluis, 58 59 & Dangles, 2001). 4 and 6 are less efficient than 5 and 6 since they do not bear a 60 catechol moiety, their abitity to transfer hydrogen atom is due to the OH groups in the

URL: http://mc.manuscriptcentral.com/gnpl Page 5 of 16 Natural Product Research

1 2 3 position C3 and C4’ and to the 2,3-double bond. Although glycosylated-Quercetin 4 5 6 bear a catechol moiety, they are less efficient in scavenging the stable radical. This 7 8 may be explained by the glycosidation at C3 which do not allow the ortho- 9 10 11 quinone/para-quinoid tautomerism induced after H-abstraction, hence the solvent 12 13 addition can not occur (Dangles et al., 1999; Alluis, & Dangles, 2001). The steric 14 15 hindrance may also reduce the free radical-scavenging activity since 1, 2 and 3 are 16 For Peer Review Only 17 18 less efficient than Luteolin (6). The lack of a 3-OH group on these compounds also 19 20 induces a weaker capacity in scavenging the ABTS•– radical. Indeed, unlike 5, 4 and 7 21 22 which exhibited a similar scavenging activity to that toward DPPH• radical, luteolin 23 24 •– 25 and especially the glycosylated-quercetin derivatives are less efficient toward ABTS 26 27 radical. Electron transfer is supposed to be involved in the ABTS•– reducing. The 3- 28 29 OH appears more useful in the electron transfer than in the hydrogen transfer since 6 30 31 • 32 and 2 scavenges about four times and three times more DPPH radical molecules than 33 34 they do with ABTS•– radical when 1 and 3 reduce about 2-fold DPPH• than ABTS•– 35 36 37 radical molecules. 38 39 Since peroxyl radicals are reduced by hydrogen transfer, the activity of the 40 41 compounds is consistent with that toward DPPH• radical. Only Myricetin (5) is not 42 43 44 among the best peroxyl radical scavengers while it exhibits the best ER50 value in 45 46 DPPH• and ABTS•– assays. 47 48 Morin (4), Myricetin (5), Luteolin (6) and Kaempferol (7) have turned out to 49 50 • •– 51 be great antioxidant compounds toward DPPH , ABTS and peroxyl radicals. These 52 53 antioxidant properties are in agreement with literature (Rice-Evans, Miller, & 54 55 Paganga, 1996; Cao, Sofic, & Prior, 1997; Tsimogiannis & Oreopoulou, 2006). 56 57 • 58 Reynoutrin (1), Quajavarin (2) and Quercitrin (3) showed a good DPPH and peroxyl 59 60 radical-scavenging activity but weak ABTS•– scavenger. This is likely due to the

URL: http://mc.manuscriptcentral.com/gnpl Natural Product Research Page 6 of 16

1 2 3 glycation at C3. As a result, the lack of 3-OH and the steric hindrance reduce the 4 5 6 antioxidant effect. A study has also provided that the sugar moiety on C3 induces 7 8 torsion of the molecule (Van Acker et al., 1996). As a result, the scavenging activity 9 10 •– 11 is reduced. However, the ABTS scavenging activity of compounds 1, 2 and 3 should 12 13 not be as weak. Studies on Rutin, a glycosylated-Quercetin, has been performed 14 15 toward this radical and have provided its good activity, even better than that of 16 For Peer Review Only 17 18 Kaempferol and Luteolin (Rice-Evans et al., 1996; Lien, Ren, Bui, & Wang, 1999). 19 20 21 1.3. Conclusions 22 Nowadays, studies on antioxidants present in plants and foods have come to be one of 23 24 25 the most popular topics in the area of food and agriculture. So, this study shows that 26 27 the leaves of P. cattleianum J. Sabine are a potential source of natural antioxidants. 28 29 30 The results indicate that the isolated compounds, Reynoutrin (1), Guajavarin (2), 31 32 Quercitrin (3), Morin (4), Myricetin (5), Luteolin (6) and Kaempferol (7) have 33 34 powerful antioxidant and radical-scavenging activities, as proven by four biological 35 36 • •- 37 assays, ALP, DPPH , ABTS and ORAC assays. In the past, the leaves of tuava tinito 38 39 were used by the Polynesians and the above results may justify their use by the 40 41 presence of antiradical and antioxidant components which may be efficient as 42 43 44 preventive agents and helpful in controlling complications in various diseases. 45 46 To our knowledge, the present investigation is the first phytochemical and 47 48 biological report on the leaves of Psidium cattleianum grown in French Polynesia, 49 50 51 among the total eight known compounds, Reynoutrin (1) and Luteolin (6) were 52 53 isolated for the first time from the genus Psidium. 54 55 56 57 Acknowledgements 58 The authors would like to thank Dr. Jean-François Butaud (IRD) for taxonomic identification of plant 59 material. The Swiss National Science Foundation (grant no. 200020-107775/1 to Prof. K. Hostettmann) 60 is gratefully acknowledged for financial support.

