Using Esterase and Laccase Enzymes to Derivatize Bioactive Plant Phenolics for Altered Chemistry

Mohammed Sherif

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy

Department of Cell and Systems Biology University of Toronto

If someone said © Copyright by Mohammed Sherif 2015 Using Esterase and Laccase Enzymes to Derivatize Bioactive Plant Phenolics for Altered Chemistry

Mohammed Sherif

Doctor of Philosophy

Department of Cell and Systems Biology University of Toronto

2015

ABSTRACT Plant phenolics have notable antioxidant activity and there is potential to improve their action by chemical modification. Two enzyme classes carry out reactions that can act on the hydroxyl moiety of phenolics. Esterase enzymes can be used in non-aqueous solvents to esterify a long chain acyl group onto the phenolic compound. Laccase enzymes can be used to form phenoxy radicals that can then couple to form larger molecular weight oligomers. Both enzymatic modifications may produce a new antioxidant with altered chemistry.

One archaeal esterase (AF1753) from Archaeoglobus fulgidus and one bacterial esterase

(PP3645) from Pseudomonas putida were assayed for activity in organic solvents. Both enzymes catalyzed hydrolysis of phenyl acetate and vinyl acetate in 98:2 (v/v) (t-amyl alcohol):buffer; with continued activity up to 96 h of reaction. However, the enzymes were not able to catalyze transesterification of 4’-hydroxyacetophenone with vinyl acetate in 9:1 (v/v) cyclohexane:(t-amyl alcohol), which was not explained by enzyme inactivation during lyophilization. Still, alanine scanning mutagenesis revealed that R37A substitution improved activity of AF1753 on long-chain p-nitrophenyl (pNP) esters.

A multicopper oxidase (SCO6712) from Streptomyces coelicolor displayed activity on a variety of phenolics including caffeic acid, ferulic acid, resveratrol, quercetin, morin,

ii kaempferol and myricetin. Among the products formed by action on flavonols were dimers of quercetin, morin, and myricetin. Quercetin and myricetin dimers showed longer retention time on reversed phase chromatography. All three dimers could be detected by 5 min of reaction but depleted by 3 h and 24 h. The TRAP and FRAP antioxidant activity of the whole reaction mixture of modified quercetin, morin, and myricetin decreased, as starting phenolic was depleted over 24 h. Accordingly, mass spectrometry was used to shed light on the molecular structure of the dimers produced from quercetin and myricetin. In both cases, mass spectrometric analyses ruled out dimer formation through the A ring of each monomer. For myricetin, the most likely linkage structure was determined to be between either two B rings or a B ring with a C ring. These predicted linkage positions are in agreement to those observed for quercetin dimers previously extracted from natural plant sources.

iii Acknowledgments

I would like to thank my supervisor Professor Emma Master for giving me the opportunity to work on this project and providing guidance. I will also thank the members of my Supervisory Committee, Professor Brad Saville and Professor Dinesh Christendat, for their advice throughout the project. I thank collaborators who gave assistance on this project including members of the group Structural Proteomics in Toronto (SPiT) in Toronto and members from Natural Resources Canada (NRCan) in Sault Ste. Marie. BioZone administration was helpful in making sure things ran smoothly. Colleagues in the Master lab helped with a lot of theoretical and technical aspects of the project. I thank my family for their support over the course of my PhD.

iv Table of Contents

LIST OF TABLES ...... VIII

LIST OF FIGURES ...... IX

LIST OF ABBREVIATIONS ...... XI

CHAPTER 1. OVERVIEW ...... 1

CHAPTER 2. LIERATURE SURVEY...... 3 2.1. Plant phenolic compounds ...... 3 2.1.1. Types and distribution ...... 3 2.1.2. Biosynthesis and role in plants ...... 7 2.1.3. Health benefits of phenolics ...... 12 2.1.4. Examples of antioxidant activity ...... 12 2.1.5. Phenolic antioxidant activity for food preservation ...... 14 2.1.6. Structure-functional correlations among phenolics with antioxidant properties 15 2.1.7. In vitro measurement of antioxidant activity ...... 17 2.1.8. Solubility considerations for antioxidant activity ...... 18 2.2. Derivatization of plant phenolics ...... 20 2.2.1. Enzymatic strategies used in plant phenolic derivatization ...... 20 2.2.2. Increasing hydrophilicity of phenolic compounds ...... 21 2.2.3. Increasing lipophilicity of phenolic compounds ...... 23 2.3. Esterases/lipases ...... 26 2.3.2. Structural features ...... 27 2.3.3. Catalytic mechanism ...... 29 2.3.4. Transesterification reactions ...... 30 2.3.5. Applied Use ...... 33 2.4. Laccases ...... 34 2.4.2. Structural features ...... 35 2.4.3. Catalytic mechanism ...... 37 2.4.4. Effect of Redox potential ...... 41 2.4.5. Applied use ...... 42 2.5. Research Hypotheses and Objectives...... 45

CHAPTER 3. CHARACTERIZATION OF SOLVENT-TOLERANT CARBOXYLESTERASES WITH ARYLESTERASE ACTIVITY ...... 46 3.1. Introduction ...... 47 3.2. Materials and methods ...... 49 3.2.1. Gene cloning and protein purification ...... 49 3.2.2. Hydrolytic activity of esterases AF1753 and PP3645 in t-amyl alcohol/water (98:2, v/v) ...... 49 3.2.3. Transesterification activity of esterases AF1753 and PP3645 in t-amyl alcohol/cyclohexane (1:9, v/v) ...... 50 3.2.4. Protein structure modeling and site-directed mutagenesis of esterase AF1753 .. 51

v 3.2.5. Activity of wild type and mutant AF1753 esterases on varying chain-length pNP esters ...... 51 3.3. Results and discussion ...... 52 3.3.1. Hydrolytic activity of esterases AF1753 and PP3645 in t-amyl alcohol/water (98:2, v/v) ...... 52 3.3.2. Transesterification activity of esterases AF1753 and PP3645 in t-amyl alcohol/cyclohexane (1:9, v/v) ...... 55 3.3.3. Activity of wild type and mutant AF1753 esterases on varying chain-length pNP esters ...... 61 3.4. Conclusions ...... 66

CHAPTER 4. BIOCHEMICAL STUDIES OF THE MULTICOPPER OXIDASE (SMALL LACCASE) FROM STREPTOMYCES COELICOLOR USING BIOACTIVE PHYTOCHEMICALS AND SITE-DIRECTED MUTAGENESIS ...... 67 4.1. Introduction ...... 68 4.2. Materials and methods ...... 69 4.2.1. Gene cloning and protein purification ...... 69 4.2.2. Site-directed mutagenesis ...... 70 4.2.3. Copper content analysis of wild-type and mutant SCO6712 laccases ...... 70 4.2.4. Substrate profile of wild-type SCO6712 and the Ser292Ala mutant laccases ...... 70 4.2.5. Kinetics of wild-type and Ser292Ala laccases on select substrates ...... 72 4.2.6. Docking of 2,6-dimethoxyphenol substrate to wild type and Ser292Ala SCO6712 laccase ...... 72 4.3. Results and discussion ...... 72 4.3.1. Effect of microaerobic cultivation on copper content and activity of SCO6712 laccase ...... 72 4.3.2. Substrate profile of wild-type SCO6712 laccase ...... 73 4.3.3. Kinetics of wild-type SCO6712 laccase on select substrates ...... 77 4.3.4. Site-directed mutagenesis ...... 79 4.4. Conclusions ...... 82

CHAPTER 5. CHARACTERIZATION OF PRODUCT FORMATION FROM ENZYMATICALLY OXIDIZED PLANT PHENOLICS AND ASSAY OF ANTIOXIDANT ACTIVITY ...... 83 5.1. Introduction ...... 84 5.2. Materials and methods ...... 85 5.2.1. HPLC-MS analysis of flavonol products after SCO6712 laccase treatment ...... 85 5.2.2. HPLC-MS analysis of flavonol dimer presence over laccase reaction time...... 85 5.2.3. Total radical-trapping antioxidant parameter (TRAP) assay of whole laccase reaction mixture ...... 85 5.2.4. Ferric reducing antioxidant power (FRAP) assay of whole laccase reaction mixture ...... 86 5.2.5. HPLC-MS/MS analysis of quercetin dimer and myricetin dimer...... 86 5.3. Results and discussion ...... 87 5.3.1. HPLC-MS analysis of flavonol products after SCO6712 laccase treatment ...... 87 5.3.2. Antioxidant assay of whole laccase reaction mixture ...... 109 5.3.3. HPLC-MS/MS analysis of quercetin dimer and myricetin dimer...... 120

vi 5.4. Conclusions ...... 124

CHAPTER 6. DISCUSSION ...... 125

CHAPTER 7. FUTURE RESEARCH ...... 137 7.1. Further characterization of esterases for transesterification potential, continuing work of Chapter 3 ...... 137 7.2. Further assessment of biochemical potential of laccase SCO6712, continuing work of Chapter 4 ...... 138 7.3. Further examination of bioactivity of flavonol dimers, continuing work of Chapter 5 ...... 139

REFERENCES ...... 141

APPENDIX 1. SUPPLEMENTAL INFORMATION FOR CHAPTER 3 ...... 159

APPENDIX 2. SUPPLEMENTAL INFORMATION FOR CHAPTER 4 ...... 162

APPENDIX 3. SUPPLEMENTAL INFORMATION FOR CHAPTER 5 ...... 163

APPENDIX 4. TRAP ANTIOXIDANT ASSAY USING LINOLEIC ACID IN PLACE OF DCFH ...... 171

APPENDIX 5. QUANTIFYING SOLUBILITY OF QUERCETIN MONOMER FOR COMPARISON TO QUERCETIN DIMER ...... 174

vii LIST OF TABLES

Table 2.1. Different classes of plant phenolic compounds (Balasundram et al., 2006)...... 3 Table 2.2. Major food sources of plant phenolic compounds (Manach et al., 2004)...... 6 Table 2.3. Tree/shrub sources of plant phenolic compounds...... 6 Table 2.4. Popular antioxidant activity assays...... 17 Table 4.1. Kinetic parameters of wild type SCO6712 and the Ser292Ala variant enzyme...... 78 Table 5.1. HPLC-MS data for quercetin, morin, myricetin, and their dimers produced after enzymatic reaction...... 88 Table 5.2. HPLC-MS data for representative intermediate molecular weight products present in late time point enzymatic reactions of quercetin, morin, and myricetin...... 99 Table 5.3. Top 10 products from late time point enzymatic reactions of quercetin, morin, and myricetin...... 100 Table 5.4. Top 10 products from 24 h reaction of quercetin without and with laccase enzyme...... 105

Table A1.1. Sequences of primers used for construction of esterase AF1753 point mutants. . 159 Table A2.1. Sequences of primers used for construction of laccase SCO6712 point mutants. 162 Table A3.1. HPLC-MS data for kaempferol and product produced after enzymatic reaction. 163 Table A3.2. Fragment ions observed for quercetin dimer #1 after HPLC-MS/MS...... 163 Table A3.3. Fragment ions observed for myricetin dimer #2 after HPLC-MS/MS...... 164 Table A3.4. Fragment ions observed for non-enzymatically produced quercetin dimer after HPLC-MS/MS...... 166 Table A3.5. Fragment ions observed for non-enzymatically produced myricetin dimer after HPLC-MS/MS...... 167

viii LIST OF FIGURES

Figure 2.1. Structures of some members of A) hydroxybenzoic acids, B) hydroxycinnamic acids, and C) stilbenes; and D) basic skeleton of all flavonoids, and specifically flavones and flavonols...... 5 Figure 2.2. Key steps in biosynthesis pathways for production of hydroxybenzoic acids and hydroxycinnamic acids...... 10 Figure 2.3. Key steps in biosynthesis pathway for production of stilbenes...... 11 Figure 2.4. Key steps in the biosynthesis pathway for production of flavones and flavonols. .. 11 Figure 2.5. Reactions involved in scavenging of peroxyl radicals by a phenolic antioxidant. ... 13 Figure 2.6. Structures of some classes of flavonoids...... 16 Figure 2.7. Possible routes to adding long-chain alkyl groups to phenolic compounds to increase lipophilicity of the phenolic...... 24 Figure 2.8. Naturally occurring dimers (n=1), trimers (n=2), and tetramers (n=3) made of successive epicatechin molecules linked to catechin...... 25 Figure 2.9. Reaction catalyzed by esterases and lipases...... 27 Figure 2.10. Prototypical α/β hydrolase fold structure...... 28 Figure 2.11. Catalytic mechanism of esterases and lipases...... 31 Figure 2.12. Reaction catalyzed by laccases...... 35 Figure 2.13. Structure of laccase TvL from Trametes versicolor. A) One of the cupredoxin domains of TvL (the protein has three such domains)...... 37 Figure 2.14. Proposed catalytic mechanisms of oxidation by laccase enzymes...... 40 Figure 2.15. Structures of closely related phenols...... 41 Figure 3.1. Known sites of action of feruloyl esterases and arylesterases...... 48 Figure 3.2. Hydrolysis reactions of A) vinyl acetate and B) phenyl acetate carried out in t-amyl alcohol/water (98:2, v/v) using esterases AF1753 and PP3645...... 54 Figure 3.3. Transesterification reaction of vinyl acetate and 4’-hydroxyacetophenone carried out in different solvent mixtures of t-amyl alcohol/co-solvent (1:9, v/v) using commercial lipase PS (from Amano)...... 57 Figure 3.4. Transesterification reaction of vinyl acetate and 4’-hydroxyacetophenone carried out in t-amyl alcohol/cyclohexane (1:9, v/v) using recombinant esterase AF1753 and PP3645...... 58 Figure 3.5. Hydrolysis reaction of ester substrate in transesterification reaction-mix...... 59 Figure 3.6. Activity of wild type and mutants of AF1753 on pNP substrates...... 63 Figure 3.7. Images of predicted protein structure of AF1753...... 65 Figure 4.1. Substrate selectivity of SCO6712...... 76 Figure 4.2. Chemical structures of natural bioactive phenolic substrates...... 77 Figure 4.3. Ribbon image of 2,6-dimethoxyphenol (2,6-DMP) docked in silico to binding pocket of SCO6712 A) wild type enzyme and B) Ser292Ala mutant...... 81 Figure 5.1. HPLC-MS chromatograms and m/z spectra for 20 min reaction samples with laccase enzyme for A) quercetin, B) morin, and C) myricetin...... 91 Figure 5.2. HPLC-MS chromatogram and m/z spectrum for 20 min reaction sample with laccase enzyme for kaempferol...... 91 Figure 5.3. HPLC-MS chromatograms for 5 min reaction sample with laccase enzyme for (A) quercetin, (B) morin, and (C) myricetin...... 94 Figure 5.4. Mass spectrum for myricetin dimer of m/z 635 (from peak D3 in chromatogram of Fig. 5.3. C); refer to Table 5.1) for 5 min reaction sample with laccase enzyme...... 94

ix Figure 5.5. HPLC-MS analysis of flavonol dimers from A) quercetin (dimer #1), B) morin, C) myricetin (dimer #2), and D) myricetin (dimer #3)...... 98 Figure 5.6. Proposed reaction scheme for production of some of the quercetin oxidation degradation products...... 104 Figure 5.7. Structures of the isomers quercetin and morin...... 108 Figure 5.8. TRAP antioxidant activity of laccase reaction-mixes from A) quercetin, B) morin, and C) myricetin reactions...... 113 Figure 5.9. TRAP antioxidant activity of laccase reaction-mixes from A) quercetin, B) morin, and C) myricetin reactions...... 117 Figure 5.10. FRAP antioxidant activity of laccase reaction-mixes from A) quercetin, B) morin, and C) myricetin reactions...... 119 Figure 5.11. Fragmentation patterns of flavonols in positive ion mode tandem mass spectrometry...... 121 Figure 5.12. Proposed fragment structures of quercetin dimer #1 (refer Table 5.1) after tandem mass spectrometry...... 122 Figure 5.13. Proposed fragment structures of myricetin dimer #2 (refer Table 5.1) after tandem mass spectrometry...... 123 Figure 6.1. Two possible mechanisms of quercetin dimer formation...... 130 Figure 6.2. Mechanism of antioxidant antagonism proposed by Peyrat-Maillard et al. (2003)...... 133 Figure 6.3. Mechanism of antioxidant antagonism that involves reaction of quercetin radicals with laccase-generated quercetin products...... 135

Figure A1.1. Protein purification of AF1753 mutants...... 161 Figure A3.2. Remaining phenolic monomer in laccase reactions, as measured by UV-Vis spectrophotometry and by intensity of HPLC-MS peak...... 170 Figure A5.1. The maximum amount of quercetin that can be dissolved in A) 50 mM sodium phosphate buffer (pH 7.4) with 0.1 M NaCl and B) 1-octanol...... 177

x LIST OF ABBREVIATIONS

2,3-DHB - 2,3-dihydroxybenzoic acid 2,6-DMP - 2,6-dimethoxyphenol AAPH - 2,2′-azobis(2-methylpropionamidine) ABTS - 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) Ala - alanine AMVN - 2,2’-azobis (2,4-dimethylvaleronitrile) Arg - arginine DCFH - 2′,7′-dichlorofluorescin DCFH-DA - 2′,7′-dichlorofluorescin diacetate DHA docosahexaenoic acid DMSO - dimethylsulfoxide DPPH - 2,2-diphenyl-1-picrylhydrazyl EDTA - ethylenediaminetetraacetic acid EPA - eicosapentaenoic acid ET - electron transfer FRAP - ferric reducing antioxidant power Glu - glutamate HAT - hydrogen atom transfer HPLC-MS - high performance liquid chromatography-mass spectrometry LDL - low density lipoprotein L-DOPA - 3,4-dihydroxy-L-phenylalanine MCO - multicopper oxidase N-HPI - N-hydroxyphthalimide NMR - nuclear magnetic resonance pNP - p-nitrophenol ROS - reactive oxygen species Ser - serine TPTZ - 2,4,6-tris(2-pyridyl)-s-triazine TRAP - total radical-trapping antioxidant parameter

xi 1

CHAPTER 1. OVERVIEW

Plant materials are a rich source of bioactive phenolic compounds. Phenolics are a common constituent of the human diet via fruits, vegetables, and beverages (Balasundram et al., 2006) but they can also be derived from forest sources such as with the monolignols, stilbenes, lignans, and certain types of flavonoids (Stevanovic et al., 2009). Phenolic compounds have well known antioxidant activities that can find use in prevention of diseases affecting human health, and in food preservation against oxidative decay (Balasundram et al., 2006). There is potential to increase the protective effect of these compounds towards lipid targets by increasing their hydrophobicity through the addition of alkyl groups. Hydrophobicity of the phenolics might also be increased by increasing their molecular weight through oligomerization. Such modifications may increase the miscibility of the bioactive compound in emulsified food oils, imparting a preservative effect (Frankel et al., 1994); or in low density lipoproteins (LDL), thereby reducing the frequency of health problems such as atherosclerosis (Vafiadi et al., 2008). The overall objective of this research project was to alter the chemistry of phenolic compounds by enzymatic modification (for example to increase hydrophobicity of the phenolics), while maintaining antioxidant activity of the phenolic. Towards this goal two enzyme types were investigated for their potential to modify phenolics. The first enzyme class, esterases, were used to try to esterify alkyl chains onto phenolic compounds and the second enzyme class, laccase (a type of multicopper oxidase), was used to oxidatively dimerize phenolic compounds via radical coupling reactions. The esterase reaction must be performed in an organic solvent to avoid hydrolytic breakdown of the desired ester product. By contrast, laccase reactions can be carried out in aqueous conditions but because of radical delocalization it is difficult to know in advance which products will be formed from the wide array of potential products. The feasibility of esterase-mediated and laccase-mediated modification of bioactive phenolics was investigated, following the literature survey (Chapter 2), in subsequent chapters. First, the chosen esterases (one bacterial (Pseudomonas putida) and one archaeal (Archaeoglobus fulgidus)) were assessed for their activity in predominantly organic solvent media. The enzymes were hydrolytically active in the organic solvent-water mixture but did not catalyze transesterification reactions. Therefore, attention was focused on the second enzymatic approach, i.e. laccase oxidation followed by oxidative coupling of phenolics to produce

2 increased molecular weight products as a means of modifying phenolic chemistry. In this case, initial work focused on evaluating determinants of activity of a bacterial laccase from Streptomyces coelicolor on a range of phenolic compounds from diverse classes, including hydroxycinnamic acids, stilbenes, flavonols, and flavones. As laccase activity was observable on the well-known antioxidant flavonols (among other compounds), the products from these reactions were analyzed. Presence of dimer products from the flavonols quercetin, morin, and myricetin was examined using HPLC-MS. Tandem mass spectrometry was used to gain initial information about the structure of the dimers of quercetin and myricetin. The change in antioxidant activity resulting from laccase action was assayed using the total radical-trapping antioxidant parameter (TRAP) assay and the ferric reducing antioxidant power (FRAP) assay on whole laccase reaction mixtures.

Summary of scholarly contributions Peer reviewed publications: Sherif, M., Waung, D., Korbeci, B., Mavisakalyan, V., Flick, R., Brown, G., Abou-Zaid, M., Yakunin, A. F., & Master, E. R. (2013). Biochemical studies of the multicopper oxidase (small laccase) from Streptomyces coelicolor using bioactive phytochemicals and site-directed mutagenesis. Microbial Biotechnology, 6(5), 588-597.

Manuscripts in preparation: Sherif, M., Wang, L., Tchigvintsev, A., Brown, G., Mavisakalyan, V., Tillier, E. R. M, Savchenko, A. V, Master, E. R, & Yakunin, A. F. Solvent-tolerant and thermophilic carboxylesterase with arylesterase activity from Archaeoglobus fulgidus. Sherif, M., Qazi, S., Abou-Zaid, M., & Master, E. R. Identification of products and antioxidant activity of reaction mixtures from treatment of four flavonols with a multicopper oxidase SLAC (small laccase) from Streptomyces coelicolor. Qazi, S., Sherif, M., Master, E. R, & Abou-Zaid, M. Tandem mass spectrometric and NMR structural characterization of quercetin dimer produced by multicopper oxidase treatment of quercetin.

CHAPTER 2. LIERATURE SURVEY

2.1. Plant phenolic compounds

2.1.1. Types and distribution

Plant phenolic compounds constitute secondary metabolites and are among the most prevalent phytochemicals, appearing in both food and non-food sources (reviewed in Balasundram et al., 2006; Manach et al., 2004). A structural categorization of plant phenolics leads to the following classes based on the configuration of the carbon skeleton (Table 2.1)

(carbon skeleton in brackets): simple phenolics (C6), hydroxybenzoic acids (C6-C1), phenylacetic acids (C6-C2), hydroxycinnamic acids (C6-C3), quinones (diverse carbon skeleton), xanthones (C6-C1-C6), stilbenes (C6-C2-C6), flavonoids (C6-C3-C6), lignans ((C6-C3)2), biflavonoids ((C6-C3-C6)2), lignins ((C6-C3)n), and tannins (diverse carbon skeleton) (Balasundram et al., 2006). The flavonoids are further subdivided into the flavones, isoflavones, flavonols, flavanones, anthocyanidins, and flavanols (Manach et al., 2004). Of the plant phenolics, the most structurally diverse group are the flavonoids (Balasundram et al., 2006) and of the flavonoids, the flavonols are the most common in foods (Manach et al., 2004). In many cases, representatives of these classes of plant phenolics can be found conjugated to monomeric and oligomeric sugars.