URL: http://mc.manuscriptcentral.com/gnpl Page 7 of 16 Natural Product Research

1 2 3 References 4 5 Alluis, B., & Dangles, O. (2001). Quercetin (=2-(3,4-dihydrophenyl)-3,5,7- 6 trihydroxy-4H-1-benzopyran-4-one) glycosides and sulfates: chemical 7 synthesis, complexation and antioxidant properties. Helvetica Chimica Acta, 8 84, 1133-1156. doi: 10.1002/1522-2675(20010516)84:5<1133::AID- 9 10 HLCA1133>3.0.CO;2-Z 11 Cai, Z.Y., & Yan, Y. (2007). Pathway and mechanism of oxidative stress in 12 Alzheimer’s disease. Journal of medical Colleges of PLA, 22, 320-325. 13 doi:10.1016/S1000-1948(07)60064-1 14 Cao, G., Sofic, E., & Prior, R.L. (1997). Antioxidant and prooxidant behaviour of 15 flavonoids: Structure-activity relationships. Free Radical Biology and 16 For Peer Review Only 17 Medicine, 22, 749-760. doi:10.1016/S0891-5849(96)00351-6 18 Cederdaum, A.I., Lu,Y., & Wu, D. (2009). Role of oxidative stress in alcohol-induced 19 liver injury. Archives of Toxicology, 83, 519-548. doi: 10.1007/s00204-009- 20 0432-0 21 Crivelaro de Menezes, T.E., Delbem, A.C.B., Brighenti, F.L., Okamoto, A.C., & 22 23 Gaetti-Jardim, Jr E. (2010). Protective efficacy of Psidium cattleianum and 24 Myracrodruon urundeuva aqueous extracts against caries development in rats. 25 Pharmaceutical Biology, 48, 300-305. doi: 10.3109/13880200903122202 26 Dangles, O., Fargeix, G., & Dufour, C. (1999). One-electron oxidation of quercetin 27 and quercetin derivatives in protic and non protic media. Journal of the 28 29 Chemical Society. Perkin Transaction 2, 7, 1387-1395. doi: 10.1039/a901460h 30 Gutteridge, J.M.C., & Halliwell, B. (2000). Free radicals and antioxidants in the year 31 2000, A historical look to the future. Annals of the New York Academy of 32 Sciences, 899, 136-147. doi: 10.1111/j.1749-6632.2000.tb06182.x 33 Kao, E.S., Hsu, J.D., Wang, C.J., Yang, S.H., Cheng, S.Y., & Lee, H.J. (2009). 34 Polyphenols extracted from Hibiscus sabdariffa L. inhibited 35 36 lipopolysaccharide-induced inflammation by improving antioxidative 37 conditions and cyclooxygenase-2 expression. Bioscience Biotechnology and 38 Biochemistry, 73, 385-390. doi: 10.1271/bbb.80639 39 Kruk, J., & Aboul-Enein, H.Y. (2007). Physical activity and cancer prevention: 40 Updating the evidence. The role of oxidative stress in carcinogenesis. Current 41 42 Cancer Therapy Reviews, 3, 81-95. doi: 10.2174/157339407780618443 43 Lien, E.J., Ren, S., Bui, H.H., & Wang, R. (1999). Quantitative structure-activity 44 relationship analysis of phenolic antioxidants. Free Radical Biology and 45 Medicine, 26, 285-294. doi:10.1016/S0891-5849(98)00190-7 46 Marques, F.A., Wendler, E.P., Sales Maia, B.H.L.N., Coffani-Nunes, J.V., Campana, 47 J., & Guerrero Jr., P.G. (2008). Volatile oil of Psidium cattleianum Sabine 48 49 from the Brazilian Atlantic forest. Journal of Essential Oil Research, 20, 519- 50 520. 51 Matés, J.M., & Sanchez-Jiménez, F.M. (2000). Role of reactive oxygen species in 52 apoptosis: Implications for cancer therapy. The International Journal of 53 Biochemistry Cell Biology, 32, 157-170. doi:10.1016/S1357-2725(99)00088-6 54 55 Petard, P. (1986). Plantes utiles de Polynésie. Papeete, Tahiti: Haere Po No Tahiti. 56 Pino, J.A., Bello, A., Urquiola, A., Marbot, R., & Martí, M.P. (2004). Leaf oils of 57 Psidium parvifolium Griseb. and Psidium cattleianum Sabine from Cuba. 58 Journal of Essential Oil Research, 16, 370-371. 59 Rice-Evans, C.A., Miller, N., & Paganga, G. (1996). Structure-antioxidant activity 60 relationships of flavonoids and phenolic acids. Free Radical Biology and Medicine, 20, 933-956. doi:10.1016/0891-5849(95)02227

URL: http://mc.manuscriptcentral.com/gnpl Natural Product Research Page 8 of 16

1 2 3 Tsimogiannis, D.I., & Oreopoulou, V. (2006). The contribution of flavonoid C-ring 4 5 on the DPPH free radical scavenging efficiency. A kinetic approach for the 6 3’,4’-hydroxy substituted members. Innovative Food Science and Emerging 7 Technologies, 7, 140-146. doi:10.1016/j.ifset.2005.09.001 8 Van Acker, S.A.B.E., de Groot, M.J., van den Berg, D.J., Tromp, M.N.J.L., den 9 Kelder, D.O., van der Vijgh, W.J.F., & Bast, A. (1996). A quantum chemical 10 11 explanation of the antioxidant activity of flavonoids. Chemical Research in 12 Toxicology, 9, 1305-1312. doi: 10.1021/tx9600964 13 Vriesmann, L.C., Petkowicz, C.L., Carneiro, P.I.B., Costa, M.E., & Beleski-Carneiro, 14 E. (2009). Acidic polysaccharides from Psidium cattleianum (Araçá). 15 Brazilian Archives of Biology and Technology, 52, 259-264. doi: 16 10.1590/S1516-89132009000200001For Peer Review Only 17 18 19 Supplementary material 20 Experimental information is submitted as supplementary material. 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/gnpl Page 9 of 16 Natural Product Research