Table 2.1. Different classes of plant phenolic compounds (Balasundram et al., 2006). Phenolic class Carbon skeleton Simple phenolics C6 Hydroxybenzoic acids C6-C1 Phenylacetic acids C6-C2 Hydroxycinnamic acids C6-C3 Quinones Diverse Xanthones C6-C1-C6 Stilbenes C6-C2-C6 Flavonoids C6-C3-C6 Lignans (C6-C3)2 Biflavonoids (C6-C3-C6)2 Lignins (C6-C3)n Tannins Diverse

The higher molecular weight classes mentioned above are oligomeric/polymeric forms of the lower molecular weight classes. Lignans are dimerized forms, while lignin is a large complex polymer, of the alcohols of hydroxycinnamic acids. The stilbenes are known to exist in oligomeric forms of the simplest structural units (C6-C2-C6) of their class (Quideau et al., 2011). As the name suggests, biflavonoids are dimeric versions of flavonoids. The tannins can be divided into the condensed tannins (carbon skeleton of (C6-C3-C6)n) and hydrolyzable tannins. Condensed tannins are oligomeric forms of flavanols while hydrolyzable tannins are composed of monomeric and polymeric hydroxybenzoic acid units esterified onto sugars (Quideau et al., 2011). The remainder of this section (Section 2.1.1) and the subsequent section (Section 2.1.2) will focus on the distribution (dietary and non-dietary), biosynthesis, and role in plants of hydroxybenzoic acids, hydroxycinnamic acids, stilbenes, and two of the flavonoids (flavones and flavonols) (Fig. 2.1). These classes of phenolics were chosen because they contain among the most well-studied and most effective antioxidant compounds. Hydroxybenzoic acids are not widely found in plant material consumed by people. The few major sources in the human diet include in certain red fruits, black radish, onion, and tea (Table 2.2) (reviewed in Manach et al., 2004). One of the more studied of the hydroxybenzoic acids, gallic acid, can be found in esterified form in the bark of Quercus stenophylla (Nishimura et al., 1984), in the flowers of Tamarix nilotica (reviewed in Van Sumere, 1989), in maple species (for example as a glycoside conjugate in leaves of Acer rubrum (Abou-Zaid & Nozzolillo, 1999) and as a methyl ester in leaves of Acer rubrum, Acer saccharinum, and Acer saccharum (Abou-Zaid et al., 2009)), (Table 2.3) and various other plant species (for an extensive list of plants containing gallic acid and other phenolics see Harborne et al., 1990). More recently, it was isolated from aerial plant parts of Pelargonium reniforme (Latté et al., 2008). Aside from being obtained from plant material after comparatively gentle solvent extraction, hydroxybenzoic acids, such as vanillic acid and syringic acid, can also be obtained upon hydrolytic treatment of lignocellulosic materials, due to oxidation and breakdown of the lignin polymer (reviewed in Garrote et al., 2004).

O OH O OH O OH O OH

OH HO OH H CO OCH 3 3 OH OH OH

salicylic acid protocatechuic acid gallic acid syringic acid

O OH O OH O OH O OH

OH OCH3 H3CO OCH3 OH OH OH OH

p-coumaric acid caffeic acid ferulic acid sinapic acid

C) OH OCH 3 OH

OH OCH 3 OH HO HO resveratrol pterostilbene pinosylvin

D) O O O OH O O flavonoids flavones flavonols

Figure 2.1. Structures of some members of A) hydroxybenzoic acids, B) hydroxycinnamic acids, and C) stilbenes; and D) basic skeleton of all flavonoids, and specifically flavones and flavonols.

Table 2.2. Major food sources of plant phenolic compounds (Manach et al., 2004). Phenolic class Major food sources Hydroxybenzoic acids red fruits black radish onion tea Hydroxycinnamic acids blueberry kiwi plum cherry apple Stilbenes grape Flavones parsley celery Flavonols onion leek broccoli blueberry

Table 2.3. Tree/shrub sources of plant phenolic compounds. Compound Tree species Reference Hydroxybenzoic acids Quercus stenophylla Nishimura et al. (1984) Tamarix nilotica Van Sumere (1989) Pelargonium reniforme Latté et al. (2008) Acer spp. Abou-Zaid et al. (2009) Hydroxycinnamic acids Tsuga heterophylla Harborne (1990) Catalpa ovata Stilbenes Veratrum formosanum Stevanovic et al. (2009) Picea abies Pinus sibirica Flavones and flavonols Eucalyptus spp. Stevanovic et al. (2009) Crataegus sp. Pinus spp. Abou-Zaid & Nozolillo (1991)

Hydroxycinnamic acids are mostly found in conjugated forms and the four most common compounds are conjugated forms of p-coumaric, caffeic, ferulic, and sinapic acids (Manach et al., 2004). As opposed to the hydroxybenzoic acids, the hydroxycinnamic acids can be found in a variety of food sources, and highest amounts have been found in blueberries, kiwis, plums, cherries, and apples (Table 2.2) (Manach et al., 2004). Similar to the hydroxybenzoic acids, the hydroxycinnamic acids can also be obtained upon hydrolytic treatment of lignocellulosic materials (Garrote et al., 2004). In addition to making up the lignin

7 polymer (in the alcohol form), the hydroxycinnamic acid ferulic acid (and its dimers) can be found esterified onto hemicelluloses (Manach et al., 2004). Stilbenes are not abundant in the human diet except from grapes and their juices (Table 2.2) (Manach et al., 2004). The most widely studied stilbene is resveratrol. Resveratrol and its glucoside have been found in Veratrum formosanum and in the bark of Picea abies, respectively, and stilbenes can notably be found from the knotwood extracts of pines (Table 2.3) (reviewed in Stevanovic et al., 2009). Flavonoids are the most structurally diverse phenolic compound in plants. Among the most well-known of the flavonoids is the flavonol quercetin because of its very strong antioxidant activity. Flavones are chiefly found in parsley and celery in the human diet (Table 2.2) (Manach et al., 2004). On the other hand, flavonols are more widely prevalent and can be found in onions, leeks, broccoli, blueberries, and other food sources (Table 2.2) (Manach et al., 2004). Additionally, flavonols can be found in leaves of forest trees including birches and eucalyptus (Stevanovic et al., 2009). Another source of flavonols (and flavones) is in trees of the family Crataegus, with 13 flavonols and 20 flavones previously identified (Stevanovic et al., 2009). Furthermore, flavonol glycosides were observed from needles of pine trees such as Pinus banksiana (Table 2.3) (Abou-Zaid & Nozolillo, 1991).

2.1.2. Biosynthesis and role in plants

The biosynthesis of plant phenolic compounds can be traced back to the shikimate pathway and the polyketide pathways with the polyketide pathway providing precursors for production of simple phenolics, while the shikimate pathway provides precursors for the other phenolic types (Harborne 1989). Starting from shikimate, phenylalanine is produced by the shikimate pathway. Phenylalanine is then deaminated to produce cinnamic acid, which is then hydroxylated to produce p-coumaric acid. As such, cinnamic acid is the first precursor to production of the other plant phenolics (Fig. 2.2) (Harborne, 1989; Dewick, 1995). Hydroxybenzoic acids are thought to be produced from cinnamic acids by removal of an acetate unit (Fig. 2.2) (Gross, 1992). However, based on tracer experiments with radiolabelled carbon, it has also been proposed that gallic acid biosynthesis could proceed via direct dehydrogenation of shikimic acid without going through a pathway involving cinnamic acid production (Gross, 1992). The hydroxycinnamic acids are formed via aromatic substitution of cinnamic acid by

8 undergoing sequential hydroxylations and methylations (Harborne, 1989; Dewick, 1995). Stilbene biosynthesis occurs by reaction of three malonyl-CoA molecules with p-coumaroyl- CoA, followed by a decarboxylation accompanied by cyclization (Fig. 2.3) (Dewick, 1995). Similar to the stilbenes, for the flavonoids in general, their biosynthesis starts by reaction of three malonyl-CoA molecules with p-coumaroyl-CoA followed by a cyclization to produce the flavanones, resulting in the basic carbon skeleton of all flavonoids (Fig. 2.4) (Heller & Forkmann, 1994; Dewick, 1995). Flavones derive from flavanones by formation of a double bond between C-2 and C-3 while flavonols are formed by first hydroxylating position 3 of flavanones followed by formation of a double bond between C-2 and C-3 (Heller & Forkmann, 1994; Dao et al., 2011). Plant phenolic compounds have been postulated to have a diverse array of functions, some representative examples being: the role of benzoic acids in photosynthesis (Van Sumere, 1989); the role of ferulic acid in regulating germination of barley seeds (Van Sumere, 1989); the role of stilbenes as antimicrobial compounds (Gorham, 1989); the role of flavones and/or flavonols in 1) protection from UV light, insects, and microorganisms, 2) hormonal control, 3) enzyme inhibition, and 4) attracting pollinators (Markham, 1989).

- O H2C COO O PEP - OH HO O PO 4 DAHPS HO DHQS DHQD + O HO O OH O OH O4P OH OH E4P O4P OH DAHP 3-dehydroquinate

- - - - O O O O O O O O

SDH SK EPSPS CH2 O O OH HO OH O P OH O P O 4 4 - OH OH OH OH O

3-dehydroshikimate shikimate shikimate 3-phosphate EPSP

O O - - - - - O O- O O O O O O O O H N O NH2 2 CH2 CS CM PAT ADT O O - OH O OH OH chorismate prephenate arogenate phenylalanine

- O OH O O O OH PAL C4H

HO OH

OH OH cinnamic acid p-coumaric acid gallic acid

Figure 2.2. Key steps in biosynthesis pathways for production of hydroxybenzoic acids and hydroxycinnamic acids. The shikimate pathway produces chorismate which goes on to produce phenylalanine. Phenylalanine is metabolized to yield hydroxycinnamic acids and hydroxybenzoic acids. Hydroxybenzoic acids can also be produced from shikimate pathway intermediates without going through phenylalanine production. Abbreviations for intermediates: PEP, phosphoenol pyruvate; E4P, erythrose-4-phosphate; DAHP, 3-deoxy-D-arabino- heptulosonate-7-phosphate; EPSP, 5-enolpyruvylshikimate-3-phosphate. Abbreviations for enzymes above arrows: DAHPS, 3-deoxy-D-arabino-heptulosonate-7-phosphate synthase; DHQS, 3-dehydroquinate synthase; DHQD, 3-dehydroquinate dehydratase; SDH, shikimate dehydrogenase; SK, shikimate kinase; EPSPS, 5-enolpyruvylshikimate-3-phosphate synthase; CS, chorismate synthase; CM, chorismate mutase; PAT, prephenate aminotransferase; ADT, arogenate dehydratase; PAL, phenylalanine ammonia-lyase; C4H, cinnamic acid 4-hydroxylase.

O SCoA O HO OH O SCoA CO O AoCSOC O 2 + 3x

O STS STS OH

OH malonyl-CoA p-coumaroyl- OH OH CoA stilbene (e.g.

resveratrol

Figure 2.3. Key steps in biosynthesis pathway for production of stilbenes. p-coumaroyl-CoA comes from the phenylpropanoid biosynthetic pathway. STS indicates stilbene synthase enzyme.

O SCoA HO OH HO OH SCoA O O CHS O CHI + 3x O OH O OH malonyl-CoA OH

p-coumaroyl- OH OH CoA F3H flavanone (e.g. naringenin) HO OH HO OH FNS O O HO OH O O HO FLS HO O O

OH OH flavonol (e.g. dihydrokaempferol OH kaempferol) flavone (e.g.

apigenin) Figure 2.4. Key steps in the biosynthesis pathway for production of flavones and flavonols. p- coumaroyl-CoA comes from the phenylpropanoid biosynthetic pathway. Enzyme abbreviations: CHS, chalcone synthase; CHI, chalcone isomerase; FNS, flavone synthase; F3H, flavanone 3- hydroxylase; FLS, flavonol synthase.

2.1.3. Health benefits of phenolics

Plant-derived phenolic compounds are implicated in a wide array of health benefits. This includes benefits against cardiovascular disease, neurodegenerative disease, cancer, and diabetes (reviewed in Scalbert et al., 2005). However, in vitro findings can not always be translated into similar in vivo effects and tests from different labs do not always give the same results (Scalbert et al., 2005). Atherosclerosis has been observed to be inhibited by consumption of food phenolics and, based on animal studies, it is thought that the phenolics mediate their effects by reducing oxidation and uptake of low density lipoprotein (LDL) by macrophages (Kaplan et al., 2001; Miura et al., 2001). However, human studies have shown mixed results, with some studies showing that consumption of tea protects against ex vivo oxidation of LDL (Ishikawa et al., 1997) while other studies showed no such benefit of tea consumption (Van het Hof et al., 1997). Animal models have also shown anticarcinogenic effects of food phenolics but the doses used in such experiments are usually much larger than typical consumption levels in the human diet, making it difficult to correlate epidemiological results to these animal models (Scalbert et al., 2005). The importance of dosage is further highlighted by the fact that low doses (less than 10 µM) of epigallocatechin gallate was found to be neuroprotective in cell culture models using the neurotoxin 6-hydroxydopamine, while higher doses of epigallocatechin gallate had cytotoxic effects (Levites et al., 2002). Use of plants by indigenous peoples for treatment of diabetes has been documented and animal model studies have also shown antidiabetic effects of plant extracts containing phenolics (Scalbert et al., 2005; Giordani et al., 2015). It is thought that one of the mechanisms by which the plant compounds reduce diabetes is through inhibition of α- glucosidase enzymes, which normally break down carbohydrates so that the sugars can be absorbed in the gut (Giordani et al., 2015).

2.1.4. Examples of antioxidant activity

Plant phenolics are well-known for their antioxidant activity, which may, in some cases, partially mediate their other health effects (reviewed in Scalbert et al., 2005). The antioxidant activity of plant phenolics is due to reaction with free radicals, but may also involve inhibition of enzymes and chelation of metal ions (Huang et al., 2005). In the case of reacting with free

13 radicals, the phenolic antioxidant sacrificially becomes oxidized to a relatively stable radical. Free radical scavenging may occur by the phenolic transferring a hydrogen atom (hydrogen atom transfer (HAT)) or by the phenolic transferring an electron (electron transfer (ET)) followed by reversibly transferring a proton (Fig. 2.5) (Wright et al., 2001). The bond dissociation enthalpy of the hydroxyl groups on phenolics will influence hydrogen atom transfer whereas the ionization potential is important for determining electron transfer (Wright et al., 2001). Also, in buffer solutions, the phenolic compound can exist in different protonation states depending on the pH. Under more basic conditions the phenolic hydroxyls can be deprotonated, and in this case the phenolic will scavenge radicals by electron transfer that is preceded by proton loss (Fig. 2.5) (Wang & Zhang, 2005).

Hydrogen atom transfer AOH + ROO.  AO. + ROOH (1)

Electron transfer followed by reversible proton transfer AOH + ROO.  AOH+ + ROO- (2) + . + AOH + H2O ⇌ AO + H3O (3) + - H3O + ROO ⇌ H2O + ROOH (4)

Deprotonated phenolic transferring an electron AOH  AO- (5) AO- + ROO.  AO. + ROO- (6)

Figure 2.5. Reactions involved in scavenging of peroxyl radicals by a phenolic antioxidant. AOH and ROO. represent a phenolic antioxidant and a peroxyl radical, respectively.

The phenolic compound may act as an antioxidant by inhibiting enzymes that produce reactive oxygen species. For example, xanthine oxidase is an enzyme that can produce the reactive oxygen species superoxide from hypoxanthine. However, the flavonols quercetin, kaempferol, and myricetin can inhibit xanthine oxidase activity as seen by inhibition of the ability of the enzyme to convert xanthine to uric acid (Selloum et al., 2001). Moreover, phenolics may chelate transition metal ions to prevent the transition metal from producing

14 reactive oxygen species. For example, ferrous iron (Fe2+) is able to generate hydroxyl radicals from hydrogen peroxide and the hydroxyl radical can then damage DNA. However, epigallocatechin-3-gallate (and also other phenolics) can prevent DNA damage induced by Fe2+ (Perron et al., 2008). Since the impact of epigallocatechin-3-gallate is reduced upon addition of ethylenediaminetetraacetic acid (EDTA), the researchers attributed the inhibition of DNA damage to the formation of phenolic/Fe2+ chelates.

2.1.5. Phenolic antioxidant activity for food preservation

Oils and fats in foods are susceptible to oxidative decay, which can lead to rancidity or “off-flavours” arising primarily from aldehyde products (Kaur & Perkins, 1991). Aside from affecting flavour, as noted above (Section 2.1.3) lipid oxidation products might also cause cardiovascular health problems (see also Addis & Warner (1991) for more on dietary lipid oxidation products). With more and more people living in cities, more food items undergo a long transit from raw material to the end consumer. Furthermore, there is increasing trend towards processed foods containing multiple ingredients, some of which are sensitive to oxidative decay. In this regard, the omega-3 polyunsaturated fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) have been proposed to have health benefits (reviewed in Mori, 2014) and in 2004 the US Food and Drug Administration allowed for qualified health claims of reduced risk of coronary heart disease for foods containing EPA and DHA (US Food and Drug Administration, 2004). Polyunsaturation makes these fatty acids particularly susceptible to oxidation reactions (Kaur & Perkins, 1991). Antioxidants are a logical choice as additives to prevent spoilage of foods containing oxidizable lipids. Among the antioxidants used most widely in industry are the phenolic compounds butylated hydroxyanisole (BHA), butylated hydroxytoluene, (BHT), butylated hydroxyquinone (TBHQ), and esters of gallic acid (Loliger, 1991). However, some of these (specifically BHA and BHT) have shown toxic effects in some animal studies, although in these cases dosages were greater than would be expected to be ingested by humans (European Food Safety Authority (EFSA), 2012). While the synthetic antioixdants BHT and BHA are still allowed by regulatory agencies, there is a continual search for new phenolic (and non-phenolic) antioxidants, particularly from natural sources such as plant food powders (for example carrot, tomato, broccoli, and beetroot) (Neacsu et al., 2015),

15 mint leaf and citrus peel extracts (Viji et al., 2015), honey (Tahir et al., 2015), and licorice extract (Zhang et al., 2014), to name a few recent works.

2.1.6. Structure-functional correlations among phenolics with antioxidant properties

General relationships between molecular structure of phenolic compounds and antioxidant activity have been previously identified. The number and location of hydroxyl groups, along with the presence of double bonds that increase the degree of conjugation, all seem to have a role in determining antioxidant activity (reviewed in Balasundram et al., 2006). Hydroxycinnamic acids generally show higher antioxidant activity than corresponding hydroxybenzoic acids, which may be due to the double bond in the propanoid group of the hydroxycinnamic acid, providing increased delocalization of the unpaired electron of the radicalized phenolic (Natella et al., 1999). Increased electron delocalization will stabilize the phenolic radical thereby making it easier to form, meaning the original phenolic is more reactive. A study was carried out with the flavonoids to investigate structural features important for antioxidant activity towards the radical of 2,2'-azinobis(3-ethylbenzothiazoline-6- sulphonate) (ABTS˙+ radical) in aqueous solution (Rice-Evans et al., 1995). Compounds with an ortho dihydroxy substitution in the B ring (quercetin and cyanidin) had higher antioxidant activity than comparable compounds (kaempferol and pelargonidin, respectively) with only a single hydroxyl group in the B ring (Fig. 2.6). Replacing hydroxyls with O-glycosides (as in 3- OH in quercetin and 7-OH in naringenin to 3-O-glycoside in rutin and 7-O-glycoside in naringin, respectively) resulted in decreased antioxidant activity. The presence of the double bond between carbon two and three in the C ring (present in quercetin but lacking in the otherwise identical taxifolin (Fig. 2.6)) resulted in higher antioxidant activity. Another study identified similar relationships between structure and antioxidant activity of flavonoids (Van Acker et al., 1996). In this case, antioxidant activity was measured for lipid peroxidation and electrochemical oxidation potentials were also measured. The researchers observed an overall qualitative correlation between antioxidant activity and oxidation potentials. They identified that compounds with an ortho dihydroxy substitution in the B ring had the highest activity and that for such compounds the rest of the molecule was relatively less important in affecting activity. Among such compounds, quercetin and myricetin showed highest activity. This suggested that, in combination with the ortho dihydroxy, the double bond between C2 and C3 along with the 3-

OH results in a very strong antioxidant, possibly due to the extensive conjugation that such a compound has (Van Acker et al., 1996). Replacing the 3-OH of quercetin with 3-O-rutinose in rutin resulted in slightly lower activity for rutin. The importance of the 3-OH became more prominent when comparing the reduction in activity of a rutin derivative that has OEtOH at 7, 3’, and 4’ positions (and hence lacks the ortho dihydroxy in ring B) to a quercetin derivative that has OEtOH at 7, 3’, and 4’ positions (Fig. 2.6), reinforcing the idea that in compounds with an ortho dihydroxy in ring B the rest of the molecular structure is relatively less important for determining antioxidant activity (Van Acker et al., 1996).

6 HO 7 OH HO OH HO OH 5 A O 8 O 4 C O O+ O HO 3 HO HO 2 2’ 6’ B 3’ 5’ R3 4’ R1 R3 R1 R3 R1 R R R 2 2 2 flavonol anthocyanidin dihydroflavonol

R1=R3=H, R2=OH; kaempferol R1=R3=H, R2=OH; pelargonidin R1=R2=OH, R3=H; taxifolin R1=R2=OH, R3=H; quercetin R1=R2=OH, R3=H; cyanidin R =R =R =OH; myricetin 1 2 3

HO OH

R3 R1 R 2 flavanone

R1=R3=H, R2=OH; naringenin

Figure 2.6. Structures of some classes of flavonoids. The ring nomenclature and carbon numbering system are shown on the flavonol skeleton.

2.1.7. In vitro measurement of antioxidant activity

There are a variety of assays that have been developed to give some measure of antioxidant activity. These assays measure hydrogen atom transfer (HAT), electron transfer (ET), or a combination of the two. Among the more popular hydrogen atom transfer assays are: oxygen radical absorbance capacity (ORAC), total radical-trapping antioxidant parameter (TRAP), total oxidant scavenging capacity (TOSC), and crocin bleaching assay (Table 2.4) (Prior et al., 2005). Among the more popular electron transfer assays is the ferric reducing antioxidant power (FRAP) assay (Prior et al., 2005). Other popular assays that assess both hydrogen atom donation and electron transfer include trolox equivalent antioxidant capacity (TEAC) and 2,2-diphenyl-1-picrylhydrazyl (DPPH) (Table 2.4) (Prior et al., 2005).

Table 2.4. Popular antioxidant activity assays. Assay name Mechanisma Reagents Biological Quantification of activity relevance ORAC HAT oxidizer (can Yes lag time until probe TRAP vary), probe oxidation and/or decrease in TOSC (can vary) slope depicting rate of probe crocin oxidation bleaching ET FeIII-TPTZ No Reduction of oxidant at FRAP HAT and ET ABTS chosen time TEAC DPPH DPPH a HAT and ET mean hydrogen atom transfer and electron transfer, respectively.

The concept behind the ORAC, TRAP, TOSC, and crocin bleaching assays is essentially the same, with the main differences being the reagents that are used and the methods of detection of reaction progress. In all cases, there is a radical generator as a source of in situ radicals, a target that acts as a probe for oxidation, and the antioxidant that inhibits oxidation of the probe by itself sacrificially becoming oxidized (Huang et al., 2005). Different versions of each method have been developed that use different radical generators and probes. Antioxidant activity can be quantified as the lag time before oxidation of the probe is initiated and/or the decrease in the rate of oxidation of the probe. In one version of the ORAC and TRAP assays,

18 both make use of 2,2′-azobis(2-methylpropionamidine) (AAPH) as a temperature-sensitive peroxyl radical generator and 2’,7’-dichlorofluorescein (DCFH) as the probe (Prior et al., 2005; Huang et al., 2005). As an alternative to AAPH and DCFH, 2,2’-azobis (2,4- dimethylvaleronitrile) (AMVN) can be used as a peroxyl radical generator in conjunction with a lipid soluble probe such as 4,4-difluoro-3,5-bis(4-phenyl-1,3-butadienyl)-4-bora-3a,4a-diaza-s- indacene (BODIPY 665/676) (Huang et al., 2005). The TOSC assay has been used with a peroxyl radical generator but also with hydroxyl radicals generated from iron/ascorbate and peroxynitrite radicals generated from 3-morpholinosydnonimine N-ethylcarbamide (SIN- 1)/diethylenetriaminepentaacetic acid (DTPA) allowing characterizations of different radicals (Regoli & Winston, 1999). The crocin bleaching assay uses crocin as the oxidizable probe. Crocin was introduced as a substitute to β-carotene because the former only undergoes radical oxidation while the latter can undergo light and heat induced oxidation (Prior et al., 2005). The FRAP, TEAC, and DPPH assays are similar to each other in that there is no probe as in the case of the purely HAT-based assays. Rather, the reaction between an oxidant and the antioxidant is measured directly, by measuring the change in oxidant concentration at a chosen time. The oxidants for the FRAP, TEAC, and DPPH are a complex of FeIII-2,4,6-tris(2-pyridyl)- s-triazine (TPTZ), 2,2’-azinobis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) radical cation, and DPPH radical, respectively. These assays are considered less biologically relevant than the purely HAT-based assays because they use oxidants that are not oxygen-based.