1 2 3 Table 1. Activities of isolated compounds from P. cattleianum in DPPH•, ABTS•-, ORAC and 4 ALP assays. 5 • a •- b c d 6 Compounds ER50 (DPPH ) ER50 (ABTS ) pEC50 (ORAC) pEC50 (ALP) 7 8 Reynoutrin (1) 0.25 ± 0.03 0.56 ± 0.12 5.37 ± 0.06 5.44 ± 0.05 9 10 11 Guajavarin (2) 0.15 ± 0.01 0.52 ± 0.05 5.55 ± 0.05 5.55 ± 0.01 12 13 Quercitrin (3) 0.22 ± 0.03 0.61 ± 0.01 5.47 ± 0.05 nd 14 15 Morin (4) 0.16 ± 0.01 0.12 ± 0.02 5.86 ± 0.06 5.84 ± 0.06 16 For Peer Review Only 17 18 Myricetin (5) 0.07 ± 0.01 0.06 ± 0.01 5.61 ± 0.06 5.62 ± 0.10 19 20 Luteolin (6) 0.05 ± 0.01 0.18 ± 0.02 6.09 ± 0.13 5.92 ± 0.10 21 22 Kaempferol (7) 0.14 ± 0.01 0.14 ± 0.01 5.84 ± 0.05 5.76 ± 0.05 23 24 25 Quercetine 26 0.09 ± 0.01 0.07 ± 0.01 5.78 ± 0.02 5.95 ± 0.04 27 (positive control) 28 29 aER is the ratio of antioxidant concentration to DPPH• concentration producing producing an 30 50 absorbance decrease of 50% at 90 min. 31 b •- 32 ER50 is the ratio of antioxidant concentration to ABTS radical concentration producing an absorbance decrease of 50% at 90 min. 33 c 34 pEC50=-logEC50, EC50 is the antioxidant concentration necessary to avoid 50% of the fluorescein 35 oxidation. d 36 pEC50=-logEC50, EC50 is the antioxidant concentration that protects ALP to 50% from peroxyl 37 radical-induced activity loss. 38 nd: not determinate 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/gnpl Natural Product Research Page 10 of 16

1 2 3 4 5

6 OH O 7 R 8 4 R 9 3 10 R2 11 OH O 12 13 14 OH 15 16 For Peer ReviewR Only 17 1 18 19 R1 R2 R3 R4 20 1 H OH H xylose 21 2 OH H H arabinose 22 3 H OH H rhamnose 23 24 4 H H OH OH 25 5 OH OH H OH 26 6 H OH H H 27 7 H H H OH 28 29 Figure 1. Structures of compounds 1-7 isolated from P. cattleianum. 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/gnpl Page 11 of 16 Natural Product Research

1 2 3 4 Antioxidant potential and radical scavenging effects of flavonoids 5 from the leaves of Psidium cattleianum grown in French Polynesia 6 7 R. Hoa, A. Violettea, D. Cressendb, P. Raharivelomananac, P.A. Carruptb, 8 a 9 K. Hostettmann 10 11 Psidium cattleianum J. Sabine (Myrtaceae) is a traditional medicinal plant in French 12 Polynesia. The leaves and the roots are used for astringent, analgesic and 13 hepatoprotective properties. These effects may be correlated with the presence of 14 antioxidant compounds. Extracts in organic solvents from the leaves of P. cattleianum 15 have been screened by a rapid DPPH TLC test. The MeOH extract exhibited significant 16 antioxidantFor activity. Peer Seven flavonoids Review (1-7) along with a benzoic Only acid were isolated by 17 bio-guided fractionation. The compounds 1-7 indicated strong antioxidant and radical- 18 scavenging activities for ALP, DPPH•, ABTS•- and ORAC assays. This study 19 demonstrates that the leaves of P. cattleianum contain main compounds with interesting 20 antioxidant and radical-scavenging activities clarified by four biological assays. These 21 findings may justify the use of the leaves in traditional medicine in French Polynesia. 22 Among the total eight known compounds, Reynoutrin and Luteolin were isolated for the 23 first time from the genus Psidium. 24 25 26 1. Experimental 27 28 1.1. General experimental procedures 29 30 TLC was performed on silica gel 60 F254 Al sheets (Merck). MPLC was performed 31 32 with Büchi 681 pump equipped with a Knauer UV detector using a RP-18 LiChropep 33 34 35 (40-63 µm; 460 x 50 mm i.d., Merck) column. The detection was performed at 365 36 37 nm. Semipreparative HPLC was performed with a LC-8 pump equipped with a SPD- 38 39 40 10A VP (Shimadzu) detector using a XTerra Prep-MS C18 OBD column (5 m, 150 x 41 42 19 mm, Waters), with detection at 365 nm using CH3CN/H2O gradients. HPLC-UV- 43 44 DAD analyses were performed on a Hewlett-Packard 1100 system equipped with a 45 46 47 photodiode array detector (Agilent Technologies) with a Nova-Pak RP-18 column (5 48 49 µm, 150 × 3.9 mm i.d., Waters) using a MeOH-H2O gradient (25:75 to 100:0 in 25 50 51 min). The detection was performed at 365 nm. UPLC was performed prior to HRMS 52 53 54 on an Acquity UPLC System (Waters) with an Acquity UHPLC BEH C18 column (1.7 55 56 m, 100 x 2.1 m i.d., Waters) using CH3CN + 0.1% FA/H2O + 0.1%FA gradient 57 58 59 (5:95 to 98:2 in 3 min). HRMS spectra were obtained on a Micromass LCT Premier 60 (Waters) using electrospray as the ion source, capillary voltage 2.8 kV, cone voltage

URL: http://mc.manuscriptcentral.com/gnpl Natural Product Research Page 12 of 16

1 2 3 40 V, MCP detector voltage 2650 V, source temperature 120°C, desolvation 4 5 6 temperature 250°C, cone gas flow 10L/h, desolvation gas flow 550 L/h. 1D and 2D 7 8 NMR spectra (1H, 13C, COSY, HSQC and HMBC) were recorded on Varian Unity 9 10 11 Inova 500 MHz spectrometer, using methanol-d4 as solvent. 12 13 14 1.2. Plant material 15 Leaves of Psidium cattleianum J. Sabine (Myrtaceae) were collected in Faa’a, Tahiti, 16 For Peer Review Only 17 18 French Polynesia, in May 2008. The plant material was taxonomically identified by 19 20 Dr. Jean-François Butaud, Institut de Recherche pour le Développement, France. A 21 22 voucher specimen (#APC05062) is deposited in the Herbarium of French Polynesia at 23 24 25 the “Musée de Tahiti et ses îles” in Punaauia (PAP, Tahiti Island). 26 27 28 1.3. Extraction and isolation of bioactive compounds 29 30 The air-dried powdered leaves of P. cattleianum (500 g) were extracted successively 31 32 with chlorofrome (3 x 1.5 L, 24 h) and methanol (3 x 1.5 L, 24 h) at room 33 34 temperature, to afford 10.2 g and 7.6 g of extracts, respectively. The methanolic 35 36 37 extract (7.6 g) was partitioned by liquid liquid extraction (LLE) between EtOAc and 38 39 H2O (1.6 L of each). The aqueous fraction was then partitioned with n-BuOH (1.6 L). 40 41 This yielded 2.8 g of EtOAc, 1.7 g of n-BuOH and 3.1 g of H O phases. The EtOAc 42 2 43 44 phase was separated on RP-18 by medium-pressure liquid chromatography (MPLC) 45 46 with a MeCN-H2O step gradient (5:95 to 100:0 in 5% steps) to afford 71 fractions. 47 48 49 This afforded directly one pure compound, 8 mg of benzoic acid. Similar fractions 50 51 were combined after HPLC examination to provide four fractions: A (142 mg), B 52 53 (171 mg), C (194 mg) and D (310 mg). Fraction B was further purified by 54 55 56 semipreparative LC with the eluent MeCN/H2O to give 3.6 mg of 1, 10 mg of 2 and 57 58 5.2 mg of 3. Fraction C was also chromatographed by semipreparative LC with the 59 60