2.1.8. Solubility considerations for antioxidant activity

In addition to presence of functional groups like hydroxyls, the solubility of the phenolic also plays a role in antioxidant efficacy in different solution conditions (reviewed in Shahidi & Zhong, 2011). It has been previously observed that, in general, non-polar antioxidants are better than their polar counterparts in protecting a lipid compound that is emulsified in an aqueous solution, whereas the polar antioxidant is found to be more effective in purely lipid systems (Porter et al., 1989; Frankel et al., 1994; Cuvelier et al., 2000). This phenomenon is proposed to occur due to preferential partitioning of the phenolic antioxidant at the interface where oxidation of the lipid is initiated (Frankel et al., 1994). In an emulsion of lipid in water, the more non- polar antioxidants would partition to the water-lipid interface and scavenge free radicals before these radicals propagate into the interior of the lipid micelle. As support for this mechanism, the

19 surfactant effectiveness of a series of acylated hydroxytyrosols (with varying acyl chain-length) correlated with their antioxidant activity in an oil-in-water emulsion, suggesting that the more effective antioxidants are those that act as surfactants and partition to the interface of the oil-in- water emulsion (Lucas et al., 2010). Notably, the antioxidant activity of the octanoate ester of hydroxytyrosol was greater than both lower chain-length and higher chain-length esters, with a parallel bell-shaped pattern observed for surfactant effectiveness (i.e. the octanoate ester had greater surfactant effectiveness than both lower chain-length and higher chain-length hydroxytyrosol esters) (Lucas et al., 2010). Therefore, this previous work also highlighted the fact that antioxidant polarity and antioxidant activity do not follow an entirely linear relationship (Lucas et al., 2010; Shahidi & Zhong, 2011). In the case of pure lipids, two different interfaces have been posited as being relevant. It was originally suggested that oxidation was initiated at the air-lipid interface and that polar antioxidants were more effective by preferentially partitioning to this interface while non-polar antioxidants were less effective because they remained soluble in the bulk lipid (Frankel et al., 1994). It was later suggested that micro- aqueous environments exist as reverse micelles and that these are the sites where oxidation is initiated (Chaiyasit et al., 2007). The polar antioxidants would preferentially partition to these reverse micelles, thereby being more effective than the non-polar antioxidants that are dissolved in the bulk lipid. The condition of emulsification of lipids is common in foods (such as milk and dressings), and biological systems (such as lipoproteins, whose oxidation is thought to be associated with cardiovascular disease (Scalbert et al., 2005)). Therefore, one chemical property of naturally occurring phenolic antioxidants that may be advantageously modified is their lipophilicity. As such, the lipophilized antioxidants can find application as food preservants; and may also be used as nutraceuticals that protect against oxidative stress of lipid components in the human body. At the same time, changes in molecular structure of the phenolic that improve lipophilicity may also cause changes to the presence of functional groups mentioned above (hydroxyls and conjugated double bonds) that have been found to be important for antioxidant activity. Therefore, it may be required to strike a balance between lipophilicity and antioxidant activity. Two enzyme types that have potential to modify chemistry of natural product phenolics (for example to increase lipophilicity) are esterases/lipase enzymes and laccase enzymes. As such, Section 2.3 and Section 2.4 will review esterase/lipase enzymes and laccase enzymes,

20 respectively; but first Section 2.2 will more broadly review approaches to plant phenolic derivatization.

2.2. Derivatization of plant phenolics

2.2.1. Enzymatic strategies used in plant phenolic derivatization

A variety of enzymes have been used to modify chemistry of plant phenolics. Enzymes can offer the advantage of stereo- and regioselectivity, which is important in fine-tuning only selected regions of the molecule. The resultant changes can affect the activity and solubility properties of the starting compound so that it may be used in novel applications. Lipases can carry out transesterification reactions using activated esters (such as vinyl esters) to produce acylated derivatives of phenolics. For example, immobilized lipases were used to acylate the phenolic hydroxyls of resveratrol (Torres et al., 2010) with vinyl acetate as the acyl donor. In this case, the authors aimed to selectively acetylate the hydroxyl at position 3 to protect it from becoming sulfated or glucuronated in the liver, thereby potentially increasing resveratrol’s bioavailability. The authors found that a lipase from Alcaligenes sp. was able to almost exclusively acetylate the 3-OH while leaving the other two hydroxyls of resveratrol intact, whereas other tested lipases showed less selectivity. Similar to the lipases, a select group of esterases have also been used to acylate phenolic hydroxyls. Topakas et al. (2003) used a feruloyl esterase to esterify hydroxycinnamic acids in the hopes of improving the lipid solubility of the compound. Another group of acylating enzymes are the aptly named acyltransferases. These enzymes typically require an acyl-Coenzyme A (acyl-CoA) substrate as the acyl donor. For example, a malonyltransferase was used to catalyze the addition of malonyl of malonyl-CoA to a free hydroxyl of glucose that is covalently linked to an anthocyanin (Suzuki et al., 2002). The anthocyanins are a type of flavonoid and are notable for the colours they impart to flowers. Upon malonylation, the anthocyanin pigment was found to be more stable (Suzuki et al., 2002). Continuing with the theme of transferases, prenyltransferases are another enzyme group that have been used for modification of bioactive phenolics. Prenylated phenylpropanoids may have anti-inflammatory and anticancer activity (Paulino et al., 2008; Messerli et al., 2009). A prenyltransferase from S. spheroides was able to catalyze prenylation of hydroxycinnamic acids, resveratrol, and some flavonoids (Ozaki et al., 2009). Another group of transferase enzymes that

21 are used in phenolic modification are the methyltransferases. Methylation of flavonoids may be important for their antifungal and antibacterial properties (Aida et al., 1996; Zhang et al., 2008). Accordingly, a methyltransferse from tomato was recombinantly expressed in E. coli and displayed activity on flavonoids including one flavanone, a dihydroflavonol, flavones, and flavonols (Cho et al., 2012). The enzyme had regiospecificity for the 3’ and 5’ positions of the flavonoids. In many cases, natural plant phenolics are found in glycosylated forms. Glycosylation can impart increased water solubility to the phenolic, among other effects. Some enzymes that hydrolyze glycosidic bonds also display transglycosylation activity. Such was the case for a maltogenic amylase from Bacillus stearothermophilus. This enzyme was able to transfer mono-, di-, and tri-glucose units from maltotriose to the flavonoid naringin (Lee et al., 1999). The major product, in which maltose had been attached to the already existing glucose of naringin, displayed not only improved water solubility but also reduced bitterness (Lee et al., 1999). Finally, oxidase enzymes including laccases and peroxidases can be used to produce homo- and heterocoupled phenolic products. Makris and Rossiter (2002) used horseradish peroxidase to oxidize quercetin producing a compound that was later identified as a quercetin dimer (Gulsen et al., 2007). Ultimately, the authors found the dimer to have reduced activity (compared to the monomer) in scavenging DPPH radical, hydroxyl free radical, and hydrogen peroxide (Gulsen et al., 2007). Nicotra et al. (2004) used a laccase from M. thermophyla to synthesize dimers of resveratrol that might have bioactivities similar to naturally occurring oligostilbenes. Nugroho Prasetyo et al. (2011) used a laccase from T. hirsuta to couple different simple phenolics such as catechol onto naringenin. The authors aimed to add hydroxyl groups that would be conjugated to the isolated C-ring hydroxyl of naringenin, thereby potentially increasing antioxidant activity (Nugroho Prasetyo et al., 2011). One commom outcome (intended or unintended) of phenolic derivatization is modification of the solubility of the phenolic. As this affects the extent to which the phenolic can access different sites in biological systems, it is a significant motivation of many derivatization processes, and so will be reviewed in the next two sections (Sections 2.2.2 and 2.2.3).

2.2.2. Increasing hydrophilicity of phenolic compounds

Phenolics from the flavonol class have been shown to have protective effects against ischaemia-reperfusion injury (Williams et al., 2011). For such an application, it is desirable to increase the water solubility of the antioxidant to administer more of it intravenously into the blood with fewer injections. One way to increase water solubility is to make a water soluble prodrug derivative of the phenolic that is converted into the parent compound in vivo. Water soluble prodrugs of flavonols were shown to have protection against sheep cardiac reperfusion injury comparable to equimolar quantities of the parent compound (Williams et al., 2011). In this case, the researchers added phosphate or adipic acid groups to the parent flavonol to make the prodrug. The prodrug can be converted back to the parent compound by action of phosphatase or esterase enzymes that are naturally present in tissues and blood. However, this particular study did not yet examine dosage increases that could potentially be realized with the water soluble prodrugs and the parent phenolics were administered in solutions containing DMSO as organic co-solvent. While DMSO is often used in experimental settings, it is undesirable for clinical applications because it may cause side effects including hemolysis (Muther et al., 1980; Santos et al., 2003) and its ability to dissolve some plastics used for intravenous administration (Marshall et al., 1984). Another approach to increasing the amount of phenolic in aqueous media is to encapsulate the compound in a carrier that has both hydrophobic core and hydrophilic exterior (reviewed for quercetin in Cai et al., 2013). In this case, the encapsulated phenolic might also be protected from undesirable in vivo metabolic modifications on route to the site of action. One group of encapsulating compounds that can be used is the cyclodextrins. A complex of quercetin and sulfobutyl ether-7β-cyclodextrin allowed for improved solubilization of quercetin in aqueous neutral buffer solution (Kale et al., 2006). The orally administered complex also showed improved tumour growth suppression in mice compared to equivalent doses of the uncomplexed quercetin (Kale et al., 2006). A third approach to increase hydrophilicity is to form nanocrystals of the desired compound. Nanocrystals are highly fine particles of the compound and have a mean particle size less than 1 µm (typically between 200 nm and 500 nm) (Keck and Muller, 2006). As expected, the increased surface area of the fine particles leads to increased dissolution rate. Researchers produced a nanocrystal formulation of the phenolic compound curcumin and this nanocrystal had a mean diameter of 250 nm compared to 22 µm for crystalline curcumin (Onoue et al.,

2010). Compared to crystalline curcumin, the curcumin nanocrystal showed improved dissolution rate in water and showed improved bioavailability in male Sprague-Dawley rats after oral administration.

2.2.3. Increasing lipophilicity of phenolic compounds

For some applications, it can be advantageous to increase lipophilicity of antioxidant phenolic compounds. If a lipid is being targeted for antioxidant protection, then a lipophilic antioxidant may be more effective than a hydrophilic one. It was previously observed that, in general, non-polar antioxidants are more effective than their polar counterparts at protecting lipids dispersed in water (reviewed in Shahidi & Zhong, 2011). In this context, lipophilic antioxidants can be exploited for preservation of emulsified lipids in foods, such as in mayonnaise, dressings, and milk. One of the methods used for increasing lipophilicity of phenolics is to add a long-chain alkyl group by way of esterification reactions of the phenolic with a long-chain acyl compound or long-chain alcohol compound (Fig. 2.7). For example, alcohols of varying chain length were esterified (using sulfuric acid as catalyst) onto caffeic acid to produce lipophilic derivatives (Aleman et al., 2015). The lipophilic caffeic acid derivatives showed better protection towards fish oil emulsified in mayonnaise or milk. It should be noted that it was not the longest alkyl chain ester derivative of caffeic acid that conferred best protection, but rather intermediate chain length (for mayonnaise) and short chain length (for milk) ester derivatives showed best protection (Aleman et al., 2015).

O + O + A) ⇌ HOR2 R1 OR2 HO R1 O R R3 3

O OH O OR4

B) + HOR4 ⇌ + H2O

OH OH

Figure 2.7. Possible routes to adding long-chain alkyl groups to phenolic compounds to increase lipophilicity of the phenolic. In A) the long-chain alkyl group is represented by R1 and is added to the phenolic as part of an acyl group. In the case of B) the phenolic contains a carboxylic group on one of its ends. The long-chain alkyl group is represented by R4 and is added to the phenolic as part of an alkoxy group.

A potential way of increasing lipophilicity is by forming dimer and higher level oligomers of the phenolic compound. Researchers had previously isolated naturally occurring dimers, trimers, and tetramers formed of successive epicatechin molecules added to catechin (Fig. 2.8) (Plumb et al., 1998). The (n-octanol)-water partition coefficient of these compounds demonstrated increasing preference for the hydrophobic n-octanol phase with increasing degree of oligomerization after the dimer (Plumb et al., 1998) (the dimer did not have a significant difference in partition coefficient compared to the monomer). However, in this study, the researchers saw decreased antioxidant activity towards iron/ascorbate induced oxidation of phospholipid liposomes with increasing lipophilicity (due to increased oligomerization) of the antioxidant compound. In a related study, catechin monomer and oligomers up to hexamer were compared for their antioxidant activity toward L-α-phosphatidylcholine liposome, using different inducers of oxidation (Lotito et al., 2000). When iron/ascorbate (which would be present in the aqueous phase) was used to induce oxidation, antioxidant activity decreased with increasing degree of oligomerization of catechin up to the pentamer, showing similar trends as seen by Plumb et al. (1998). However, when 2,2’-azobis (2,4-dimethylvaleronitrile) (AMVN,

25 which would be present in the lipid phase) was used to induce oxidation, antioxidant activity increased with increasing degree of oligomerization of catechin up to the pentamer (Lotito et al., 2000). These previous examples demonstrate that the more lipophilic phenolic derivative is not always the better antioxidant for emulsified lipids. Changing the inducer of oxidation can change the pattern of antioxidant activity, and increasing the molecular weight of the antioxidant can be beneficial up to a certain point but further increases in molecular weight may become detrimental.

OH OH epicatechin HO O H OH n OH H OH OH

HO O catechin H H OH OH

Figure 2.8. Naturally occurring dimers (n=1), trimers (n=2), and tetramers (n=3) made of successive epicatechin molecules linked to catechin.

The addition of alkyl groups via esterification has been carried out using acid catalysts. For example, esters of caffeic acid were synthesized using sulfuric acid as catalyst (Aleman et al., 2015), while rosmarinic acid esters were produced using the strongly acidic sulfonic resin Amberlite IR-120H (Panya et al., 2012). The use of acid catalyst represents harsh reaction conditions. On the other hand, the use of enzymes to synthesize novel lipophilic derivatives of phenolics is an alternative approach that avoids the use of strong acids. In addition, naturally occurring oligomeric phenolics are present for a limited subset of the phenolic classes including flavones, flavanols, and hydroxycinnamic acids. Enzymatic catalysis can allow the synthesis of additional novel oligomeric compounds not readily available from environmental sources.

Esterases and lipases are enzymes that can catalyze esterification reactions while laccases can catalyze oxidation that leads to oligomerization.

2.3. Esterases/lipases

Broadly speaking, esterases and lipases are enzymes that catalyze the hydrolysis of ester bonds (Fig. 2.9). Esterases and lipases belong to the general class of enzymes called ester hydrolases (EC 3.1) by the Nomenclature Committee of the International Union of Biochemistry and Molecular Biology (NC-IUBMB) (2010b). The ester hydrolase class is further divided to result in more classes, among of which are carboxylic-ester hydrolases (i.e. esterases and lipases) (EC 3.1.1), thioester hydrolases (EC 3.1.2), phosphoric monoester hydrolases (EC 3.1.3), and others. Carboxylic-ester hydrolases (EC 3.1.1) are then divided into carboxylesterases (EC 3.1.1.1), arylesterases (EC 3.1.1.2), triacylglycerol lipases (EC3.1.1.3), and others. A less formal but often used distinction for the carboxylic-ester hydrolases is to refer to them simply as either “esterases” or “lipases”. In this case, esterases (for example EC 3.1.1.1) are differentiated from lipases (for example EC 3.1.1.3) in the tendency of esterases to prefer short-chain substrates while lipases show activity on both short-chain and long-chain substrates. From this point on, the terms esterase and lipase will be used rather than the NC-IUBMB terminology. In addition to classification based on reaction substrates, these enzymes have also been classified into families based on amino acid sequence features. For example, the carbohydrate active enzyme (CAZy) classification system, which focuses on enzymes acting on carbohydrates, comprises a carbohydrate esterase class that is divided into 16 families (Lombard et al., 2013). Likewise, in the ESTerases and alpha/beta-Hydrolase Enzymes and Relatives (ESTHER) classification system, esterases and lipases (along with other hydrolases) have been classified within 148 families based mainly on sequence features and any available biological data (Lenfant et al., 2013).

O O

+ HOR3 ⇌ + HOR2 R OR R OR 1 2 1 3

Figure 2.9. Reaction catalyzed by esterases and lipases. In the case of hydrolysis, R3 is a hydrogen atom so that HOR3 is water and the final products are a carboxylic acid and an alcohol. In the case of transesterification, R3 is an alkyl group so that HOR3 is an alcohol and the final products are a new ester and an alcohol.

2.3.2. Structural features

Esterases and lipases are characterized by an α/β-hydrolase fold structure, which is defined as a central β-sheet (as opposed to α/β barrel) of eight β-strands connected and surrounded by six α-helices (Fig. 2.10) (Ollis et al., 1992). Different hydrolases show variability around this prototypical structure in terms of the number of β-strands and α-helices. For example, a lipase from Bacillus subtilis and a lipase from Pseudomonas cepacia are both composed of six β-strands (Van Pouderoyen et al., 2001; Kim et al., 1997b), while an esterase from Pseudomonas fluorescens is made of seven β-strands (Kim et al., 1997a).

Figure 2.10. Prototypical α/β hydrolase fold structure. A) Image adapted from Ollis et al. (1992). α helices and β strands represented by cylinders and arrows, respectively. Dark circles in loop regions after β5, β7, and β8 show positions of catalytic residues serine, aspartate/glutamate, and histidine, respectively. B) 3-Dimensional structure of P. fluorescens esterase showing α/β hydrolase fold. Catalytic triad residues serine, aspartate, and histidine are shown as sticks coloured red, blue, and magenta, respectively. Oxyanion hole residues are shown as sticks coloured in cyan.

The catalytic residues form a triad and are found in the order of serine, aspartate, and histidine in the primary sequence of the enzyme. In some cases, glutamate is present in place of aspartate, as in the case of a feruloyl esterase from Pleurotus eryngii (Nieter et al., 2014). The catalytic nucleophilic serine is typically contained in the consensus sequence Gly-X-Ser-X-Gly. The catalytic serine is located on a sharp γ-like turn (termed the nucleophilic elbow) of the enzyme secondary structure, going from β5 to the following alpha helix (αC) in the prototypical α/β hydrolase structure (Fig. 2.10) (Ollis et al., 1992). An important structural feature of the enzyme, which helps to stabilize the tetrahedral intermediate of the substrate formed during catalysis, is known as the oxyanion hole. This oxyanion hole is formed by the main-chain amide hydrogens of two amino acid residues. One of these residues is right after the nucleophilic serine in the nucleophilic elbow, while the second residue is located at a loop going from β3 to αA in the prototypical α/β hydrolase structure (Fig. 2.10) (Ollis et al., 1992). A structural feature that is unique to lipases as opposed to esterases is the presence of a lid that covers the enzyme active site. This lid is formed from α-helical segments of the protein that are mobile to allow access of the substrate to the active site (Brady et al., 1990; Brzozowski et al., 2000; Grochulski et al., 1994; Kim et al., 1997b). The presence of the lid and its movement have been proposed as an explanation for the phenomenon of interfacial activation of lipases (Brzozowski et al., 1991). Interfacial activation is the increase in lipase activity that is observed when the enzyme is present in a solution where the substrate concentration is high enough to form a separate phase (Verger, 1997). Upon opening of the lipase lid, hydrophobic patches on the underside of the lid become exposed and may be stabilized by interaction with the hydrophobic substrate phase at the solution interface (Brzozowski et al., 1991). Additionally, the lipase active site becomes exposed for catalysis. A recent experiment showed the feasibility of altering substrate preference of Rhizopus chinensis lipase to favour short chain substrates by swapping the lipase lid with a hydrophilic lid-like region from Aspergillus niger esterase (Yu et al., 2014).

2.3.3. Catalytic mechanism

The main catalytic residue of esterases and lipases is a nucleophilic serine (Ser) with the other two residues of the catalytic triad increasing nucleophilicity of this serine (Fig. 2.11) (reviewed in Jaeger et al., 1999). The histidine (His) side chain can abstract a proton from the

30 hydroxyl of the serine side chain resulting in a more nucleophilic serine. The resulting positive charge on the histidine is stabilized by the negatively charged aspartate (Asp) side chain. The side chain oxygen of the nucleophilic serine attacks the carbonyl carbon of the ester substrate forming a tetrahedral intermediate. The negatively charged oxygen of the tetrahedral intermediate is stabilized by the oxyanion hole of the enzyme. The catalytic histidine can then donate a proton to the alcohol oxygen of the original ester substrate resulting in collapse of the tetrahedral intermediate, release of the alcohol of the original ester substrate, and formation of an acyl-enzyme intermediate. The catalytic histidine can then abstract a proton from another nucleophile (water or an alcohol) which can then attack the carbonyl carbon of the acyl-enzyme intermediate, resulting in formation of a carboxylic acid or a new ester and regenerating the enzyme for catalysis of a new substrate.

2.3.4. Transesterification reactions

As mentioned in the above section (Section 2.3.3), the product from esterase and lipase catalysis can be a carboxylic acid or a new ester. The presence of water will favour the production of a carboxylic acid and water will be plentiful if the reaction is carried out in aqueous solvent. The use of organic solvents as the reaction media can result in esterification or transesterification products being formed. However, the reaction rate can be affected by carrying out enzymatic reactions in non-aqueous solvents. Furthermore, even if the medium is primarily non-aqueous, some water is required to maintain enzyme activity. Organic solvents can affect enzymatic reaction rate in a number of ways (reviewed in Dordick, 1992). Organic solvents may form more favourable interactions with the substrate than water forms with the substrate, resulting in more stabilization of the substrate. This leads to less propensity of the substrate to partition into the enzyme active site where catalysis can occur.

Stabilization of the substrate was proposed as the cause for increased apparent Km values of a peroxidase for phenolic substrates in water-immiscible organic solvents vs. aqueous media (Ryu & Dordick, 1989). Additionally, the authors observed that the peroxidase still maintained, and in some cases increased, its maximum turnover capacity (Vmax) in organic solvents despite the increased Km.

O O O O O O NH NH HN NH HN HN

- 1 2 R O O O 1 ⇌ R OR ⇌ O 1 2 O R1 OR OR 2 O H 3 -O Asp Ser HN N+ H OH Ser

Ser N NH His

His ⇌ 3

O O NH O HN

R OR - 1 3 4 O R OR ⇌ 1 3 O HN N+ H Ser His

Figure 2.11. Catalytic mechanism of esterases and lipases. Image adapted from Jaeger et al. (1999). Dashed lines indicate stabilizing hydrogen bonds. In step 1, the nucleophilic serine (Ser) attacks the ester substrate to produce the tetrahedral intermediate, which after releasing the OR2 group in step 2, gives the acyl-enzyme intermediate. In step 3, a new alcohol group (HOR3) reacts with the acyl-enzyme intermediate. In step 4, the nucleophilic serine is regenerated as the product is released. The net reaction is R1COOR2 + HOR3 ⇌ R1COOR3 + HOR2.

Another important factor of organic solvents is their ability to interact with water. Enzymes have a strong interaction with a layer of water close to their surface that seems to be important for maintaining activity by facilitating mobility of the enzyme active site (Dordick, 1992). It was shown that enzymes suspended in organic solvents with the same level of water were able to bind different amounts of water depending on the hydrophobicity of the solvent; and that activity of the enzyme correlated with the content of enzyme-bound water rather than with the total water content of the media (Zaks & Klibanov, 1988). Furthermore, certain organic solvents are able to strip away some of the tightly bound water of the enzyme to the extent that the organic solvent can interact with water (Gorman & Dordick, 1992). Generally, organic solvents that are more hydrophilic (as measured by the octanol-water partition coefficient (log P)), or have higher dielectric constants are better able to disrupt the interaction of water with the enzyme and hence remove water from the enzyme. The physical state of the enzyme in organic solvents would often be a suspension of particles. This raises the possibility of diffusion of the substrate from the bulk solvent to the enzyme active site acting as a limiting factor for reaction rate. Theoretically, the substrate would have to firstly diffuse from the bulk solvent through a boundary film (that separates the enzyme particle surface and the bulk solvent) onto the enzyme particle surface, and secondly, diffuse through the channels of the enzyme particle to the enzyme active sites (Kamat et al., 1992). However, diffusion effects may not be critical in many cases and can be compensated for by agitation of the reaction media and using small enzyme particles (Wescott & Klibanov, 1993; Klibanov, 1997). The process of separating water from the enzymes to form the enzyme particles is often achieved by lyophilization. The lyophilization step can cause changes to the structure of proteins, resulting in diminished activity (Klibanov, 1997; Pikal-Cleland & Carpenter, 2001; Bhatnagar et al., 2007). It has been suggested that loss of water during lyophilization may lead to loss of hydrogen bonds that maintain the native protein state; and that lyoprotectants can promote stability be engaging in hydrogen bonds with the protein in place of lost water (Pikal-Cleland & Carpenter, 2001). At the same time, researchers managed to solubilize subtilisin BPN’ enzyme in non- polar and polar organic solvents using a surfactant at low enough concentrations to avoid micelle formation while promoting enzyme solubilization (Wangikar et al., 1997). Using such an enzyme preparation in octane allowed to achieve an increase in activity (kcat/Km) of over two

33 orders of magnitude compared to a suspension of lyophilized enzyme in octane. Additionally, the researchers observed changes in tertiary structure of the subtilisin BPN’ enzyme after solubilization and incubation in the polar tetrahydrofuran (THF). They proposed this was caused by THF penetrating into the protein structure and disrupting native interactions between residues of the enzyme, which suggests that direct interaction between organic solvent and enzyme is another factor that can influence enzyme activity.

2.3.5. Applied Use

Lipases are used more in industry than esterases and this might be due to the perception that lipases generally show better organic solvent tolerance, enantioselectivity, and substrate range (Bornscheuer, 2002). Among the esterases, feruloyl esterases are popularly investigated for potential use in breakdown of plant polymers for the biofuels industry (reviewed in Faulds, 2010), in processing of animal feeds to improve digestibility, and in release and modification of ferulic acid for food and nutraceutical applications (reviewed in Topakas et al., 2007). Lipases are commonly used in the detergent industry due to their ability to degrade fats in clothing stains (Sharma & Kanwar, 2014). For example, in 1995 Genencor International (now a subsidiary of DuPont) introduced two bacterial lipases for detergent applications. Additionally, lipases can be used on textiles to facilitate dyeing by removing lubricant residues that come from the machinery used for manufacturing, thus improving wettability and absorbency of the textile (Duran & Duran, 2000). In the food industry, lipases are used to carry out fatty acid exchange reactions of triacylglycerols in foods. One prominent case is the use of lipases to convert oils such as palm oil to a product that is closer in characteristic to cocoa butter, as cocoa butter is desirable for its ability to be solid at room temperature but quickly melt in the mouth (Jala et al., 2012), and a procedure was patented by Unilever in the early 1980s for such use of lipases (Coleman & Macrae, 1981). Lipase catalyzed transesterification can also be applied to produce flavour and fragrance ingredients such as 2-phenethyl esters (Li et al., 2014). Lipases can also be used for biosensor applications. For example, a biosensor for assessing levels of dimethoate pesticide was based on measuring inhibition of lipase due to dimethoate (Wei et al., 1997). Another biosensor incorporating lipases was developed to measure target DNA (Chen et al., 2014). In this case, binding of target DNA to a hairpin-

34 structured probe DNA causes linearization of the probe. This leads to exposure of an ester bond linking one end of the DNA probe with ferrocene, such that a lipase can act on the ester bond and release ferrocene; the free ferrocene can then be electrochemically detected. Finally, lipases can be used on their biologically natural targets in biosensors that measure serum levels of lipids. For example, lipase immobilized on nanoporous gold can cleave serum triglycerides to release fatty acids that cause a change in pH. The pH change can be detected electrochemically (Wu et al., 2014).

2.4. Laccases

Laccases (EC 1.10.3.2) belong to the group of proteins commonly called multicopper oxidases. They are characterized by their ability to oxidize ortho- and para- substituted phenolic compounds but they often also show activity on aniline compounds such as p-phenylenediamine (NC-IUBMB, 2010a) (Fig. 2.12). Similar to certain esterases, a group of laccases has also been categorized in the CAZy database, in this case into the auxiliary activity 1 (AA1) family (Lombard et al., 2013). Laccases use copper atoms to carry out the oxidation of their substrates with oxygen as the electron acceptor, to produce the oxidized substrate and water as the final products. The oxidized substrate is initially a radical and can often undergo further reactions with or without enzyme involvement. In some cases, radicalized phenolics can undergo both synthesis into larger molecular weight products and breakdown into smaller molecular weight products. For example, the flavonol quercetin was oxidized using a chemical radical to yield quercetin breakdown products and also a quercetin dimer product, so there can be competing breakdown and oligomerization in the same reaction (Zhou & Sadik, 2008). Perhaps the most popular demonstration of the dual synthetic and degradative role played by these enzymes involves the plant polymer lignin. Laccases are thought to play a role in lignin formation as well as lignin degradation. Among the very early works in this regard, Leonowicz et al. (1985) treated low and high molecular weight lignosulfonate fractions with a fungal laccase from Trametes versicolor. They observed the enzyme polymerized the low molecular weight fraction but degraded the high molecular weight fraction. Others have used lignin-model dimer compounds to simulate reactions with lignin. Kawai et al. (1988) used a laccase from Coriolus versicolor to break down model dimers. On the other hand, Maarit et al. (2009) observed tetrameric products or aldehyde products being produced after treatment of dimers with fungal

35 laccases from Melanocarpus albomyces and Trametes hirsuta but no degradation products were observed in this study. The authors noted that other researchers before them had observed breakdown products instead of the aldehyde products and proposed that the difference in products could be due to different reaction conditions (such as reaction temperature and enzyme dosage) (Maarit et al., 2009).

. O2 + 4ROH 2H2O + 4RO Figure 2.12. Reaction catalyzed by laccases. ROH represents a phenolic hydroxyl group. Besides phenolic compounds laccases often also show activity on aniline compounds such as p- phenylenediamine.

2.4.2. Structural features

Laccases are made up of protein folds known as cupredoxin domains. The cupredoxin domain is a greek key β-barrel structure which is formed from two β-sheets. Typically, eight β- strands make up the two β-sheets, with the two β-sheets in a sandwich arrangement in relation to each other (Fig. 2.13A). This fold was first identified in the copper-binding proteins azurin (Adman et al., 1978) and plastocyanin (Colman et al., 1978) and subsequently identified in the copper-binding protein cupredoxin (Adman et al., 1989), which resulted in the current name of the fold. Laccases typically contain three cupredoxin domains but examples of laccases with only two such domains have been identified in Streptomyces coelicolor (Skálová et al., 2009) and Streptomyces sviceus (Gunne et al., 2014). The two-domain laccases seem to make up for the “missing” domain by forming homotrimers as the functional unit (Gunne et al., 2014). Laccases contain four copper atoms divided into three types (Fig. 2.13B and C). The different types of copper are distinguished by their electronic properties, which are influenced by their particular environment in the protein. One of the coppers is known as the type 1 copper and is responsible for the blue colour seen from solutions of these enzymes. Type 1 copper is contained in domain 3 of laccases (Sharma et al., 2007) and is coordinated by two histidines, a cysteine, and an additional residue that varies between different proteins (Claus, 2004). The type 1 copper is where the electron is first transferred from the substrate to the enzyme. The second two copper types (type 2 and type 3) are positioned into a trinuclear cluster in the protein (Fig.

2.13C) and are contained between domain 1 and 3 of the enzyme (Sharma et al., 2007). The single type 2 copper displays an electron paramagnetic resonance (EPR) signal, whereas the two type 3 coppers do not show an EPR signal due to antiferromagnetic coupling between them (Claus, 2004). The type 2 and each type 3 copper are coordinated by two and three histidine residues, respectively.

A) B)

type 2 Cu type 3 Cu

type 1 Cu

Figure 2.13. Structure of laccase TvL from Trametes versicolor. A) One of the cupredoxin domains of TvL (the protein has three such domains). Four β-strands make up the left side and four β-strands make up the right side of the barrel. B) Entire polypeptide chain showing location of copper atoms (brown) and coordinating residues (cyan). C) Close up showing only the copper coordinating residues. The type 1 copper is at the bottom while the trinuclear cluster is on top. Image was generated with PyMOL using structure having PDB code 1GYC.

2.4.3. Catalytic mechanism

The overall reaction catalyzed by laccases is the oxidation of a substrate accompanied by the reduction of oxygen. Four electrons are taken from the substrate and transferred to oxygen. The oxidation of the substrate occurs at the type 1 copper atom, after which the electrons are sent to the trinuclear copper centre, where oxygen is reduced to water (Claus, 2004). Two different models have been proposed for the transfer steps of the electrons from the type 1 copper to the trinuclear copper cluster (Solomon et al., 1996). In one scenario, after the type 1 copper becomes reduced by substrate, the type 1 copper transfers an electron to the type 2 copper (Fig 2.14 steps 1 and 2 leading to pathway A). The type 1 copper is then reduced a second time by a second substrate molecule, leading to both the type 1 and type 2 coppers concurrently transferring one electron to the type 3 coppers (Fig 2.14 steps 3 and 4 in pathway A). The type 1 copper is reduced a third time by a third substrate molecule, and as in the initial reduction, the electron is transferred to the type 2 copper (Fig 2.14 steps 5 and 6 in pathway A). Finally, type 1 copper is reduced by a fourth substrate molecule to give the fully reduced enzyme (Fig. 2.14 step 7). In this case, the two type 3 coppers take part in a single two-electron transfer step (i.e. when the type 1 and type 2 coppers concurrently each transfer one electron to the type 3 coppers). In the second scenario, after the type 1 copper becomes reduced by substrate, the type 1 copper does three one-electron transfers to the trinuclear copper cluster (although it is not readily clear in what order the trinuclear coppers are reduced) (Fig. 2.14 pathway B). In this case the two type 3 coppers take part in two one-electron transfer steps. Oxygen is proposed to be reduced to water in two two-electron steps as follows (Fig. 2.14) (Solomon et al., 1996). The fully reduced enzyme reacts with oxygen to produce a

38 hydroperoxide-enzyme intermediate, in which the hydroperoxide forms a bridge between type 2 copper and one type 3 copper (Fig 2.14 step 8). In this first step, the two type 3 coppers transfer their electrons to oxygen to form the hydroperoxide. The proton, required to form the hydroperoxide from oxygen, is obtained from the solvent water. Upon abstracting a proton from the solvent water, a hydroxide ion is formed that bridges the two type 3 coppers. In the second step, the type 1 and type 2 coppers transfer their electrons to the hydroperoxide (Fig 2.14 step 9). The hydroperoxide consequently becomes water and a hydroxide ion (two protons are also consumed in this step). This newly formed hydroxide ion bridges the type 2 and one type 3 copper. The coppers are now all in their oxidized forms and can be reduced in another cycle of substrate oxidation. In the process of reducing the coppers, the bridging hydroxides, which were formed during oxygen reduction, are converted into water (two protons being consumed in the process).

T2 T3 T1

Cu(II) 1 Cu(II) 2 Cu(II) OH Cu(II) Cu(II) OH Cu(I) + + Cu(II) S S Cu(II) H H2O OH OH

Cu(II) Cu(II) A 3 4 Cu(I) OH Cu(II) Cu(I) OH Cu(I) + + Cu(II) S S Cu(II) H H2O

Cu(I) Cu(I) Cu(II) Cu(II) 3 Cu(II) Cu(I) 4 B + + Cu(II) S S Cu(II) H H2O OH OH

Cu(I) Cu(I) A 5 Cu(II) Cu(II) Cu(II) Cu(I) S+ Cu(I) S Cu(I) 6

Cu(I) Cu(I) 5 Cu(I) Cu(II) Cu(I) Cu(I) B S+ Cu(II) S Cu(II)

Cu(I) Cu(I) 7 8 Cu(I) Cu(II) Cu(I) Cu(I) + Cu(I) S S Cu(I) O2 + H2O

Cu(II) Cu(II) 9 Cu(I) OH Cu(I) Cu(II) OH Cu(II) + Cu(II) 2H H2O Cu(II) OOH OH

Figure 2.14. Proposed catalytic mechanisms of oxidation by laccase enzymes. The figure shows the reduction and oxidation reactions presumed to occur with the 4 copper cofactors of a protein monomer. T1, T2, and T3 indicate the type 1, type 2, and type 3 coppers, respectively. S and S+ represent the reduced and oxidized, respectively, forms of a laccase substrate. Alternative pathways A and B represent two different proposed oxidation schemes. These alternative pathways become distinct after step 2 and then both follow the same sequence after step 6. In the final step (step 9), the enzyme is returned to the form it was in at the beginning of the sequence, resulting in one complete cycle. See text for details. Image adapted from Solomon et al. (1996).

2.4.4. Effect of Redox potential

Redox potential of the substrate and the laccase enzyme is an important parameter determining the activity of the laccase on any particular substrate. Redox potential is a measure of the ability of a substance to be oxidized/reduced and the value is usually expressed in terms of the potential for the reduction reaction. Electron donating functional groups at ortho or para positions to phenolic hydroxyls result in reduced redox potential (meaning higher oxidizability) of the molecule, and this may be part of the reason for higher reaction rates observed for the laccase-catalyzed oxidation of such substrates in comparison to phenolic molecules with electron withdrawing functional groups (Xu, 1996). For example, the fungal laccase from -1 -1 Polyporus pinsitus showed kcat values of 1600 ± 300 min , 3300 ± 100 min , and 3100 ± 100 min-1 on phenol, 2-hydroxyphenol, and 2,6-dimethoxyphenol, respectively (Fig. 2.15) (Xu,

1996). The higher kcat correlated with higher redox potential values for the two ortho substituted phenols, which have the electron donating hydroxy and methoxy substituents.

OH H3CO OCH3 OH OH OH phenol 2-hydroxyphenol 2,6-dimethoxyphenol Figure 2.15. Structures of closely related phenols. Fungal laccase from Polyporus pinsitus -1 -1 -1 showed kcat values of 1600 ± 300 min , 3300 ± 100 min , and 3100 ± 100 min on phenol, 2- hydroxyphenol, and 2,6-dimethoxyphenol, respectively (Xu, 1996).

The type 1 copper of the laccase enzyme is the copper responsible for the initial electron abstraction from the substrate, so factors affecting redox potential of this copper have been investigated. It was seen that laccases with high redox potential have a relatively larger distance between type 1 copper and one of the histidines that coordinates to that copper (Piontek et al., 2002; Garavaglia et al., 2004). By comparing crystal structures, Piontek et al. (2002) observed that the distance from type 1 copper to Nδ2 of His458 was 0.17 Å longer in laccase from Trametes versicolor (which has a standard reduction potential of 800 mV) compared to the

42 distance from type 1 copper to the corresponding nitrogen in laccase from Coprinus cinereus (which has a standard reduction potential of 550 mV). This greater distance means the electrons from the coordinating nitrogen are farther away from the copper; and this was proposed to result in greater electron deficiency of the copper, leading to a greater propensity of the type 1 copper to abstract electrons (i.e. a higher redox potential of the type 1 copper) (Piontek et al., 2002). The presence of a methionine residue that coordinates to the type 1 copper via an axial bond has also been indicated as important in affecting type 1 copper redox potentials. Substitution of Met502 in the CotA laccase from Bacillus subtilis to leucine or phenylalanine resulted in increased redox potential of the type 1 copper from 455 mV to 548 mV or 515 mV, respectively (Durao et al., 2006); while replacement of Phe463 of TvL laccase from Trametes villosa with methionine resulted in a decrease of the redox potential from 790 mV to 680 mV (Xu et al., 1999). The influence of the axial ligand has also been observed for other copper containing proteins besides laccases. Rusticyanin is a copper oxidase enzyme containing only type 1 copper. Mutation of the axial methionine in rusticyanin to leucine resulted in increased redox potential (final values of 613 mV to 798 mV, depending on the pH); to values similar to what are seen for fungal laccases, which themselves typically have a leucine instead of a methionine at the axial position in their wild type forms (Hall et al., 1999). On the other hand, mutating the axial methionine of rusticyanin to a glutamine resulted in reduced redox potential (final values of 472 mV to 563 mV) (Hall et al., 1999). The presence of an axial glutamine would represent the wild type situation of a type 1 copper oxidase known as stellacyanin, which is known to have a comparatively low redox potential of 184 mV. It was proposed that leucine and glutamine respectively form weaker and stronger interactions than methionine with the type 1 copper of the mutant rusticyanin; and those weaker and stronger interactions respectively increased and decreased the redox potential of the enzyme (Hall et al., 1999).

2.4.5. Applied use

Laccases have the potential to be exploited as industrial catalysts for a variety of applications. The only byproduct of their reaction being water and their ability to function on a wide range of substrates makes them ideal enzyme candidates. In addition, laccases can be used in conjunction with a mediator to increase the range of substrates acted upon. The laccase

43 initially oxidizes the mediator and then the mediator can act on secondary compounds that are inaccessible to the enzyme due to steric hindrance. Biotechnological use of laccases can require large quantities of enzyme and recombinant production technologies can be employed for this purpose. However, enzymes produced recombinantly do not always incorporate a full complement of four copper atoms per protein monomer (Galli et al., 2004; Durao et al., 2008a). One approach used to overcome this is to incubate the purified enzyme with copper to incorporate additional copper into the enzyme, as was done for the fungal laccase from Coprinus cinereus recombinantly expressed in Aspergillus oryzae (Bukh & Bjerrum, 2010). However, for some laccases the presence of copper may be required for correct folding of the protein during protein production, so another approach is to add copper into the cell culture media and grow the cells under low oxygen (microaerobic) conditions, as was done for the CotA laccase from Bacillus subtilis recombinantly produced in E. coli (Durao et al., 2008a). The low oxygen condition is presumed to facilitate intake of more copper into the E. coli cells by causing limitations in the copper efflux ability of the cells (Durao et al., 2008a). Finally, the growth of cells coexpressing the desired laccase along with copper binding chaperones has been used to increase copper incorporation. Copper binding chaperones bind copper and deliver it to their target enzymes. The CopZ copper binding chaperone was coexpressed with CotA laccase in E. coli leading to an increase of copper incorporation by the laccase (Gunne, et al., 2013). Laccases can play a role in bioremediation of pollutants. By way of catalyzing coupling reactions, laccases can be used to remove pollutants such as 2,4,6-trinitrotoluene (TNT), methylphenols, chlorophenols, and aromatic amines (reviewed in Kudanga et al., 2011). Contaminated soils can be detoxified by using laccase to couple the pollutants to humic compounds such as guaiacol; and phenolic contaminants in wastewater can be removed by oxidizing them so that they become oligo/polymers that are insoluble and can be filtered away (Kudanga et al., 2011). In the food industry, laccases can find a place in altering rheological characteristics of dough. In this case, laccases may be working by crosslinking the polymers of flour via the phenolic components of these polymers, resulting in improved dough strength made from such flour (Kudanga et al., 2011). Fruit juices may display undesirable haziness and browning due to interactions between phenolics and other compounds in the juice. Laccases can be used to

44 prevent these undesirable effects by oxidizing the phenolics which can then be removed by filtration (Giovanelli & Ravasini, 1993). As laccases are implicated in lignin formation by plants and lignin degradation by microorganisms, it is not surprising to find reports trying to harness laccases for modification of lignocellulose-based materials. In the pulp and paper industry, laccases can find use in producing materials with increased wet strength (Kudanga et al., 2011). Laccases can act on pulp, in conjunction with mediator compounds, to catalyze cross-linking reactions of lignin within fibres, leading to greater bonding between fibres. This greater inter-fibre bonding is thought to be the reason for increased wet strength. In addition to strength improvements, laccases can be used to impart other desirable properties onto lignocellulose materials by catalyzing grafting reactions. Grafting of phenolic compounds onto handsheets made of softwood kraft pulp resulted in the handsheets having antimicrobial activity; and grafting of lauryl gallate onto thermomechanical pulp resulted in increased hydrophobicity of the pulp (Kudanga et al., 2011). Finally, laccases have been used to catalyze formation of coloured polymeric phenolic compounds, which have then been used as dyes for lignocellulose-based fabrics such as flax fabrics (Kim et al., 2008). However, in this case, large loss of colour after washing and mechanical stress suggested no covalent bonds had formed between the coloured polymeric phenolics and the fibre. Aside from the cross-linking of compounds onto lignin-based materials, laccases have been employed for the delignification of pulps for applications such as paper making (Bourbonnais & Paice, 1992). The redox potential of laccases (approximately 0.8 V in the best cases) does not allow them to act directly on non-phenolic aromatic groups of lignin and the phenolic portion makes up no more than 20 % of the lignin (Canas & Camarero, 2010). Therefore, a mediator compound is first oxidized by the laccase and then that mediator oxidizes the lignin. In the early works by the group of Paice, ABTS was used as a synthetic mediator. Since then, other synthetic mediators containing -NO reactive groups have been used, such as 1- hydroxybenzotriazole (HBT) and N-hydroxyphthalimide (N-HPI) (Canas & Camarero, 2010). To minimize costs and use less toxic compounds, there have been efforts to identify low-cost natural mediators of laccase activity. These efforts have been inspired by the observation that phenolic compounds, from initial stages of lignin degradation by fungi, could serve as natural mediators of laccases during later stages of lignin degradation. Camarero et al. (2007) reported

45 the first use of natural phenolics (acetosyringone and syringaldehyde) to mediate laccase action on paper pulp lignin.

2.5. Research Hypotheses and Objectives

Taking an enzymatic approach towards modifying chemistry of phenolic phytochemicals lead to the following hypotheses for this research project: a) Organic solvent-stable arylesterases will be capable of carrying out synthetic reactions in hydrophobic organic media. b) The fairly shallow and exposed trimeric binding site of the small laccase (SCO6712 from Streptomyces coelicolor) can accept bioactive phenolics, including flavonols, as substrates for oxidation. Activity of this enzyme can be improved by using recombinant growth conditions that promote copper incorporation into the enzyme. c) Among the products produced by laccase action on flavonols will be dimerized flavonols. Since the B ring has been identified as the most reactive part of flavonols, the B ring of these flavonols will be involved in forming the dimer linkage. d) Laccase treatment of flavonols will produce new antioxidant products that have increased lipophilicity compared to the original flavonol.

The following specific objectives stemmed from the above hypotheses: a) Characterize the ability of one bacterial (PP3645 from Pseudomonas putida) and one archaeal (AF1753 from Archaeoglobus fulgidus) esterase, both stabilized by organic solvent, to also demonstrate activity in organic solvent. b) Characterize the ability of a bacterial laccase (SCO6712 from Streptomyces coelicolor) to act on bioactive phenolics as substrates for oxidation and the effect of varying cell cultivation conditions on protein copper content. c) Determine if oligomerized products can be produced from laccase SCO6712 action on flavonols and examine the structures of these products. d) Determine overall changes in the antioxidant activity that result from action of laccase SCO6712 on flavonols.

CHAPTER 3. CHARACTERIZATION OF SOLVENT-TOLERANT

CARBOXYLESTERASES WITH ARYLESTERASE ACTIVITY

Parts of this chapter are in preparation for submission:

Sherif, M., Wang, L., Tchigvintsev, A., Brown, G., Mavisakalyan, V., Tillier, E. R. M., Savchenko, A. V., Master, E. R., and Yakunin, A. F. Solvent-tolerant and thermophilic carboxylesterase with arylesterase activity from Archaeoglobus fulgidus.