URL: http://mc.manuscriptcentral.com/gnpl Page 13 of 16 Natural Product Research

1 2 3 same eluent MeCN/H O to afford 4.7 mg of 4, 4.3 mg of 5, 2 mg of 6 and 2.9 mg of 4 2 5 6 7. The separations were monitored using HPLC. 7 8 9 1.4. Radical-Scavenging Activity (DPPH•) TLC Assays 10 • 11 A TLC autographic assay of radical-scavenging activity using the stable DPPH 12 13 radical was applied for extract screening. After application of 100 g of the samples 14 15 on silica gel 60 F Al plates (Merck), development was with Ethyl acetate-formic 16 For254 Peer Review Only 17 18 acid-glacial acetic-water (100:11:11:26) for the MeOH extract. Plates were 19 20 thoroughly dried for complete removal of solvents. A solution of 2,2-diphenyl-1- 21 22 • 23 picrylhydrazyl radical (DPPH , 2 mg/mL in MeOH) was then sprayed. Inhibitors 24 25 appeared as yellow spots against a purple background (Cuendet, Hostettmann, 26 27 Potterat, & Dyatmiko, 1997). 28 29 30 31 1.5. Total Antioxidant Activities 32 33 1.5.1 Antioxidant assays 34 Alkaline phosphatase (ALP) from calf intestinal mucosa (EC.3.1.3.1), 4- 35 36 37 methylumbelliferyl phosphate disodium salt (4-MUP), α,α’-azodiisobutyramidine 38 39 dihydrochloride (AAPH), glycine, magnesium chloride hexahydrate, sodium 40 41 42 hydroxide, sodium chloride and dimethylsulfoxide (DMSO) of microbiological 43 44 quality were purchased from Fluka (Buchs, Switzerland). 2,2-diphenylpicrylhydrazine 45 46 (DPPH•), 2,2’-azinobis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt 47 48 •- 49 (ABTS ) and potassium persulfate were purchased from Sigma-Aldrich (Buchs, 50 51 Switzerland). Fluorescein (FL) was purchased from Siegfried Handel (Zofingen, 52 53 Switzerland). Water solutions were prepared in demineralised and purified water 54 55 56 obtained with the Elix 3 Millipore water purifying system. 57 58 The fluorescence was monitored using a Bio Tek FLX 800 microplate 59 60 fluorescence reader, the absorbance was measured using a Bio-Tek Power wave X

URL: http://mc.manuscriptcentral.com/gnpl Natural Product Research Page 14 of 16

1 2 3 microplate absorbance reader. Both are used with KC4 v3.3 software (Bio Tec 4 5 6 Instruments Inc., Winooski, USA). The dose-response curves were plotted using 7 8 GraphPad prism version 5.00, GraphPad software (San Diego California, USA) 9 10 11 •- 12 1.5.2. ABTS assay 13 The radical scavenging ability was evaluated spectrophotometrically by the ABTS•- 14 15 decolourization, the green radical becoming clear after reduction, in microplates based 16 For Peer Review Only 17 18 on a well-known technique (Re et al., 1999). 19 20 The mass of the compound and the ER50 parameter, namely the ratio of 21 22 •- 23 antioxidant concentration to ABTS radical concentration, both producing an 24 25 absorbance decrease of 50% at 90 min, were determinated by a dose-response curve. 26 27 28 • 29 1.5.3. DPPH assay 30 The capacity of test compounds to scavenge the stable radical 2,2 diphenyl-1- 31 32 picrylhydrazyl (DPPH•) was assessed spectrophotometrically by measuring the 33 34 • 35 decolourization of DPPH thanks to a microplate assay (Brand-Williams, Cuvelier, & 36 37 Berset, 1995). The purple radical becomes yellow DPPH-H upon reduction by the test 38 39 40 compound. The mass necessary and the ER50 parameter, namely the ratio of 41 • 42 antioxidant concentration to DPPH concentration, producing an absorbance decrease 43 44 45 of 50% at 90 min were determinated by a dose-response curve. 46 47 48 1.5.4. ALP assay 49 The antioxidant activity of a compound was determined by its ability to preserve the 50 51 52 catalytic effectiveness of the enzyme alkaline phosphatase (ALP) despite the presence 53 54 of peroxyl radicals generated by 2,2’-asobis(2-methylpropionamidine dihydrochloride 55 56 57 (AAPH). The reaction followed to assess the catalytic activity of ALP was the 58 59 enzymatic dephosphorylation of 4-methylumbelliferyl phosphate (4-MUP) to 60 fluorescent 4-methylumbelliferone (4-MU). Enzymatic hydrolysis rates of 4-MUP