M. Sherif carried out activity assays in organic solvent media, structural modelling of AF1753, and purification as well as kinetic characterization of AF1753 mutants. L. Wang performed stability tests in organic solvents, whereas G. Brown performed gene cloning and mutagenesis, and initial broad activity scans of alanine mutants were performed by A. Tchigvintsev.

3.1. Introduction

Esterase enzymes can catalyze the hydrolysis of ester linkages to produce a carboxylic acid and an alcohol. Under non-aqueous solvent conditions they can also catalyze the reverse reaction of esterification. In addition, they can catalyze transesterification, whereby an alcohol and an ester are reacted to produce a new alcohol and a new ester. Phenolic compounds contain hydroxyl groups that can be esterified or transesterified onto long-chain acyl groups to alter the chemistry of the phenolic, for example to increase the lipophilicity of the phenolic. The resultant lipophilized phenolic may display improved antioxidant activity towards lipids in an emulsified medium. For example, Vafiadi et al. (2008) used an esterase to produce lipophilic derivatives of sinapic acid, which displayed higher antioxidant activity towards low density lipoprotein (LDL) distributed in an aqueous phase. The challenge of carrying out enzymatic esterification or transesterification lies in the requirement to minimize water in the reaction mixture to avoid hydrolysis breaking down the desired ester product. This often necessitates running the enzyme reaction in non-aqueous solvents such as organic solvents or ionic liquids but such solvents can affect conformational mobility of the enzyme resulting in reduced activity (Klibanov, 1997). Since enzymes are purified in aqueous buffers, the process of removing water from the enzyme can also reduce enzyme activity (Klibanov, 1997). Nonetheless, enzymes have successfully been utilized for catalysis in organic solvents (examples given in the review by Klibanov, 2001). While the synthetic potential of lipases (which carry out the same reactions as esterases but favour longer-chain substrates) is often harnessed (Allevi et al., 1998; Chebil et al., 2007; Sabally et al., 2006; more examples given in the review by Figueroa-Espinoza & Villeneuve, 2005), there are comparatively fewer examples of the use of esterases to extend the valuable health benefits of phenolic compounds. Exceptions include transesterification of phenolic acid methyl esters with n-butanol by a feruloyl esterase from Fusarium oxysporum (Topakas et al., 2003), and esterification of glycerol with sinapic acid by a feruloyl esterase from Aspergillus niger (Vafiadi et al., 2008). However, in these two cases the enzyme used was a feruloyl esterase (EC 3.1.1.73), which is an enzyme that is known to act at the carboxylic end of hydroxycinnamic acids (Williamson et al., 1998) but not necessarily at the phenolic hydroxyl (Fig. 3.1). On the other hand, arylesterases (EC 3.1.1.2) are known to act at phenolic hydroxyl positions (Fig. 3.1), with phenyl acetate being a common substrate used in screening for

48 arylesterase activity (Fenster et al. (2003) and Shaw et al. (1994)). Arylesterases may therefore have applicability to a wider range of phenolic substrates compared to the feruloyl esterases.

O OH feruloyl esterase site of action

OH arylesterase site of action

Figure 3.1. Known sites of action of feruloyl esterases and arylesterases.

Previous characterization in our lab by Lijun Wang revealed a group of bacterial and archaeal esterases that showed arylesterase activity, and also showed organic solvent stability as assessed by pre-incubating in organic solvent followed by transferring to an aqueous solution to measure activity (Wang, 2009). These enzymes showed greater stability in hydrophobic (high log P) organic solvents compared to hydrophilic (low log P) organic solvents. As pre-incubation stability may not necessarily translate into activity in the presence of the organic solvent, my aim was to determine the activity of one of the archaeal (from Archaeoglobus fulgidus) and one of the bacterial (from Pseudomonas putida) enzymes in the presence of predominantly organic solvent. Their UniProt accession numbers are O28521 and Q88GS3 for the archaeal and bacterial enzyme, respectively. In this report the enzymes will be referred to using their gene names of AF1753 and PP3645, respectively. Based on activity and solvent stability work, these two enzymes represent the most promising candidates from a group of bacterial and archaeal enzymes. In the current work, the enzymes showed hydrolytic capability in organic solvent but were not able to carry out transesterification. As the lack of transesterification seemed due to limitations of the substrate binding pockets of the enzymes rather than general enzyme inactivity, alanine scanning mutagenesis of AF1753 (the more active enzyme of the two enzymes) was performed and mutants were characterized in an attempt to find mutations leading to higher activity on bulky long-chain p-nitrophenyl (pNP) esters.

3.2. Materials and methods

3.2.1. Gene cloning and protein purification

Briefly, genes were PCR amplified using the genomic DNA purchased from the ATCC, a proof reading DNA polymerase (Pfx polymerase), and the following PCR cycles: denaturation at 95 °C for 15 sec, annealing at 53 °C for 30 sec, and elongation at 68 °C for 1 min. PCR products were flanked by BseRI restriction sites to facilitate cloning into p15Tv-L (GenBank accession EF456736). Ligation products were transformed into E. coli DH5α by electroporation. Plasmids containing genes of interest were transferred to E. coli BL21- CodonPlus(DE3)-RIPL and E. coli transformants were cultured at 37 °C in a Terrific Broth (TB) medium. After the culture reached OD600 ~1, protein expression was induced using final concentration 100 mg/L isopropyl-β-D-thiogalactopyranoside (IPTG) and cell growth was continued at 16 °C overnight. Cells were harvested after overnight growth, suspended in binding buffer (300 mM NaCl, 50 mM HEPES pH 7.5, 5 % glycerol, 5 mM imidazole), passed through one freeze-thaw cycle, and then lysed by sonication. Cell extracts were cleared by centrifugation and incubated with Ni-affinity resin (Qiagen) for 2 h. The resin was then washed with 200 mL of washing buffer (300 mM NaCl, 50 mM HEPES, pH 7.5, 5 % glycerol, 30 mM imidazole) and eluted with approximately 5 mL elution buffer (300 mM NaCl, 50 mM HEPES, pH 7.5, 5 % glycerol, 250 mM imidazole). Protein concentrations were measured using the Bradford assay and their purities were evaluated by 12 % SDS-PAGE. Purified proteins were dialyzed against 10 mM Britton Robinson buffer at pH 8 (10 mM acetic acid, 10 mM phosphoric acid, 10 mM boric acid, pH adjusted with sodium hydroxide) containing 0.1 M NaCl , flash frozen in liquid nitrogen and then stored at -80 °C.

3.2.2. Hydrolytic activity of esterases AF1753 and PP3645 in t-amyl alcohol/water

(98:2, v/v)

The hydrolysis reaction was performed using 50 mM phenyl acetate or 50 mM vinyl acetate in t-amyl alcohol/buffer (98:2, v/v) where the buffer was 50 mM of Britton Robinson buffer (pH 9 and pH 8 for AF1753 and PP3645, respectively). AF1753 and PP3645 were present at 30 µg/mL and 100 µg/mL, respectively, and final reaction volumes were 200 µL. Reactions were started by mixing 196 µL of vinyl acetate or phenyl acetate prepared in t-amyl

50 alcohol with 4 µL of enzyme solution prepared in 50 mM Britton Robinson buffer. The reactions were run in triplicates and incubated at 37 oC for 96 h. Reaction samples (10 µL) were taken every 24 h and mixed with 190 µL of water (in the case of phenyl acetate) or 190 µL of a bromothymol blue (BTB) pH indicator solution (in the case of vinyl acetate) in sodium phosphate buffer (pH 7.3). After addition of BTB the sample solution would contain 5 mM sodium phosphate buffer (pH 7.3) and 0.15 mM BTB. The alcohol product of phenyl acetate hydrolysis was phenol, which was detected by spectrophotometry at 270 nm (Zeller, 1956). The carboxylic acid product of vinyl acetate hydrolysis was acetic acid, which was detected by measuring absorbance of the pH indicator BTB at 616 nm (Banerjee et al., 2003).

3.2.3. Transesterification activity of esterases AF1753 and PP3645 in t-amyl

alcohol/cyclohexane (1:9, v/v)

Enzymes were frozen as droplets using liquid nitrogen, transferred to 2 mL glass vials, and lyophilized for two days at -88 oC and 0.25 mbar. Dried enzymes were stored at 4 oC and used within three days. Transesterification reaction between 4’-hydroxyacetophenone and vinyl acetate was tested using the lyophilized enzymes. The reaction was performed using 50 mM 4’- hydroxyacetophenone and 250 mM vinyl acetate in t-amyl alcohol/cyclohexane (1:9, v/v) solvent. AF1753 and PP3645 were present at 2.5 mg/mL and 5 mg/mL, respectively, and final reaction volumes were 1 mL. Reactions were started by adding 1 mL of 50 mM 4’- hydroxyacetophenone and 250 mM vinyl acetate prepared in t-amyl alcohol/cyclohexane (1:9, v/v) solvent to lyophilized enzyme. The reactions were run in replicates of 4 and incubated rotating top-to-bottom at 8 rpm at 37 oC for 146 h. Reaction samples were taken at 0 h, 3 h, 5 h, 22 h, 47 h, and 146 h for product detection. To measure loss of 4’-hydroxyacetophenone by spectrophotometry at 270 nm, 10 µL of reaction sample was mixed with 190 µL t-amyl alcohol to make a first dilution. Then 10 µL of the first dilution was mixed with 90 µL of t-amyl alcohol to make a final dilution and absorbance was measured. To measure production of hydrolysis product (i.e. acetic acid) using BTB in a similar manner to what was done for hydrolysis reactions above (Section 3.2.2), 7 µL of reaction sample was mixed with 140 µL of 0.15 mM BTB solution in 5 mM sodium phosphate buffer pH 7.3. The mixture was spun down for 1 min to separate the organic and

51 aqueous layers, and 100 µL of the aqueous layer was transferred to a microtitre plate to measure absorbance at 616 nm. Production of transesterification product (i.e. 4’-acetoxyacetophenone) was also followed by thin layer chromatography (TLC). Reaction sample (1 µL) was spotted on a polyester-backed silica gel TLC plate containing a fluorescent indicator (Sigma). The plate was run for 20 min using 1:9 (v/v) acetone:cyclohexane mobile phase and viewed under 254 nm UV light. The substrate 4’-hydroxyacetophenone and the product 4’-acetoxyacetophenone were visualized as dark spots against a green colour of the fluorescent indicator.

3.2.4. Protein structure modeling and site-directed mutagenesis of esterase AF1753

The protein structure of esterase AF1753 was modeled based on an arylesterase from Pseudomonas fluorescens (pdb entry 1va4; 24.7 % sequence identity to AF1753) using the Protein Homology/analogY Recognition Engine (Phyre2) software (Kelley & Sternberg, 2009). Based on the model of AF1753, 12 residues predicted to be from the acyl binding region (Arg37, Glu244), alcohol binding region (Leu31, Tyr72, Met99, Phe124, Lys129, Leu189, Glu190), and catalytic triad (Ser98, Asp214, His243) of the enzyme were chosen and alanine point mutants generated. Site-directed mutagenesis was performed using the above plasmid containing AF1753 gene and the QuikChange site-directed mutagenesis kit from Stratagene. The sequences of primers used are shown in Supplemental Table A1.1. Mutant proteins were overexpressed in E. coli BL21(DE3) and purified as described above for the wild-type protein. PyMOL software (Version 1.7.1.0) was used to visualize protein structure and model effect of alanine point mutations on the structure of the protein binding pocket.

3.2.5. Activity of wild type and mutant AF1753 esterases on varying chain-length pNP

esters

The following pNP ester substrates were used to test activity of partially purified wild type and 12 point mutants of esterase AF1753 (carbon chain length of substrate acyl group in brackets): pNP acetate (C2), pNP propionate (C3), pNP butyrate (C4), pNP valerate (C5), pNP octanoate (C8), pNP decanoate (C10), pNP myristate (C14), pNP palmitate (C16). Reaction conditions were: 0.1 mL final volume, 2 µg/mL esterase, 1 mM substrate (0.5 mM for pNP palmitate due to solubility limits), 25 mM Britton Robinson buffer pH 8, 5 % (v/v) DMSO for

52 substrate solubilization, and 37 oC. Reactions were run in triplicates and absorbance readings were taken at 410 nm every 15 s up to 5 min.

3.3. Results and discussion

3.3.1. Hydrolytic activity of esterases AF1753 and PP3645 in t-amyl alcohol/water

(98:2, v/v)

The stabilizing effects of organic solvents on enzymes may end up leading to an active site that is not flexible enough to carry out catalysis (Dordick, 1992). Therefore, the hydrolytic activity of the two esterases in this study was examined in organic solvent. The solvent t-amyl alcohol was chosen because both enzymes were previously shown to retain over 50 % activity after pre-incubation in it (Wang, 2009). Also, the miscibility of t-amyl alcohol with water allows for avoiding the lyophilization of the enzymes and still maintaining a one-phase reaction with low levels of water introduced with the enzymes. Both enzymes AF1753 and PP3645 displayed hydrolytic activity in t-amyl alcohol/water (98:2, v/v) and activity continued even up to the last sampling time point of 96 h (Fig. 3.2). Work carried out by Lijun Wang demonstrated that after pre-incubating at 37 oC for 5 h in 50 % (v/v) t-amyl alcohol and then transferring to an aqueous medium to measure activity, PP3645 and AF1753 had 164 ± 26 % and 62 ± 41 % the activity, respectively, of the same enzyme that had been pre-incubated at 37 oC for 5 h in 50 mM potassium phosphate buffer (pH 7) (Wang, 2009). While Lijun Wang’s work showed that PP3645 was stabilized by 50 % t-amyl alcohol it was not yet clear if this stabilization leads to an inactive enzyme in the presence of t-amyl alcohol. The work in this thesis demonstrates that PP3645 is indeed active in the presence of t-amyl alcohol. Despite that Lijun Wang’s work showed AF1753 lost approximately 40 % activity after 5 h pre-incubation in 50 % t-amyl alcohol, in this thesis work AF1753 displayed a fairly consistent and linear rate of activity on phenyl acetate for up to 96 h in 98 % t-amyl alcohol (Fig 3.2). It should be kept in mind that the pre-incubation stability value of 62 % from Lijun Wang’s work has a large error value (± 41). Nonetheless, it may be that AF1753 is more stable in 98 % t-amyl alcohol than in 50 % t-amyl alcohol. Alternatively, it is possible that above a certain threshold the enzyme is stabilized by the organic solvent relative to being pre-incubated in aqueous solution, where the particular threshold value being dependent on the enzyme and solvent. It would be interesting to pre-

53 incubate PP3645 and AF1753 at levels above 50 % organic solvent to see if there are increases in pre-incubation stability. Both enzymes hydrolyzed vinyl acetate to a greater extent than phenyl acetate. After 96 h of reaction, PP3645 achieved only 2 % substrate conversion with phenyl acetate, whereas it achieved 20 % substrate conversion with vinyl acetate. On the other hand, AF1753 achieved 16 % and 44 % substrate conversion with phenyl acetate and vinyl acetate, respectively, after 96 h of reaction. This may be due to AF1753 having greater affinity for aryl esters compared to PP3645.

O O A) + H2O ⇌ + HO O OH vinyl acetate

O O + H2O ⇌ + B) O OH HO phenyl acetate

Vinyl acetate; PP3645 Vinyl acetate; AF1753

Phenyl acetate; PP3645 Phenyl acetate; AF1753 50

Substrate conversion (%) conversion Substrate 10

0 0 20 40 60 80 100

Time (h)

Figure 3.2. Hydrolysis reactions of A) vinyl acetate and B) phenyl acetate carried out in t-amyl alcohol/water (98:2, v/v) using esterases AF1753 and PP3645. n=3; error bars indicate standard deviation.

It was previously observed that enzymes in t-amyl alcohol with low levels of water were able to carry out catalysis (Zaks and Klibanov, 1988). One of the tested enzymes (yeast alcohol oxidase) required a minimum of approximately 5 % water for catalysis, while the other enzyme (mushroom polyphenol oxidase) required a minimum of approximately 1 % water (Zaks and Klibanov, 1988). The researchers determined that it was the amount of water bound to the enzyme that influenced activity and that the water content of the enzyme varied with different organic solvents that had the same water content. Hydrophobic organic solvents allowed the enzyme to retain more water than hydrophilic solvents with the same solvent water content. An enzyme with less bound water was hypothesized to have less flexibility (Zaks and Klibanov, 1988). This decreased flexibility may account for decreased activity in organic solvent but may also lead to enzyme stabilization (Klibanov, 2001). This stabilizing effect may be part of the reason that the esterases in this study were able to demonstrate activity for up to 96 h incubation in 98 % t-amyl alcohol. However, further testing will be required to compare activity levels after pre-incubation in water versus 98 % organic solvent for up to 96 h, to verify that the enzymes are stabilized under such conditions.

3.3.2. Transesterification activity of esterases AF1753 and PP3645 in t-amyl

alcohol/cyclohexane (1:9, v/v)

The transesterification activity of AF1753 and PP3645 was tested using a hydrophobic solvent-mix of t-amyl alcohol/cyclohexane (1:9, v/v). Preliminary experiments with 5 mg/mL of a commercial lipase as a positive control (Nicolosi et al., 1993) showed this solvent-mix to promote highest transesterification activity compared to an intermediate hydrophobicity solvent (100 % t-amyl alcohol) and a hydrophilic solvent-mix (t-amyl alcohol/DMSO (1:9, v/v)) (Fig. 3.3). The log P values for these three solvents are -1.22, 0.89, and 3.2 for DMSO, t-amyl alcohol, and cyclohexane, respectively (Wang 2009); with greater log P being one measure of greater hydrophobicity. This pattern of activity of the lipase agrees with previous studies that show that in some cases more hydrophobic solvents lead to greater activity due to less stripping of water from the enzyme (Klibanov, 1997). However, in other cases the level of partitioning of the substrate (Ps) and product (Pp) between water and the organic solvent were found to be important predictors of enzyme activity rather than the log P of the solvent (Yang and Robb,

1991, 1994). This was explained as Ps values that favour partitioning of an optimum concentration of substrate (i.e. the maximum substrate concentration before substrate inhibition reduces enzyme activity) to the aqueous phase surrounding the enzyme and a high Pp value that favours partitioning of the product away from the enzyme to the bulk solvent, leading to higher reaction rate (Yang and Robb, 1994). In the case of the arylesterases, neither lyophilized AF1753 nor lyophilized PP3645 displayed transesterification activity above background levels in t-amyl alcohol/cyclohexane (1:9, v/v) even after 146 h of reaction (Fig. 3.4). On the other hand, there was significant hydrolytic activity detected for both enzymes in the transesterification reaction mix (Fig. 3.5), with the required water for hydrolysis presumably coming from moisture absorbed from the air by the enzyme and/or organic solvent. This indicates that the enzymes still maintained activity in the water immiscible non-aqueous solvent after the lyophilization. It should be noted that, as expected, the reaction with lipase showed the transesterification product but it also showed the hydrolysis product, indicating that both transesterification and hydrolysis could occur in the same reaction mix (Fig. 3.3 and 3.5). Therefore, the absence of transesterification product with the esterases may be a consequence of lower ability of 4’-hydroxyacetophenone to access their active sites when the acyl-enzyme intermediate is formed with the acyl portion of vinyl acetate. In this way, only water molecules would have access to the active site and hydrolysis, but not transesterification, would occur.

O O O O + + HO O ⇌ HO O

vinyl acetate 4’-hydroxyacetophenone 4’-acetoxyacetophenone

Cyclohexane t-amyl alcohol DMSO A 100

70 B 60 50 40 C 30 20

Conversion of phenolic phenolic of Conversion (%) to product substrate 10 1 2 3 4 5 0 -10 0 20 40 60 80 Time (h)

Figure 3.3. Transesterification reaction of vinyl acetate and 4’-hydroxyacetophenone carried out in different solvent mixtures of t-amyl alcohol/co-solvent (1:9, v/v) using commercial lipase PS (from Amano). The TLC images on the right side show results for 72 h reaction samples for A) cyclohexane, B) t-amyl alcohol, and C) DMSO as co-solvents. In the TLC image the top spot is ester product while the bottom spot is phenolic substrate. The first lane from left is a standard mixture of 50 mM substrate and 50 mM product, the next two lanes are reaction samples without enzyme, and the last two lanes are reaction samples with enzyme. n=4; error bars indicate standard deviation.

O O O O HO O HO O

vinyl acetate 4’-hydroxyacetophenone 4’-acetoxyacetophenone

PP3645 AF1753 100 90

80 70 60 50

Conversion of phenolic phenolic of Conversion (%) to product substrate 20

0 0 50 100 150 -10 Time (h)

Figure 3.4. Transesterification reaction of vinyl acetate and 4’-hydroxyacetophenone carried out in t-amyl alcohol/cyclohexane (1:9, v/v) using recombinant esterase AF1753 and PP3645. The TLC image on the right side shows results for 146 h reaction samples. In the TLC image the top spot is the ester product while the bottom spot is the phenolic substrate. The first lane from left is a standard mixture of 50 mM substrate and 2 mM product, the next two lanes are reaction samples without enzyme, the next two lanes are reaction samples with PP3645 enzyme, and the last two lanes are reaction sample with AF1753 enzyme. n=4; error bars indicate standard deviation.

O O + H2O ⇌ + HO O OH

PP3645 AF1753 Lipase PS 35

Conversion of ester ester of Conversion (%) acid to acetic substrate 5

0 0 20 40 60 80 100 120 140 160 -5 Time (h)

Figure 3.5. Hydrolysis reaction of ester substrate in transesterification reaction-mix. The reaction-mix contains vinyl acetate and 4’-hydroxyacetophenone in t-amyl alcohol/cyclohexane (1:9, v/v). n=4; error bars indicate standard deviation.

A possible reason for reduced accessibility of the arylesterase active sites might be found by looking to previously identified effects of solvents on enantioselectivity. The level of enantioselectivity of an enzyme may change by changing from one organic solvent to another (Wescott & Klibanov, 1994). The enzyme subtilisin Carlsberg was used to carry out transesterification of chiral sec-phenethanol with vinyl butyrate in different organic solvents and it was observed that the enantioselectivity of the enzyme increased with decreasing dielectric constant of the solvent (Fitzpatrick & Klibanov, 1991). The research group carried out subsequent modeling studies of the binding of sec-phenethanol to the enzyme which showed that the R enantiomer makes more close contacts (which may indicate greater steric hindrance) with the enzyme binding site compared to the S enantiomer (Fitzpatrick et al., 1992). Taken together, these results supported the possibility that in solvents with low dielectric constants, the greater rigidity of the enzyme causes it to have greater difficulty accepting the R enantiomer into the active site compared to the S enantiomer. This would then manifest as a greater loss in reactivity towards the R enantiomer compared to the S enantiomer (Wescott & Klibanov, 1994). In a similar manner, the esterase enzymes used in this study may have had a more constrained active site compared to the commercial lipase, which would then be compounded by the effects of the organic solvent. This may partly explain the inability of the esterases to carry out both hydrolysis and transesterification. Their more constrained active sites would permit access to water (to carry out hydrolysis) but not the more bulky phenyl substrate (to carry out transesterification). Therefore, alanine point mutants in the binding pocket of AF1753 were generated in an effort to identify amino acid positions that impact enzyme activity towards medium to long-chain esters. An alternative way to test if the esterase active site is less accommodating could be to use less bulky non-phenolic alcohol substrates, such as methanol or ethanol as opposed to 4’-hydroxyacetophenone, in the transesterification reaction. However, observation of transesterification activity on such substrates would not clarify if the limitation for the 4’-hydroxyacetophenone substrate is in fact accessibility to the active site. Additionally, the antioxidant phytochemical phenolics that are intended to be ultimately used for transesterification reactions have multiple substitutions and are even more bulky than 4’- hydroxyacetophenone. Therefore, from a practical perspective, it would be desirable to identify if the esterase can be made to transesterify 4’-hydroxyacetophenone as a relatively simple phenolic model substrate.

3.3.3. Activity of wild type and mutant AF1753 esterases on varying chain-length pNP

esters

Nickel-affinity protein purification yielded partially purified enzyme variants. The expected molecular weight of AF1753 is 30 kD and protein gels of partially purified wild type and mutant variants show the presence of the expected enzyme along with two major contaminating bands at approximately 25 kD and approximately 70 kD (Supplemental Fig. A1.1). To identify best candidates for further protein purification, activity assays were run with partially purified enzymes. In this case, relative activity of each enzyme was assessed, to identify those with improved activity on longer chain substrates. Whereas the wild type enzyme showed approximately 20 % relative activity on longer chain (greater than 8 carbon acyl group) pNP substrates compared to pNP acetate, the relative activity of two mutants, Arg37Aal and Glu244Ala, on longer chain pNP substrates compared to pNP acetate was above 50 % (Fig. 3.6). Additionally, mutant Arg37Ala showed higher activity than wild type for pNP esters with longer than 8 carbon acyl groups. However, even though the relative activity of Glu244Ala on long-chain pNP substrates was increased compared to the wild type enzyme, the measured activity of Glu244Ala on all tested substrates was consistently lower than wild type. Looking at the model of the protein structure, the side chain of Glu244 is predicted to contribute to one side of the acyl binding pocket of the enzyme near the pocket entrance (Fig. 3.7A), while Arg37 is predicted to contribute to the inner face of the acyl binding pocket (Fig. 3.7A). Upon mutation of Glu244 to alanine, it is anticipated that more room would be available to accommodate longer acyl groups; notably, a tunnel forming a second entrance to the active site is also predicted (Fig. 3.7B). Similarly, upon mutation of Arg37 to alanine, the acyl binding pocket is expected to widen and again a second entrance to the active site is predicted (Fig. 3.7C). Using the CASTp server (http://sts.bioe.uic.edu/castp/) to quantify the volume of the substrate binding pockets (Dundas et al., 2006) of the predicted AF1753 structure leads to values of 619.68 Å3, 1260.87 Å3, and 1201.80 Å3 for the wild type, Arg37Ala, and Glu244Ala, respectively; indicating almost twice as much space is available in the mutants. However, it is also likely that the newly formed tunnel for the Glu244Ala mutant poorly positions the acyl moiety of the ester linkage relative to the catalytic serine (Ser98) (Fig. 3.7B).

In Glu244Ala, as the catalytic serine attacks the carbonyl carbon from the opposite side of the carbonyl oxygen to form the acyl-enzyme tetrahedral intermediate, the carbonyl oxygen is directed towards the newly formed tunnel. This leaves less space available to accommodate the acyl group of the substrate. On the other hand, in the case of Arg37Ala, the new tunnel is not positioned on the same line as the line of attack of the catalytic serine to the carbonyl carbon of the substrate. Therefore, the tunnel seems favourably positioned in relation to Ser98 to allow taking advantage of the increased space to accommodate the acyl group of the substrate (Fig. 3.7C). While the arylesterases used in this study do not appear as promising tools for transesterification, the observation that the enzymes carried out hydrolysis in organic solvent leads to the idea that, as an alternative to mutagenesis, other organic solvent-tolerant arylesterases with preferential activity on long-chain substrates can be tested for transesterification reactions of phenolics in organic solvent. In this regard, the thermostable arylesterase from Sulfolobus solfataricus was shown to have better binding and activity on mid chain-length esters such as pNP caprylate (8-carbon chain) over short-chain ester pNP butyrate (4-carbon chain) (Park et al., 2008). In addition, this enzyme displayed approximately 90 % residual activity after 1 h pre-incubation in 90 % water-miscible organic solvents. As such, this enzyme may represent a good candidate to compare with AF1753.

A) 200 WT Leu31Ala Arg37Ala Tyr72Ala Met99Ala Phe124Ala

Lys129Ala Leu189Ala Glu190Ala 150 Glu244Ala

100

50 its activity on C2 (%) its activity

Activity of mutant of relative to Activity

0 C2 C3 C4 C5 C8 C10 C14 C16 pNP ester substrate

WT Arg37Ala Glu244Ala 10

Specific activity (µmol/min/mg) activity Specific 0 C2 C3 C4 C5 C8 C10 C14 C16 pNP ester substrate

Figure 3.6. Activity of wild type and mutants of AF1753 on pNP substrates. A) Activity of each mutant on each substrate relative to that mutant’s activity on pNP acetate. B) Specific activity of wild type and two mutants on pNP substrates. Labels on horizontal axes indicate length of carbon chain on acyl group of pNP ester substrate. n=3; error bars indicate standard deviation.

R37 E244 90o

R37

E244 90o

wild type wild type active site active site entrance entrance B) new access to new access active site to active A244 R37 site 90o

acyl R37

O O A244 90o + - N O O wild type wild type active site active site entrance entrance C) new access to new access active site to active E244 A37 90o site

A37

E244 90o

wild type wild type active site active site entrance entrance

Figure 3.7. Images of predicted protein structure of AF1753. Left side images and right side images show surface and ribbon views, respectively. A) Wild type enzyme, B) Glu244Ala mutant, and C) Arg37Ala mutant. In green are the catalytic Ser96, His243, and Asp214 (only Ser96 shown in surface views for clarity). In magenta are the Glu244, Arg37, and the corresponding mutants Ala244 or Ala37. For the ribbon view, in yellow are the oxyanion hole residues His29 and Gly30; in cyan are residues that line the substrate binding pocket; and in black are the residues that line the entrance to a second tunnel that becomes available to access the active site upon mutation of Arg37 or Glu244 to alanine. Mutation was modeled using PyMOL mutagenesis function. Surface images are shown with part of the front surface cut away to reveal the inner surface (coloured black) of the substrate binding pocket. The normal wild type entrance to the binding pocket is at the bottom of the images (shown in bracketed region for ribbon view). For both mutants, a tunnel forming a second entrance (top left of image) to the binding pocket is predicted (entrance shown in bracketed region for ribbon view). The skeletal formula of a pNP ester, at the expected position that would be occupied by a substrate undergoing hydrolysis in the substrate binding pocket, is drawn above the surface image of each mutant. For ribbon images, residues 132 - 140 and 178 - 190 were removed to show a clear view of the active site. Images were generated with PyMOL software.

3.4. Conclusions

In this work, both PP3645 and AF1753 enzymes displayed hydrolytic activity in t-amyl alcohol/water (98:2, v/v) and also in t-amyl alcohol/cyclohexane (1:9, v/v), indicating that use of the enzymes in organic solvents does not eliminate activity. To the best of my knowledge, this work is the first demonstration of activity of arylesterases in over 90 % organic solvent media. However, neither enzyme carried out transesterification in the t-amyl alcohol/cyclohexane (1:9, v/v) solvent mix. The catalysis of hydrolysis but not transesterification suggested that the enzyme binding pockets are not accommodating enough to allow access of the phenolic (for transesterification) to compete with access of water (for hydrolysis) to the active site. Therefore, site-directed mutagenesis was used to expand the active site of the more active arylesterase AF1753. Assays using partially purified alanine point mutants of esterase AF1753 demonstrated mutant variants (Arg37Ala and Glu244Ala) with improved relative activity on longer chain pNP esters compared to the relative activity of the wild type. Future experiments should assay the transesterification ability of the Arg37Ala mutant.

CHAPTER 4. BIOCHEMICAL STUDIES OF THE MULTICOPPER OXIDASE

(SMALL LACCASE) FROM STREPTOMYCES COELICOLOR USING BIOACTIVE

PHYTOCHEMICALS AND SITE-DIRECTED MUTAGENESIS

Parts of this chapter are published in:

Sherif, M., Waung, D., Korbeci, B., Mavisakalyan, V., Flick, R., Brown, G., Abou-Zaid, M., Yakunin, A.F., and Master, E. R. (2013). Biochemical studies of the multicopper oxidase (small laccase) from Streptomyces coelicolor using bioactive phytochemicals and site-directed mutagenesis. Microbial Biotechnology, 6(5), 588-597.

M. Sherif carried out protein production and purification, as well as all activity and kinetic assays on different substrates. Gene cloning and site-directed mutagenesis was performed by G. Brown and B. Korbeci. Inductively coupled plasma atomic emission spectrometry was done by the Department of Chemistry University of Toronto.

4.1. Introduction

Laccases are multicopper oxidases (MCOs) that use copper cofactors and oxygen to oxidize aromatic and non-aromatic compounds, producing water as a byproduct. The catalytic site contains four copper atoms coordinated to various amino acid residues (often histidine) within the protein (Sharma et al., 2007). Although most laccases characterized to date were isolated from fungi, bacterial MCOs with laccase activity offer advantages in terms of heterologous expression and purification. With their ability to act on aromatic compounds, laccases represent an alternative means to esterases for altering chemistry of phenolics. One of the advantages of using laccases over esterases is that the reaction can be carried out under aqueous conditions. However, one of the drawbacks is that there is a greater diversity of products that can arise from oxidation compared to transesterification. Upon action of laccases on phenolic compounds, a phenoxy radical is formed which can then react to produce a variety of downstream products. Among the products are dimer and higher level oligomers of the phenolic (Adelakun et al., 2012; Carunchio et al., 2001; Ponzoni et al., 2007; Gavezzotti et al., 2014). These higher molecular weight products may display increased hydrophobicity, which may lead to enhanced antioxidant effects towards emulsified lipids. For example, Lotito et al. (2000) found that naturally occurring catechin oligomers (up to pentamer) had higher antioxidant protection than the catechin monomer in protecting against AMVN-induced oxidation of L-α-phosphatidylcholine liposomes in buffer. Furthermore, the (n- octanol)-water partition coefficient of these compounds demonstrated increasing hydrophobicity with increasing degree of oligomerization after the dimer (Plumb et al., 1998) (the dimer did not have significant difference in partition coefficient compared to the monomer). I used a previously identified and characterized bacterial MCO from Streptomyces coelicolor (previously called SLAC for Small LACcase) (Machczynski et al., 2004; Skálová et al., 2009) that has laccase activity to evaluate its potential for derivatization of plant phenolic compounds. The UniProt accession number for the enzyme is Q9XAL8 and from here on I will refer to it using its gene name of SCO6712. This enzyme contains two cupredoxin domains as opposed to the typical three cupredoxin domains seen for most laccases. Based on crystal structure analysis and activity assay after protein gel, this enzyme seems to form a trimer as the main active unit (Skálová et al., 2009). However, earlier studies on this enzyme using activity

69 assay after protein gel electrophoresis suggested that a dimer may also be an active form (Machczynski et al., 2004). Large-scale application of laccases is challenged by the cost of enzyme production, as well as enzyme inhibition at typical industrial process conditions (Kunamneni et al., 2008). While most laccases now in use were isolated from mesophilic fungi, bacterial laccases could present a source of low-cost enzymes (Santhanam et al., 2011). In addition to being readily expressed in recombinant hosts and purified, thermostable laccases from bacteria may represent particularly robust industrial catalysts. As was previously reported, the ease of production in E. coli, thermostability, and reported activity on 2,6-dimethoxyphenol (DMP) at alkaline pH, mean that SCO6712 could be an ideal candidate for numerous biotechnology applications, including organic syntheses (Machczynski et al., 2004). Accordingly, in the present study, SCO6712 was selected for detailed biochemical characterization on a broad range of compounds, including bioactive phytochemicals. The availability of the SCO6712 structure also facilitated site- directed mutagenesis, which was performed to confirm key residues involved in copper binding, and to predict those that could be modified to alter the reactivity or substrate selectivity of this enzyme. Finally, since microaerobic cultivation conditions were previously shown to promote the incorporation of copper ions by CotA when recombinantly expressed in E. coli (Durão et al., 2008a), the effect of cultivation conditions on the yield and specific activity of SCO6712 was evaluated.

4.2. Materials and methods

4.2.1. Gene cloning and protein purification

Genes were PCR amplified using the genomic DNA purchased from the ATCC, a proof reading DNA polymerase (Pfx polymerase), and the following PCR cycles: denaturation at 95 °C for 15 sec, annealing at 53 °C for 30 sec, and elongation at 68 °C for 1 min. PCR products were flanked by BseRI restriction sites to facilitate cloning into p15Tv-L (GenBank accession EF456736). Ligation products were transformed into E. coli DH5α by electroporation. Plasmids containing genes of interest were transferred to E. coli BL21(DE3) Gold strain (Stratagene) and E. coli transformants were cultured under standard aerobic conditions with shaking (200 r.p.m.) at 37 °C in a Terrific Broth (TB) medium. After the culture reached OD600

~1, protein expression was induced using final concentration 100 mg/L isopropyl-β-D- thiogalactopyranoside (IPTG). CuCl2 (final 1 mM) was also added and cell growth was continued at 16 °C overnight with agitation of the culture during growth (aerobic conditions) or without agitation of the culture during growth (microaerobic conditions). Cells were harvested after overnight growth, suspended in binding buffer (300 mM NaCl, 50 mM HEPES pH 7.5, 5 % glycerol, 5 mM imidazole), passed through one freeze-thaw cycle, and then lysed by sonication. Cell extracts were cleared by centrifugation and incubated with Ni-affinity resin (Qiagen) for 2 h. The resin was then washed with 200 mL of washing buffer (300 mM NaCl, 50 mM HEPES, pH 7.5, 5 % glycerol, 30 mM imidazole) and eluted with approximately 5 mL elution buffer (300 mM NaCl, 50 mM HEPES, pH 7.5, 5 % glycerol, 250 mM imidazole). Protein concentrations were measured using the Bradford assay and their purities were evaluated by 12 % SDS-PAGE. Purified proteins were dialyzed against 50 mM HEPES buffer (pH 7.5) containing 0.1 M NaCl , flash frozen in liquid nitrogen and then stored at -80 °C.

4.2.2. Site-directed mutagenesis

Site-directed mutagenesis was performed using the above plasmid containing SCO6712 gene and the QuikChange site-directed mutagenesis kit from Stratagene. The sequences of primers used are shown in Supplemental Table A2.1. Mutant proteins were overexpressed in E. coli BL21(DE3) and purified as described above for the wild-type protein. Transformants were cultivated in the presence of 1 mM CuCl2.

4.2.3. Copper content analysis of wild-type and mutant SCO6712 laccases

Wild-type and point mutants of SCO6712 laccase were buffer-exchanged into 10 mM HEPES buffer (pH 7.5) using Nanosep (Pall) polyethersulfone ultracentrifugation tubes (10 kD cutoff). Copper content of buffer-exchanged proteins (in the retentate) was then determined using inductively coupled plasma atomic emission spectrometry.

4.2.4. Substrate profile of wild-type SCO6712 and the Ser292Ala mutant laccases

The following (mostly phenolic) substrates were used to test activity of wild-type SCO6712 laccase: ABTS; sodium ferrocyanide; N-hydroxyphthalimide (N-HPI); phenol; 4-

71 hydroxybenzyl alcohol; acetovanillone; 2,3-dihydroxybenzoic acid (2,3-DHB); 2,6- dimethoxyphenol (2,6-DMP); gallic acid (3,4,5-trihydroxybenzoic acid); syringaldehyde; syringic acid (4-hydroxy-3,5-dimethoxybenzoic acid); 3-(3,4-dihydroxyphenyl)-L-alanine (L- DOPA); p-coumaric acid; o-coumaric acid; m-coumaric acid; caffeic acid; ferulic acid; resveratrol; apigenin; quercetin; morin; kaempferol; myricetin; and rutin. Activity was tested generally at a range of pH 2 to 10; with some substrates (apigenin, quercetin, morin, kaempferol, myricetin, and rutin) not testable at pH lower than 7 due to substrate solubility limitations. Reaction conditions were: 0.2 mL final volume, 18.5 µg/mL laccase, 1 mM substrate, 50 mM Britton Robinson buffer, and 60 oC. Reactions for substrates N-HPI; 4- hydroxybenzyl alcohol; acetovanillone; and 2,3-DHB contained 5 % (v/v) ethanol for substrate solubilization. Reactions for gallic acid; syringaldehyde; syringic acid; p-coumaric acid; caffeic acid; ferulic acid; resveratrol; apigenin; quercetin; morin; kaempferol; myricetin; and rutin contained 5 % (v/v) dimethyl sulfoxide (DMSO) for substrate solubilization. Reactions were performed in triplicate and were initiated by adding 180 µL substrate solutions to 20 µL enzyme; controls without enzyme were also included. Reactions were run for 20 min and product formation or substrate depletion was measured using spectrophotometry by initially scanning from 250 nm to 790 nm. Reactions showing change in absorbance were subsequently followed at a specific wavelength at the optimal pH. Wavelengths and molar absorption coefficients used to measure -1 -1 oxidative activity were: ABTS at pH 4, ε420 = 36 mM cm (Johannes & Majcherczyk, 2000); -1 -1 sodium ferrocyanide at pH 5, ε405 = 0.9 mM cm (Machczynski et al., 2004); 2,6-DMP at pH -1 -1 -1 -1 8, ε468 = 14.8 mM cm (Machczynski et al., 2004); L-DOPA at pH 8, ε475 = 3.6 mM cm -1 -1 (Laufer et al., 2006); syringic acid at pH 8, ε270 = 9.3 mM cm (calculated from standard -1 -1 curve); caffeic acid at pH 7, ε330 = 16.3 mM cm (calculated from standard curve); ferulic acid -1 -1 at pH 8, ε287 = 18.8 mM cm (calculated from standard curve); resveratrol at pH 8, ε308 = 26.8 -1 -1 -1 -1 mM cm (calculated from standard curve); quercetin at pH 7, ε385 = 17.2 mM cm -1 -1 (calculated from standard curve); morin at pH 7, ε385 = 18.4 mM cm (calculated from -1 -1 standard curve); kaempferol at pH 8, ε385 = 15.1 mM cm (calculated from standard curve); -1 -1 and myricetin at pH 7, ε385 = 15.9 mM cm (calculated from standard curve). Compounds without a previously reported extinction coefficient for product were followed by measuring substrate loss using a standard curve generated at the chosen wavelength. Substrates for which

72 activity was seen with the wild-type laccase were subsequently tested with the Ser292Ala mutant laccase.

4.2.5. Kinetics of wild-type and Ser292Ala laccases on select substrates

Kinetic parameters were obtained at 60 oC using ten substrate concentrations (0.1 mM to 10 mM for ABTS and 2,6-dimethoxy phenol; 0.1 mM to 3.5 mM for quercetin; and 0.1 mM to 4 mM for morin and myricetin) at the optimal pH for each substrate. Reactions were performed in triplicates and were initiated by adding 90 μL substrate solutions to 10 μL enzyme (final 100 μL reaction volume with 6 µg/mL laccase enzyme) and initial rates were obtained by measuring product formation at 0 min, 1 min, 3 min, 5 min, and 10 min. Kinetic parameters were calculated using the Michaelis-Menten equation (GraphPad Prism5 Software).

4.2.6. Docking of 2,6-dimethoxyphenol substrate to wild type and Ser292Ala SCO6712

laccase

The protein structure of laccase SCO6712 (pdb entry 3cg8) and the ligand 2,6- dimethoxyphenol (pdb entry 3fu7) were used for docking analysis. PyMOL software (Version 1.7.1.0) was used to model Ser292Ala point mutation on the structure of the laccase protein. Docking was performed with AutoDock.4 software (Morris et al., 1998). Ligand and protein were prepared for docking using AutoDockTools 1.5.6. Crystallographic water and copper atoms (both considered non-bonded atoms by the software) were removed from the protein and all hydrogen atoms were added. All hydrogen atoms were also added for the ligand. AutoDockTools 1.5.6 added gasteiger charges for the ligand and merged non-polar hydrogens for both ligand and protein. AutoGrid4 was used to pre-calculate pairwise interaction energies between atoms in ligand and protein. A Lamarckian genetic algorithm was then used by AutoDock.4 for determining optimal docked configurations between ligand and protein.

4.3. Results and discussion

4.3.1. Effect of microaerobic cultivation on copper content and activity of SCO6712

laccase

The copper content of purified MCOs can be incomplete and depends on the cultivation medium, temperature, and oxygen concentration (Galli et al., 2004; Durao et al., 2008a). Therefore, as previously done for CotA (Durao et al., 2008a), SCO6712 was prepared from microaerobic cultivations (growing the E. coli overnight at 16 oC without agitation of the cell culture during growth after IPTG induction) in an effort to increase the content of copper ions in the purified enzyme. Indeed, it was found that microaerobic growth conditions increased the copper content in SCO6712 from 0.5 to 1.2 moles of Cu2+ per mole of enzyme, when compared to enzyme from aerobic cultivation conditions (with continuous shaking of cells after IPTG induction). Increased copper content correlated with an increase in specific activity on 1 mM ABTS, from 110 ± 10 nmol min-1 mg-1 to 980 ± 30 nmol min-1 mg-1. These results confirm that similar to previous studies with CotA (Durao et al., 2008a; Mohammadian et al., 2010), microaerobic cultivation conditions promote the incorporation of Cu2+ in SCO6712.

4.3.2. Substrate profile of wild-type SCO6712 laccase

To date, activity measurements of SCO6712 have been reported for only a handful of compounds, including 2,6-dimethoxyphenol (2,6-DMP) (Machczynski et al., 2004). In the current study, SCO6712 was tested for oxidase activity against 24 different substrates, including natural, bioactive phenolic compounds. Consistent with results reported in Machczynski et al (2004), SCO6712 prepared herein oxidized 2,6-DMP, sodium ferrocyanide, and L-DOPA (Fig. 4.1). The current analysis also confirmed comparable substrate consumption by SCO6712 of resveratrol, quercetin, morin, myricetin, and kaempferol relative to ABTS after 20 min of reaction (Fig. 4.1). Overall, SCO6712 showed approximately 30 times higher substrate consumption after 20 min toward the best substrate sodium ferrocyanide (783 ± 17 µM) as compared with the poorest substrate 2,6-DMP (27.0 ± 0.2 µM). SCO6712 activity on different substrates was pH-dependent and optimum activities were detected between pH 4 and 8 (Fig. 4.1). The activity of SCO6712 on phenolic substrates was highest at alkaline pH, which is consistent with previous analyses of SCO6712 on 2,6-DMP (Machczynski et al., 2004), and the predicted influence of pH on the redox potential of phenol substrates (Xu 1997). In particular, new insight gained from this work as compared to the previous analysis of SCO6712 by Machczynski et al. (2004) includes the ability of SCO6712 to oxidize comparatively large phenolic substrates with multiple hydroxyl groups (like quercetin) in addition to smaller

74 phenolics (like syringic acid). SCO6712 activity on polyaromatic compounds may be explained by the potential of unpaired electrons to delocalize after oxidation of the phenolic compound, leading to stabilized radical products. The electronic contribution of hydroxyl and/or methoxy substituents also generates phenoxy moieties that are more easily oxidized (Xu, 1996). This phenomenon could explain why SCO6712 oxidized caffeic acid and ferulic acid but not p- coumaric acid, which lacks a hydroxyl or methoxy substituent ortho to the single hydroxyl (Fig. 4.2). Consistent with this hypothesis, reported redox potential values at pH 7.4 for p-coumaric acid, ferulic acid and caffeic acid are 0.66 V, 0.50 V, and 0.36 V, respectively (Jørgensen & Skibsted, 1998). Similarly, SCO6712 activity was observed on kaempferol but not apigenin. Apigenin differs from kaempferol by lacking a hydroxyl group adjacent to the carbonyl carbon (Fig. 4.2) and the reported redox potentials of kaempferol and apigenin at pH 7.4 are 0.39 V and 0.71 V, respectively (Jørgensen & Skibsted, 1998).