URL: http://mc.manuscriptcentral.com/gnpl Page 15 of 16 Natural Product Research

1 2 3 were determined by a continuous spectrofluorimetric assay. The detailed description 4 5 6 of the method has been reported previously (Bertolini, Novaroli, Carrupt, & Reist, 7 8 2007). The presence of a possible inhibition of compounds on the ALP activity was 9 10 11 verified for each one by replacing the oxidant by glycine buffer. 12 13 The percentage of ALP protection by tested compounds was calculated 14 15 according to the following equation: 16 For Peer Review Only 17 (hasample− ha ox) 18 % ALP protection =100 × 19 (hanon ox− ha ox) 20 21 with the hydrolytic activity of oxidized samples (hasample) and oxidized 22 23 controls (haox) as well as of non oxidized controls (hanon ox). 24 25 The EC50 value representing the antioxidant concentration necessary to protect 26 27 28 50% of ALP activity and the mass of antioxidant compound necessary to protect 50% 29 30 of ALP activity were determinate by a dose-response curve. 31 32 33 1.5.5. ORAC assay 34 35 ORAC assay was carried out with respect to the same conditions as ALP assay. The 36 37 only difference was 186 µL of a fluorescein solution at 6.10-8 M used instead of the 38 39 40 solutions of protein and substrate. The fluorescence was read at λEx 485 ± 20 nm, λEm 41 42 528 ± 20 nm to obtain the fluorescence value of oxidized samples (Fluosample) and 43 44 oxidized controls (Fluo ) as well as non oxidized controls (Fluo ). 45 FLox FLnon ox 46 47 The more efficient was the antioxidant activity of a compound, the higher was 48 49 the percentage of remaining fluorescein. This last one was calculated by the following 50 51 52 equation: 53 54  Fluosample− Fluo FLox  55 % remaining fluorescein =100 ×  FluoFLnonox − Fluo FLox 56   57 The EC50 value representing the antioxidant concentration necessary to protect 58 59 50% of fluorescein and the mass of antioxidant compound to avoid 50% of the 60

URL: http://mc.manuscriptcentral.com/gnpl Natural Product Research Page 16 of 16

1 2 3 fluorescein oxidation, namely a fluorescence loss of 50% were determinate by a dose- 4 5 6 response curve. 7 8 The measurement time of 90 min for the ABTS•- and DPPH• assays was 9 10 11 chosen with respect to the oxidation time in ALP assay. 12 13 14 References 15 Bertolini, F., Novaroli, L., Carrupt, P.A., & Reist, M. (2007). Novel screening assay 16 For Peer Review Only 17 for antioxidant protection against peroxyl radical-induced loss of protein 18 function. Journal of Pharmaceutical Sciences, 96, 2931-2944. doi 19 10.1002/jps.20881 20 Brand-Williams, W., Cuvelier, M.E., & Berset, C. (1995). Use of a free radical 21 method to evaluate antioxidant activity. LWT-Food Science and Technology, 22 28, 25-30. doi:10.1016/S0023-6438(95)80008-5 23 24 Cuendet, M., Hostettmann, K., Potterat, O., & Dyatmiko, W. (1997). Iridoid 25 glucosides with free radical scavenging properties from Fagraea blumei. 26 Helvetica Chimica Acta, 80, 1144-1152. doi: 10.1002/hlca.19970800411 27 Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M., & Rice-Evans, C. 28 (1999). Antioxidant activity applying an improved ABTS radical cation 29 30 decolorization assay. Free Radical Biology and Medicine, 26, 1231-1237. 31 doi:10.1016/S0891-5849(98)00315-3 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/gnpl