Na+ Na+ N N N OH 90 + 4- + HO + Na Fe Na NH4 wild type OH - OH O3S N N 80 OH N Ser292Ala OH O

HO O S N OH OH 70 C2H5 O O HO OH N O O OH N HO OH HO OH 60 O OH HO OH S N C2H5 O OH HO OH OH 50 O

- O O3S + NH4 OCH 40 3 HO OH O OH OH OH 30 O OH OH

Substrate loss (% of initial) loss (% Substrate NH2 H3CO OCH3 H CO OH 20 3 OH HO OH 10 H3CO

ABTS

Morin

L-DOPA

2,6-DMP

Myricetin Quercetin

Sodium

Ferulic acid Ferulic

Resveratrol

Caffeic acid Caffeic

Kaempferol Syringic acid Syringic ferrocyanide pH 4 5 8 8 8 7 8 8 7 7 8 7

Figure 4.1. Substrate selectivity of SCO6712. Substrate oxidation by purified wild-type SCO6712 (dark grey bars) and variant Ser292Ala (light grey bars) on various substrates. Wild-type SCO6712 and the Ser292Ala variant were prepared using microaerobic growth conditions. Substrate depletion (µM) was measured after 20 min at 60 oC and the optimal pH for the reaction. Initial substrate concentrations were 1000 µM. n = 3; errors indicate standard deviation.

OH OCH OH 3 OH OH H O HO HO O O O

p-coumaric acid caffeic acid ferulic acid

OH OH

HO O HO O

OH OH O OH O apigenin kaempferol

Figure 4.2. Chemical structures of natural bioactive phenolic substrates. Structures are shown for p-coumaric acid, caffeic acid, ferulic acid, apigenin, and kaempferol.

4.3.3. Kinetics of wild-type SCO6712 laccase on select substrates

The kinetic parameters for SCO6712 were obtained at 60 oC for ABTS, 2,6-DMP, quercetin, morin and myricetin (Table 4.1); it was not possible to determine kinetic parameters for other substrates due to low solubility or high background absorbance of the compound at required substrate concentrations. The Km values for SCO6712 were comparable to those reported for the CueO multicopper oxidase from E. coli (YacK) (Kim et al., 2001); Km and kcat values of SCO6712 for 2,6-DMP were also consistent with those reported by Machczynski et al.

(2004). However, in this study SCO6712 showed approximately two times lower Km and three times lower kcat for 2,6-DMP compared with the previous work (Machczynski et al., 2004) suggesting that the SCO6712 in this study has a stronger binding affinity but lower maximum turnover of this compound. This could be due to differences in copper content of SCO6712 produced in this work compared to earlier reports. For instance, since type 2 and type 3 copper ions are coordinated by neighboring SCO6712 monomers, and that the coordination of type 2 and type 3 copper ions is predicted to promote the formation of the substrate binding site (Skálová et al., 2009), copper content is likely to influence the folding and flexibility of the substrate cleft. The Km of SCO6712 for quercetin, morin, and myricetin (1.2 ± 0.4 mM; 1.5 ±

0.2 mM; and 0.9 ± 0.3 mM, respectively) were comparable to those for ABTS and 2,6-DMP (0.8 ± 0.1 mM and 1.1 ± 0.2 mM, respectively) (Table 4.1), indicating that SCO6712 effectively binds and oxidizes higher molecular weight phenolics. Notably, SCO6712 displayed highest kcat values with myricetin (8.1 ± 0.8 s-1), which is consistent with the contribution of small, electron donating substitutions on electron transfer kinetics (Xu, 1996). Similar kcat values with ABTS, quercetin and morin, but lower ABTS transformation after 20 min (Fig. 4.1), suggests products of ABTS oxidation might be inhibitory.

Table 4.1. Kinetic parameters of wild type SCO6712 and the Ser292Ala variant enzyme. Lighter shaded regions in quercetin, morin, and myricetin molecules indicate identical chemical structures. n = 3; errors indicate standard deviation.

-1 Substrate pH Enzyme Km (mM) kcat (s ) kcat/Km (mM-1s-1) ABTS C2H5 4 wt 0.8 ± 0.1 7.7 ± 0.3 10 ± 1.0 C H 2 5 N S292A 2.4 ± 0.2 11.1 ± 0.3 4.6 ± 0.4 N N N S -O3S -O3S S 2,6-DMP H3CO 8 wt 1.1 ± 0.2 4.0 ± 0.2 3.5 ± 0.6

HO S292A 0.12 ± 0.02 0.93 ± 0.02 8 ± 1

H3CO

Quercetin HO O 7 wt 1.2 ± 0.4 5.2 ± 0.7 4.3 ± 1.5 OH a HO S292A 4.0 ± 2.0 13.0 ± 4.0 3.0 ± 0.1 O HO OH Morin HO O 7 wt 1.5 ± 0.2 5.3 ± 0.4 3.6 ± 0.6 OH HO S292A 1.6 ± 0.4 4.0 ± 0.4 2.5 ± 0.7 O OH OH Myricetin HO HO O 7 wt 0.9 ± 0.3 8.1 ± 0.8 9.0 ± 3.0 OH a HO S292A 3.0 ± 2.0 13.0 ± 4.0 3.1 ± 0.4 O HO OH a kinetic values are estimates since measurements could only be acquired under non-saturating conditions due to reaching the solubility limits of the substrate.

4.3.4. Site-directed mutagenesis

Previous studies with CotA demonstrated the ability to manipulate the redox potential and catalytic efficiency of this enzyme through substituting amino acids involved in coordinating or stabilizing the type 1 copper site (Durao et al., 2006; Durao et al., 2008b). Continuing from previous work by a Master’s student in our lab, 17 amino acids were individually mutated to alanine, and proteins produced under microaerobic conditions (without agitation of the cell culture during growth), to identify the residues important for SCO6712 activity, as well as to confirm key residues involved in copper binding, and to predict those that could be targeted to alter the substrate specificity of this enzyme (Supplemental Table A2.1). Notably, Tyr229Ala and Tyr230Ala substitutions in aerobically produced enzymes caused increased SCO6712 activity on ABTS by more than 10 times compared to the wild-type enzyme also prepared using aerobic growth conditions. The Tyr230 and Tyr229 residues occupy most of the shallow trimeric substrate binding site at the surface of the enzyme (Skálová et al., 2009). The copper content of the Tyr230Ala mutant produced under aerobic conditions was 2.3 moles of Cu2+ per mole of enzyme, 4.5 times higher than the wild-type enzyme produced under aerobic conditions and twice that of wild-type SCO6712 produced under microaerobic conditions. Despite having higher copper content than wild-type SCO6712 produced using microaerobic conditions, the activity of Tyr230Ala with 1 mM ABTS was 1.2 µmoles min-1 mg- 1, which is comparable to the activity of wild-type SCO6712 produced under microaerobic conditions. These results suggest that the Tyr230Ala substitution likely reduces SCO6712 activity on ABTS, but that increased copper binding can compensate for the loss in activity. In this case, further substitution of Tyr230 might identify residues that promote the cooperative loading of copper ions and production of fully activated forms of the enzyme. While microaerobic cultivation conditions did not increase the copper content of the Tyr230Ala mutant, the molar ratio of copper to protein in Ser292Ala and Glu228Ala increased from 0.5 and 0.4 in aerobic preparations, to 1.5 and 0.9 in microaerobic preparations. In these cases, the increase in copper content also led to higher enzyme activities, where the activity of Ser292Ala on 1 mM ABTS at pH 4.0 increased more than 15 times (from 38 ± 0.2 nmol min-1 mg-1 to 670 ± 6 nmol min-1 mg-1), and the corresponding activity of Glu228Ala increased more than 4 times (from 34 ± 1 nmol min-1 mg-1 to 150 ± 6 nmol min-1 mg-1).

Similar patterns of copper incorporation between wild-type SCO6712 and the Ser292Ala mutant provoked an interest to further evaluate the substrate selectivity of Ser292Ala using compounds evaluated with the wild-type. Moreover, the similar increase in copper content of Ser292Ala and wild-type SO6712 prepared from microaerobic conditions allowed more direct assessment of the amino acid substitution on substrate selectivity rather than copper binding. As in reactions containing wild-type SCO6712, sodium ferrocyanide followed by polyaromatic compounds were transformed to highest extents after 20 min in reactions containing Ser292Ala (Fig. 4.1). However, kinetic parameters of Ser292Ala differed from wild-type SCO6712. Specifically, the catalytic efficiency of Ser292Ala was slightly decreased on ABTS and slightly increased on 2,6-DMP (Table 4.1). Notably, the Km of Ser292Ala for 2,6-DMP was an order of magnitude lower than the Km of the wild-type enzyme for 2,6-DMP. Further, while the catalytic efficiency of wild-type SCO6712 was similar on quercetin, morin, and myricetin, accurate kinetic parameters could not be determined for Ser292Ala on quercetin and myricetin since the substrate solubility became limiting before maximum turnover rates were achieved. Thus, the apparent Km of Ser292Ala for quercetin and myricetin would be larger than the Km of wild type SCO6712 with these substrates. Taken together, these results suggest that Ser292 localizes near the ortho position of bound phenolic substrates. In this case, substitution of Ser292 to smaller amino acids is likely to promote binding to compounds with very bulky functional groups at the ortho position, such as in the case of 2,6-DMP. On the other hand, a comparatively smaller hydroxyl group at the ortho position of substrates may form favourable interactions (such as hydrogen bonding) with the hydroxyl group of the Ser292 side chain. Substitution of Ser292 to alanine may thus lead to weaker binding to substrates such as quercetin and myricetin. Examination of the position of 2,6-dimethoxyphenol docked in silico to the binding pocket of SCO6712 enzyme further supports the idea that Ser292 is located near the ortho position of phenolic substrates (Fig. 4.3). Based on the kinetic results, binding of SCO6712 to morin, which has a hydroxyl at the meta position, does not seem to be affected by mutating Ser292 to alanine, possibly because the meta hydroxyl of morin forms binding interactions with other residues of the enzyme.

T1 Cu

H293 Y230

S292 2.1Å 2,6-DMP

T1 Cu

H293 Y230

A292 4.3Å 2,6-DMP

Figure 4.3. Ribbon image of 2,6-dimethoxyphenol (2,6-DMP) docked in silico to binding pocket of SCO6712 A) wild type enzyme and B) Ser292Ala mutant. Docking was performed using AutoDock.4 software (Morris et al., 1998). The three polypeptide chains of the SCO6712 trimer are shown in different colours. The 2,6-DMP docked to wild type and Ser292Ala mutant is shown in cyan and yellow, respectively. Type 1 copper is shown as a brown sphere. Dotted lines between 2,6-DMP and H293 or Y230 show polar contacts. Distance from methyl group of 2,6-DMP to either Ser292 or Ala292 is also shown. Protein structure image was generated using PyMOL.

The predicted interactions of Ser292 with ortho substituents may be an indication that the natural substrate for this enzyme contains ortho dihydroxy substitution but not ortho methoxyl substitution. In the context of lignin degradation, this may mean that SCO6712 enzyme would not act directly on lignin since the phenolic monomeric building blocks of lignin do not contain ortho hydroxyl and contain either ortho methoxyl or no substitution at the ortho position (Lange et al., 2013). However, it has recently been suggested that bacterial laccases may have a role in lignin degradation (Bugg et al., 2010). In fact, a very recent study used a Streptomyces coelicolor knockout mutant lacking SCO6712 laccase to identify a possible role for the enzyme in lignin degradation (Majumder et al., 2014). In the same work, the authors determined that even though the purified SCO6712 was able to oxidize a dimeric lignin model compound containing an ortho methoxy group, the activity was low, leading the authors to suggest that the natural substrate must be different in shape or chemical nature. It is possible then, that the natural substrate of the laccase is a lignin degradation product with ortho dihydroxy substitution (Bugg et al., 2010) (possibly derived from action of other oxidase enzymes such as peroxidases), which acts as mediator of laccase during later stages of lignin degradation.

4.4. Conclusions

A bacterial laccase known as SLAC (SCO6712) showed activity on a range of compounds (caffeic acid, ferulic acid, resveratrol, quercetin, morin, kaempferol and myricetin, and others) that include phenolics from different classes (hydroxybenzoic acids, hydroxycinnamic acids, stilbenes, and flavonoids), suggesting that SCO6712 could be applied in organic syntheses involving bioactive phytochemicals. Mutation studies indicate that saturating mutagenesis of Ser292, as well as Tyr230, is a feasible approach to extending the catalytic efficiency of SCO6712 on bioactive compounds. In particular Ser292 may be involved in favourable interactions with the ortho hydroxyl groups of phenolic substrates, while it hinders interactions with phenolics that contain an ortho methoxyl group.

CHAPTER 5. CHARACTERIZATION OF PRODUCT FORMATION FROM

ENZYMATICALLY OXIDIZED PLANT PHENOLICS AND ASSAY OF

ANTIOXIDANT ACTIVITY

Parts of this chapter are in preparation for submission:

Sherif, M., Qazi, S., Abou-Zaid, M., and Master, E. R. Characterization of product formation from enzymatically oxidized plant phenolics and assay of antioxidant activity.

M. Sherif performed all antioxidant assays and HPLC-MS/MS analyses, with support through helpful discussions with S. Qazi and M. Abou-Zaid.

5.1. Introduction

Based on the lack of synthetic product formation by the arylesterases, it was decided to investigate the products formed from laccase action on the flavonols. The hydroxyl groups of phenolics can be oxidized by oxidase enzymes to form phenoxy radicals that can then react with each other to form oligomers. For example, action of oxidoreductase enzymes is thought to play a role in the formation of the lignin polymer from the constituent phenolic monolignols (Boerjan et al., 2003). Also, laccase-like enzymes have been shown to oxidatively polymerize flavonoids to cause seed coat browning in Arabidopsis, with possible protective functions of the polymerized flavonoids against biotic and abiotic stress (Pourcel et al., 2005). From a biotechnological perspective, multicopper oxidases hold promise for the modification of flavonoids to produce larger molecular weight oligomers that partition more effectively into lipids (Lotito et al., 2000) and thus impart better antioxidant activity towards emulsified lipids (Shahidi & Zhong, 2011). Few reports demonstrate the use of multicopper oxidase on flavonols and even less report the characterization of the reaction products. One exception includes laccase modification of rutin and analysis of resulting dimers and trimers (Uzan et al., 2011). Phenolic compounds such as the flavonoids have antioxidant properties. Antioxidant activity can arise from radical scavenging via hydrogen atom transfer or electron transfer (Wright et al., 2001). Antioxidant activity may also involve metal ion chelation (Perron et al., 2008) preventing such ions from initiating oxidation reactions. In the radical scavenging mechanism, the phenolic antioxidant is sacrificially oxidized, itself becoming a relatively stable radical. Some general relationships relating molecular structure to antioxidant activity have been observed for different classes of phenolic compounds (reviewed in Balasundram et al., 2006). In general, the number and position of hydroxyl groups along with the location of double bonds turn out to be key determinants affecting antioxidant activity by their stabilizing effects on the resultant phenoxy radical that is produced after sacrificial oxidation. Herein, I describe using HPLC-MS to identify dimers of quercetin, morin, and myricetin as one of the main observable products of laccase catalyzed oxidation of these flavonol substrates. Additionally, I have examined changes in antioxidant activity of the whole enzyme reaction mixture over enzymatic reaction time, as a way of assessing overall changes in antioxidant activity as a result of laccase treatment.

5.2. Materials and methods

5.2.1. HPLC-MS analysis of flavonol products after SCO6712 laccase treatment

Phenolic substrates quercetin, morin, kaempferol, and myricetin (1 mM) were individually reacted with SCO6712 enzyme for 20 min as mentioned in Chapter 4. Reactions were filtered through 0.2 µm syringe filters with GHP (hydrophilic polypropylene) membranes (Pall). Separation of substrates and products was carried out with an Acquity (Waters) UPLC BEH C18 (2.1 mm x 50 mm; 1.7 µm) reversed-phase column. Total run time was 20 min at 0.15 mL/min flow rate and 10 µL of sample was injected into the column. Eluents were 0.1 % formic acid (A) and 100 % acetonitrile (B). Gradient elution was as follows: 95 % to 85 % A (0 to 1 min), 85 % to 80 % A (1 to 8 min), 80 % to 50 % A (8 to 10 min), 50 % to 20 % A (10 to 12 min), 20 % to 95 % A (12 to 13 min), and 95 % A (13 to 20) min. Mass spectrometry was run in positive ion mode on an Exactive (Thermo Scientific) mass spectrometer using an electrospray ionization source and an Orbitrap mass analyzer. Spectra were analyzed using Xcalibur software (Thermo Scientific) and also using small molecule analysis mode of SIEVE software (Thermo Scientific).

5.2.2. HPLC-MS analysis of flavonol dimer presence over laccase reaction time

Quercetin, morin, and myricetin (1 mM) were individually reacted with SCO6712 enzyme for 5 min, 3 h, and 24 h using same reaction conditions as mentioned in Chapter 4. Reactions were filtered through 0.2 µm syringe filters with GHP (hydrophilic polypropylene) membranes (Pall) and HPLC-MS was carried out as mentioned in the preceding paragraph (Section 5.2.1).

5.2.3. Total radical-trapping antioxidant parameter (TRAP) assay of whole laccase

reaction mixture

TRAP assay was run using a procedure modified from that previously reported (Valkonen & Kuusi, 1997). The TRAP antioxidant assay was carried out in 50 mM pH 7.4 sodium phosphate buffer with 50 mM 2,2′-azobis(2-methylpropionamidine) (AAPH) and 40 µM 2′,7′-dichlorofluorescin (DCFH) at 37 oC. DCFH was first generated from 2′,7′-

86 dichlorofluorescin diacetate (DCFH-DA) by mixing 324 µL of 1 mM stock DCFH-DA in DMSO with 81 µL of 50 mM sodium hydroxide and leaving to sit for 30 min at room temperature to hydrolyze. This DCFH solution was then diluted by adding 675 µL of MilliQ water to make 300 µM DCFH. In a microtitre plate 37.5 µL of 200 mM sodium phosphate buffer was mixed with 73.5 µL MilliQ water. To this was added 20 µL of 300 µM DCFH and 4 µL of SCO6712 laccase reaction sample (5 min, 20 min, 40 min, 1 h, 3 h, 5 h, and 24 h reactions carried out as mentioned in Chapter 4 and remaining phenolic monomer substrate measured spectrophotometrically as mentioned in Chapter 4). Before the antioxidant assay was started by addition of 15 µL AAPH (500 mM stock), the microtitre plate was pre-incubated at 37 oC for 20 min. Absorbance of oxidized DCFH was followed at 504 nm every 1 min for 120 min at 37 oC.

5.2.4. Ferric reducing antioxidant power (FRAP) assay of whole laccase reaction

mixture

The FRAP assay was run following the method of Benzie and Strain (1996). Working FRAP reagent was prepared by mixing 300 mM sodium acetate buffer (pH 3.6), 10 mM 2,4,6- tris(2-pyridyl)-s-triazine (TPTZ) in 40 mM HCl, and 20 mM iron (III) chloride hexahydrate

(FeCl3·6H2O) in MilliQ water in a 10:1:1 ratio. This working FRAP reagent was incubated at 37 oC. Ascorbic acid standards were prepared at 0 mM, 0.1 mM, 0.2 mM, 0.4 mM, 0.6 mM, 0.8 mM, and 1 mM concentrations in MilliQ water. Standard (10 µL ascorbic acid) or laccase reaction sample (10 µL of a five times dilution of 5 min, 20 min, 40 min, 1 h, 3 h, 5 h, and 24 h SCO6712 laccase reactions carried out as mentioned in Chapter 4 and remaining phenolic monomer substrate measured spectrophotometrically as mentioned in Chapter 4) was pipetted into a microtitre plate and 300 µL working FRAP reagent was added. Absorbance of reduced ferric-tripyridyltriazine (FeIII-TPTZ) complex was followed at 593 nm every 2 min for 10 min at 37 oC.

5.2.5. HPLC-MS/MS analysis of quercetin dimer and myricetin dimer

HPLC-MS was carried out as described above (Section 5.2.1). Tandem mass spectrometry was carried out using nitrogen collision gas at 35 eV collision energy. A quadrupole mass filter was used to isolate ions of a specific m/z before fragmentation. Spectra

87 were analyzed using Xcalibur software (Thermo Scientific). Going through the list of m/z for the fragment spectra of quercetin dimer and myricetin dimer, the Formula Finder feature of the program Molecular Weight Calculator Version 6.49 (http://omics.pnl.gov/software/molecular- weight-calculator) was used to aid in correlating m/z value to possible combinations of hydrogen, carbon, and oxygen. The following atomic weights were used in these calculations: proton = 1.00727649 amu, 1H hydrogen atom = 1.0078246 amu, 12C carbon atom = 12 amu, 16O oxygen atom = 15.994915 amu, and 13C carbon atom = 13.0033548 amu. Masses for proton, 1H, 12C, and 16O are the default monoisotopic mass values in the Molecular Weight Calculator software and are sourced from the IUPAC standard atomic weights and uncertainties (Coplen 1996). Mass for 13C was from De Laeter et al. (2003).

5.3. Results and discussion

5.3.1. HPLC-MS analysis of flavonol products after SCO6712 laccase treatment

Previous studies in which phenolic compounds are treated with oxidase enzymes demonstrate that often-formed products are dimers and higher level oligomers of the original phenolic. For example, Uzan et al. (2011) formed dimers and trimers of rutin after treatment of rutin with laccases from Pycnoporus fungi. Oligomeric flavonoids may have application as better preservatives of emulsified lipids against oxidative decay (Lotito et al., 2000). In this study, dimerized forms of quercetin, morin, and myricetin were identified as one of the major products after twenty minutes of reaction with SCO6712 enzyme; with positive ion mode m/z values of 603, 575, and 633 respectively (Table 5.1; Fig. 5.1). No dimer of kaempferol was identified after 20 min reaction. Instead, the major product for reaction with kaempferol had an m/z value of 271 indicating a breakdown product of the starting monomer (m/z 287) (Supplemental Table A3.1; Fig. 5.2). The quercetin dimer of m/z 603 and the myricetin dimer of m/z 633 had longer retention times than the corresponding monomers on reversed-phase HPLC (Fig. 5.1), suggesting these dimer products will have increased hydrophobicity compared to the starting monomer. Attempts were made to quantify the solubility of the starting monomer to eventually compare to purified dimer but challenges were encountered along the way (see Appendix 5 for details).

Table 5.1. HPLC-MS data for quercetin, morin, myricetin, and their dimers produced after enzymatic reaction. Mass spectra collected in positive ion mode. Compound m/z Retention Lost atoms from the monomers to form time (min) dimers (based on observed m/z value)a Quercetin 303.049 12.15 -- Quercetin dimer #1 603.076 12.59 2 H’s Quercetin dimer #2 603.076 12.90 2 H’s Morin 303.050 11.78 -- Morin dimer 575.082 11.54 2 H’s, 1 C, and 1 O Myricetin 319.045 11.32 -- Myricetin dimer #1 633.051 11.50 4 H’s Myricetin dimer #2 633.051 11.71 4 H’s Myricetin dimer #3 635.067 10.21 2 H’s a. H=hydrogen, C=carbon, O=oxygen

RT: 0.00 - 20.01 SM: 7B A) 100 S NL: D1 4.59E6 90 m/z= 303.0300- 80 303.0600+ 603.0600- 603.0900 MS 70 Qrc_May15_rxn 1 60 50

Relative abundance Relative

Relative Abundance Relative 30

20 10 0 0 2 4 6 8 10 12 14 16 18 20 Time (min) Qrc_May15_rxn1 #3697 RT: 12.64 AV: 1 NL: 4.26E6 T: FTMS {1,1} + p ESI Full ms [200.00-1000.00] 603.0771 100 90 80

60 585.0670 50 40

Relative abundance Relative

Relative Abundance Relative 30

20 219.1018 650.3754 10 301.0346 567.2784 694.4016 479.2258 782.4539 864.4987 928.5542 0 200 300 400 500 600 700 800 900 1000 m/z

B) RT: 0.00 - 20.01 SM: 7B 100 NL: S 1.28E7 90 m/z= 303.0300- 80 303.0600+ 575.0700-

575.0900 MS 70 Mrn_May15_rxn 1 60

50 40

Relative abundance Relative

Relative Abundance Relative 30

20 D

0 0 2 4 6 8 10 12 14 16 18 20 Time (min)

Mrn_May15_rxn1 #3371 RT: 11.54 AV: 1 NL: 2.27E6 T: FTMS {1,1} + p ESI Full ms [200.00-1000.00] 575.0831 100

80 70 60

40 219.1749

Relative Abundance Relative abundance Relative 30 20 273.0399 407.0406 393.0613 10 425.0511 285.0764 557.0730 629.0029 711.0561 790.4880 901.0346 983.0999 0 200 300 400 500 600 700 800 900 1000 m/z C) RT: 0.00 - 20.01 SM: 7B 100 NL: D2 3.94E6 90 m/z= 319.0300- 80 319.0600+ 633.0400-

633.0700 MS 70 Mrctn_May15_rx n1 60

Relative abundance Relative Relative Abundance Relative D1 30 S 20 10

0 0 2 4 6 8 10 12 14 16 18 20 Time (min) Mrctn_May15_rxn1 #3405 RT: 11.65 AV: 1 NL: 3.77E6 T: FTMS {1,1} + p ESI Full ms [200.00-1000.00] 633.0523 100

70 60

Relative abundance Relative

Relative Abundance Relative 30 20 505.2628 655.0339 10 219.1748 461.2367 605.0575 683.0532 799.0439 319.0455 919.0682 981.0656 0 200 300 400 500 600 700 800 900 1000 m/z

Figure 5.1. HPLC-MS chromatograms and m/z spectra for 20 min reaction samples with laccase enzyme for A) quercetin, B) morin, and C) myricetin. S indicates peak for starting phenolic substrate and D indicates peak for dimer product. The number after D indicates a particular dimer (refer Table 5.1).

RT: 0.00 - 20.01 SM: 7B 100 NL: P 6.71E6 90 m/z= 271.0500- 80 271.0700+ 287.0400-

287.0600 MS 70 kmpf_May15_rxn 1 60

Relative abundance Relative

Relative Abundance Relative 30

20 10 0 0 2 4 6 8 10 12 14 16 18 20 Time (min)

kmpf_May15_rxn1 #3698 RT: 12.64 AV: 1 NL: 6.48E6 T: FTMS {1,1} + p ESI Full ms [200.00-1000.00] 271.0602 100

Relative abundance Relative

Relative Abundance Relative 30 677.0933

20 437.0508 10 371.2281 523.2526 717.0863 295.1907 628.3915 804.4969 904.5146 961.1278 0 200 300 400 500 600 700 800 900 1000 m/z Figure 5.2. HPLC-MS chromatogram and m/z spectrum for 20 min reaction sample with laccase enzyme for kaempferol. P indicates product peak. Peak for substrate almost completely overlaps with that for product and is too small to be visible.

Based on the m/z value of 603, quercetin dimer would have been formed by addition of two monomers with accompanying loss of two hydrogens, and this is not surprising since the formation of the new dimer bond would involve the loss of at least one bond originally present in each monomer to maintain four bonds for each carbon and two bonds for each oxygen. Based on the m/z of 575, morin dimer may have been formed by addition of two monomers with accompanying loss of two hydrogens and one CO, while myricetin dimer of m/z 633 would have been formed by addition of two monomers with accompanying loss of four hydrogens. The production of morin dimer suggests that the ortho dihydroxy substitution on the B ring of flavonols is not a necessity for oxidative dimer formation, since morin has a meta dihydroxy on its B ring. However, oxidation of kaempferol (with only a single hydroxyl on its B ring) did not lead to kaempferol dimers, suggesting that at least a dihydroxy substitution on the B ring is required for flavonol dimer formation. This may mean that oxidatively forming a stable dimer of flavonols requires formation of a dioxane linkage, which is the dimer linkage reported for most previously purified quercetin dimers (Hirose et al., 1999; Krishnamachari et al., 2004; Tram et al., 2005; Veverka et al., 2013). The presence of dimers of quercetin, morin and myricetin was assessed for varying reaction time (5 min, 3 h, and 24 h) to determine if the dimer accumulates to a maximum steady level or if it depletes due to undergoing further reactions. In addition to giving some insight into the reactivity of the dimer, this information will be helpful in determining a suitable reaction time for product purification. The dimer was found to be detectible after 5 min reaction time (Fig. 5.3). In the case of myricetin, a dimer with m/z 635 was also identified and would have been formed by addition of two monomers with accompanying loss of two hydrogens (Table 5.1; Fig. 5.4). Looking at the HPLC-MS chromatograms for 5 min reaction samples of quercetin and myricetin with laccase SCO6712 enzyme, it appears there were different isomeric forms of the quercetin (m/z 603) and myricetin (m/z 633) dimers eluting at slightly different times (Fig. 5.3).

A) RT: 0.00 - 20.01 SM: 7B 100 D1 NL: 3.14E6 90 m/z= 603.0600- 80 603.0900 MS Qrc_5min_r 70 D2 xn1 60 50

Relative Abundance Relative

Relative abundance Relative 30

20 10 0 0 2 4 6 8 10 12 14 16 18 20 Time (min)

RT: 0.00 - 20.01 SM: 7B B) 100 NL: D 1.55E6 90 m/z= 575.0700- 80 575.0900 MS Mrn_5min_rx 70 n1 60 50 40

Relative abundance Relative

Relative Abundance Relative 30

0 0 2 4 6 8 10 12 14 16 18 20 Time (min) RT: 0.00 - 20.01 SM: 7B C) 100 NL: D2 2.49E6 90 m/z= 633.0400- 80 633.0700+ D1 635.0630- 635.0700 MS 70 Mrctn_5min_rxn 1 60

50 D3 40

Relative abundance Relative

Relative Abundance Relative 30 20 10

0 0 2 4 6 8 10 12 14 16 18 20 Time (min)

Figure 5.3. HPLC-MS chromatograms for 5 min reaction sample with laccase enzyme for (A) quercetin, (B) morin, and (C) myricetin. For clarity, chromatograms are showing peaks only for the dimer products and omit substrate peaks. The number after D indicates a particular dimer (refer to Table 5.1).

Mrctn_5min_rxn1 #2985 RT: 10.21 AV: 1 NL: 8.38E5 T: FTMS {1,1} + p ESI Full ms [200.00-1000.00] 635.0672 100 651.0617 90 80

50 439.0300 578.3752 40

Relative abundance Relative

Relative Abundance Relative 30 668.0891 20 371.1017 456.0567 217.1075 696.4382 894.0798 10 561.3492 279.1594 813.0582 949.0750 0 200 300 400 500 600 700 800 900 1000 m/z Figure 5.4. Mass spectrum for myricetin dimer of m/z 635 (from peak D3 in chromatogram of Fig. 5.3. C); refer to Table 5.1) for 5 min reaction sample with laccase enzyme.

The dimers produced after 5 min in the enzymatic reactions subsequently become depleted by 3 h reaction time as indicated by HPLC-MS peak intensities (Fig. 5.5). Part of this disappearance may be due to formation of higher degree oligomers in the reaction as the dimer reacts with unreacted monomers. For quercetin, morin, and myricetin, after 5 min reaction with enzyme, the HPLC-MS peak intensities for substrate monomer were 68 % ± 7 %, 86 % ± 4 %, and 46 % ± 5 %, respectively, of initial substrate. By 3 h reaction with enzyme, these values had fallen to 0.8 % ± 0.6 %, 0.9 % ± 0.2 %, and 0.03 % ± 0.01 %. While the HPLC-MS data for 3 h and 24 h enzymatic reactions do not show peaks that correspond to trimers, there are peaks with m/z values that correspond to species having a molecular weight that is intermediate between trimer and dimer and also intermediate between dimer and monomer (Table 5.2). Furthermore, species with m/z values that correspond to a molecular weight less than the monomer could be detected (Table 5.2). Taken together, these results suggest that dimers and trimers are produced but subsequently undergo cleavage reactions. Among the top 10 detected products at late time points (3 h and 24 h) in enzymatic reactions, most have a molecular weight intermediate between that of the monomer and dimer (Table 5.3). The presence of these products at late time points suggests that such molecular weight products are more stable products than the dimers produced at early time points. This has implications in limiting the extent of polymerization that can be achieved with these flavonols under these reaction conditions.

A) 4.0

3.5

3.0

2.5

2.0

1.5

1.0

(millions of counts) 0.5

of chromatogram Height peak 0.0 -enz +enz -enz +enz -enz +enz

initial 5 min 3 h 24 h

B) 1.8

1.6

1.4

1.2

1.0

0.8

0.6

(millions of counts) 0.4

Height of chromatogram of chromatogram Height peak 0.2

0.0 -enz +enz -enz +enz -enz +enz initial 5 min 3 h 24 h

C) 3.5

3.0

2.5

2.0

f counts) 1.5

1.0

(millions o 0.5

of chromatogram Height peak 0.0 -enz +enz -enz +enz -enz +enz initial 5 min 3 h 24 h

D) 1.0 0.9 0.8

0.7

0.6

0.5

0.4

0.3

(millions of counts) 0.2 of chromatogram Height peak 0.1 0.0 -enz +enz -enz +enz -enz +enz initial 5 min 3 h 24 h

Figure 5.5. HPLC-MS analysis of flavonol dimers from A) quercetin (dimer #1), B) morin, C) myricetin (dimer #2), and D) myricetin (dimer #3). Samples were from reactions of laccase with substrate after 5 min, 3 h, and 24 h of reaction. Initial represents the time when the substrate is first mixed with buffer (without enzyme) along with approximately 20 min of handling time on ice before sampling for HPLC-MS. -enz and +enz indicate reactions without and with enzyme, respectively. Refer to Table 5.1 for different dimers of the same substrate. n=3; error bars indicate standard deviation.

Table 5.2. HPLC-MS data for representative intermediate molecular weight products present in late time point enzymatic reactions of quercetin, morin, and myricetin. Mass spectra collected in positive ion mode. Reactant Reaction time m/z of Retention Intensity Molecular weight point (h) product time (min) (x 105) rangea Quercetin 3 301.034 12.2 7.43 monomer and dimer and monomer and dimer and monomer and dimer and monomer and dimer and monomer and dimer and monomer and dimer and

100

Table 5.3. Top 10 products from late time point enzymatic reactions of quercetin, morin, and myricetin. Reaction products are ordered in increasing molecular weight. Mass spectra collected in positive ion mode. Reactant Reaction time m/z of Retention Intensity Molecular weight point (h) product time (min) (x 105)a range Quercetin 3 301.034 12.20 7.43 monomer and monomer and dimer and monomer and monomer and

101

Table 5.3. cont... Reactant Reaction time m/z of Retention Intensity Molecular weight point (h) product time (min) (x 105)a range Morin 24 453.046 6.04 1.52 >monomer and monomer and dimer and monomer and dimer and

A wide range of oxidative degradation products have been previously observed for quercetin and related compounds using chemical, electrochemical, enzymatic, and autoxidation systems (Krishnamachari et al., 2004; Zenkevich et al., 2007; Zhou & Sadik, 2008). It is proposed that degradation processes are initiated by addition of molecular oxygen or water to the central C ring of the flavonol (Krishnamachari et al., 2004), which can be expected to increase the polarity of the resulting compounds. In this work, most of the enzymatic products that could correspond to a cleavage reaction of the dimer have a retention time earlier than that of the dimer in reversed phase HPLC (Tables 5.1 and 5.3), suggesting they are indeed more polar. Following from the reaction mechanism suggested by Krishnamachari et al. (2004), the identity of some of the late-formed enzymatic products for quercetin can be putatively assigned

102 to a reaction sequence initiated by addition of oxygen to the C ring of the flavonol monomer radical to produce compounds (2) and then (3) (Fig. 5.6A). The flavonol is then cleaved at its C ring to produce the depside compound (4), which can lose carbon monoxide and carbon dioxide to produce compound (5). Compound (5) can then rearrange to produce compound (6). Compounds (4) and (6) can be oxidized at their ortho dihydroxy moieties to produce compounds (7) and (8), respectively. Compounds (7) and (8) can correspond to species with m/z 333 and 261, respectively, that are produced after 24 h reaction of quercetin with enzyme (Table 5.3). Oxygen in compound (2) may also react with a different carbonyl on quercetin to produce compound (9), which subsequently breaks down to produce compound (10) with loss of carbon monoxide (Fig 5.6B). Compound 10 can then be oxidized at its ortho dihydroxy position to produce compound (11), which can correspond to the species with m/z 305 produced after 24 h reaction of quercetin with enzyme (Table 5.3). The unoxidized forms of compounds (7) and (11) (i.e. compounds (4) and (10), respectively, (Fig 5.6)) have been previously proposed as quercetin oxidation products based on mass spectrometric analysis (Zhou & Sadik, 2008). However, a quercetin oxidation product with molecular weight of compound (8) was not previously reported. This may be due to the fact that previous works have used shorter incubation times (typically not longer than 3 h) and/or lower incubation temperatures (Schreier & Miller, 1985; Krishnamachari et al., 2004; Zenkevich et al., 2007; Zhou & Sadik, 2008). Longer incubation times and higher temperatures in this work might also be the reason for observing oxidized forms of compounds (4) and (10) (i.e. compounds (7) and (11)).

103

(A) OH OH OH HO HO HO

HO O HO O HO O C O-O O O-O O O O OH O OH O OH O OH (1) (2) (3) OH OH OH HO HO HO

CO, CO2 HO O HO O HO OH O OH O O OH O OH OH O H (4) (5) (6) oxidation oxidation

O O O O

HO O HO OH O (7) O (8) O OH O OH OH

(B) OH OH OH HO HO HO CO HO O HO O HO O O-O O O O O O OH O OH OH OH OH O

(2) (9) (10) O O

HO O O oxidation (11) O OH OH

104

Figure 5.6. Proposed reaction scheme for production of some of the quercetin oxidation degradation products. Compound (1) is a radical formed after initial oxidation of quercetin. Compounds (7), (8), and (11) can correspond to products with m/z 333, 261, and 305, respectively, identified from 24 h reactions of quercetin with enzyme (Table 5.3).

105

While the non-enzymatic reaction does eventually achieve complete reaction of quercetin by 24 h, comparison of the product distribution for reactions without and with laccase shows that the enzyme not only accelerates the reaction but also leads to a slightly different product profile (Table 5.4). The enzymatic reaction is known to produce water as the byproduct of oxidation while non-enzymatic autoxidation has been shown to produce hydrogen peroxide as the reduced product from oxygen (Canada et al., 1990). This hydrogen peroxide may play a partial role in the different product profiles. For example, some of the hydrogen peroxide may decompose to produce hydroxyl radicals that then react with quercetin radicals, leading to breakdown of quercetin and its oligomers. However, it cannot be ruled out that the different product profile may simply be a consequence of the fact that the initial products in the reaction without enzyme were formed at a later time than with enzyme so the non-enzymatic initial products have not had as much time to undergo further reactions. In this way, even though the starting material is completely depleted after 24 h in both reactions without and with enzyme, corresponding reactions may not yet be terminated.

Table 5.4. Top 10 products from 24 h reaction of quercetin without and with laccase enzyme. Reaction products are ordered in increasing m/z value. Reaction without enzyme Reaction with enzyme m/z Retention Intensity m/z Retention Intensity time (x 105) time (x 105) (min) (min) 245.045 7.35 1.91 261.040 11.19 1.53 289.034 3.57 2.38 301.034 12.20 4.83 301.034 12.20 6.36 305.029 11.64 1.51 307.045 7.36 1.60 333.024 5.44 1.23 409.055 9.82 2.83 453.045 11.81 3.33 453.045 11.81 5.98 453.045 10.56 2.96 453.045 10.56 4.76 481.039 8.47 1.90 481.039 8.47 2.07 481.040 9.74 1.32 481.040 9.74 1.85 547.051 10.78 1.59 483.055 7.44 2.43 619.035 9.61 1.28

It is possible that the dimer can be acted on directly by the enzyme to produce dimer radicals that undergo coupling to form trimers, which subsequently break down. Additionally,

106 the dimer may be undergoing non-enzymatic chemical oxidation (as was seen for the monomers, discussed above), to form the reactive radicals that undergo coupling. The dimers were also produced without the presence of enzyme, albeit after longer (3 h) incubation times for quercetin and morin to give 84 % ± 14 % and 5.5 % ± 0.6 %, respectively, of peak intensity of dimer from 5 min enzymatic reaction (Fig. 5.5); and in a similar manner to enzymatically produced dimers, the non-enzymatically produced dimers become depleted by 24 h reaction time. Notably, the 24 h morin reactions without enzyme showed only 2.3 % ± 0.2 % of HPLC- MS peak intensity of dimer from 5 min enzymatic reaction, even though the HPLC-MS peak intensity for remaining amount of morin substrate was comparable at 86 % ± 4 % of initial and 82 % ± 4 % of initial for 5 min enzymatic and 24 h non-enzymatic reactions, respectively. This may mean that the dimer can accumulate to greater maximum extent in the presence of enzyme. An initially formed morin radical may potentially react by dimerization or by breakdown reactions. Dimerization reactions might be more favoured, and hence compete with, breakdown reactions if the concentration of initially formed radicals is greater at any particular point in time. The enzyme is able to cause the accumulation of more morin radicals (before the radicals undergo breakdown reactions) and this may be the reason for the greater amount of morin dimer produced by enzyme than without enzyme. Chemical autoxidation of flavonols has been widely observed previously, particularly at more basic pH conditions (Canada et al., 1990; Dangles et al., 1997; Zenkevich et al., 2007; Zhou et al., 2008). While it may at first sight be thought that the initial reaction is a simple and direct transfer of electrons from the flavonol to molecular oxygen, it has been suggested that thermodynamic and electrochemical considerations (redox potential of quercetin is 0.33 V and redox potential of dioxygen is -0.16 V at pH 7) preclude this from occurring (Hajji, et al., 2006). Instead, it is proposed that trace levels of metal ions facilitate the process (Hajji, et al., 2006). The benefit of reduced reaction time and increased dimer product quantity would need to be weighed against the cost of enzyme production to see if the enzyme has any advantage from an economic perspective. In this respect, bacterial laccases may offer advantages in less costly recombinant production of large enzyme quantities. One potential added benefit of using the enzyme (rather than simply incubating in the reaction buffer without enzyme) is that it may be possible to mix two different phenolics (one of which is not easily autoxidized, such as morin) to form heterodimeric species; whereas such species would be almost non-existent in a purely

107 autoxidation system since the less easily autoxidized phenolic would not accumulate enough radicals to compete with homodimerization of the more easily autoxidized substrate. While it is not obvious if heterodimeric flavonoids would have application as more effective antioxidants, they may have bioactivity beyond antioxidant activity, as mentioned in Section 5.3.2. Morin is an isomer of quercetin with the only difference that morin has an m-dihydroxy substitution in the B ring compared to the o-dihydroxy substitution of quercetin (Fig. 5.7). Formation of different molecular weight dimer products between morin and quercetin might therefore be attributed to the fact that the dimer formation involves the B ring of one or both of these compounds, which in turn might be a consequence of the fact that the site of action of the SCO6712 enzyme is the B ring hydroxyls of these compounds. While there are no previous examples, to the best of my knowledge, of quercetin dimers produced after laccase treatment, previous research groups have used mushroom polyphenol oxidase (tyrosinase) and horseradish peroxidase to oxidize quercetin (Makris & Rossiter, 2002). One of the major products that eluted later than quercetin in reversed-phase HPLC was proposed to be a dimer formed between the B-ring hydroxyls and the C2-C3 carbons (Gulsen et al., 2007); the same compound as isolated from onion skins (Tram et al., 2005) (Fig. 5.7). Makris & Rossiter (2002) also carried out oxidation of morin but did not determine molecular weight of the major product. However, the product eluted earlier than morin in reversed-phase HPLC, as was seen in this study, leaving open the possibility that it is the same compound as in this study.

108

OH HO OH OH HO O HO O

OH OH OH O OH O

quercetin morin OH HO

HO O O OH O O OH OH O O OH

quercetin dimer OH Figure 5.7. Structures of the isomers quercetin and morin. Structure of quercetin dimer previously isolated from onion skins.

Presence of an appreciable amount of myricetin dimer having m/z 635 even in the 5 min reaction solution in the absence of enzyme (Fig. 5.5) suggests that it is rapidly autoxidized, which is consistent with the fact that the tri-hydroxy group on the B ring facilitates oxidation by virtue of the electron donating properties of phenolic hydroxyls increasing electron density of neighbouring phenolic hydroxyls (Xu, 1996; Balasundram et al., 2006). The myricetin dimer of m/z 633 could represent a further oxidation of the dimer having m/z 635 to result in the formation of a quinone with loss of two hydrogens. While there are no previous reports of myricetin dimers, Meotti et al. (2008) used Human leukocyte myeloperoxidase to oxidize myricitrin (a glycoside conjugate of myricetin). They observed the formation of a compound with m/z 925 (in negative ion mode) that could correspond to a dimer of myricitrin. They also observed another product with an m/z 923 (in negative ion mode) that is less than the first dimer by two m/z units and was proposed to be a further oxidation product of the dimer having m/z 925. The authors carried out an additional experiment, which contained glutathione in the

109 myeloperoxidase-myricitrin reaction mix. In this case, an additional product was formed with an m/z value representing a conjugate of myricitrin and glutathione (m/z 768). The authors proposed that after initial abstraction of an electron from the myricitrin monomer, a quinone is formed as an intermediate before dimer formation. This quinone can then react with another myricitrin or with glutathione to form a dimer or a myricitrin-glutathione conjugate, respectively. It is possible that the dimer can undergo another round of oxidation to form a quinone of the dimer.

5.3.2. Antioxidant assay of whole laccase reaction mixture

In the course of producing new products with varied solubilities, it is conceivable that the inherent radical scavenging capacity as tested in single-phase aqueous solutions is reduced. For example, the phenolic hydroxyl of one monomer may react with a carbon atom of a second monomer to form a dimer linked by carbon-oxygen bond, and the conversion of phenolic hydroxyls to ethers has been correlated to reduced antioxidant activity (Van Acker et al., 1996). In this case, one of the monomers making up the dimer could have reduced activity so that the dimer has reduced antioxidant activity per monomer unit. Even if the dimer (and higher level oligomers) on its own has equal to or better inherent radical scavenging capacity than the monomer, the dimer (and higher level oligomers) might not make up for the lack of inherent radical scavenging capacity of some of the other products that are produced during laccase treatment, such as quinones. In other words, the starting phenolic is being put through a process that transforms it into a mix of products in which some of these products might be better antioxidants but some might be worse. Accordingly, to examine the effect of enzyme treatment on the total inherent radical scavenging capacity, I have carried out antioxidant assays with the whole laccase reaction mixtures. In this way, the net effect of enzyme treatment on inherent radical scavenging capacity can be tested. In the total radical-trapping antioxidant parameter (TRAP) antioxidant assay, 2,2′- azobis(2-methylpropionamidine) (AAPH) is a temperature-sensitive radical generator that produces peroxyl radicals in situ which can oxidize a target probe compound (Prior et al., 2005). Among the probes used is 2′,7′-dichlorofluorescin (DCFH), which when oxidized displays an increase in absorbance at 504 nm (Valkonen & Kuusi, 1997). An antioxidant will compete with DCFH for radicals so that the oxidation of DCFH is inhibited until the antioxidant is spent, at

110 which point DCFH oxidation proceeds at rapid rate. The lag time until DCFH oxidation begins can be used as a measure of antioxidant activity. The peroxyl radical produced by AAPH is thought to be quenched by the antioxidant donating a hydrogen atom (Prior et al., 2005). The hydrogen atom donation (HAT) mechanism of scavenging the oxygen-based radicals in the TRAP assay is representative of the scavenging of reactive oxygen species (ROS) present in biological systems (Prior et al., 2005). Therefore, the TRAP assay was used for this study as it has relevance for assessing nutraceutical potential. Notably, this assay has been used for measuring the antioxidant potential of human blood serum (Valkonen & Kuusi, 1997). Researchers have also used AAPH (without DCFH) in an erythrocyte suspension and quantified antioxidant activity based on haemolysis (Chalas et al., 2001). The TRAP results with the reaction mixtures for quercetin, morin, and myricetin showed a decrease in lag time as the starting flavonoid was consumed (Fig. 5.8). All the different products working together as a whole were not able to make up for loss in antioxidant activity due to the loss of starting phenolic. This acknowledges that there are at least some products being produced that have reduced antioxidant activity (in terms of the induced lag time) compared to the starting substrate. At the same time, comparing myricetin reaction without enzyme for 3 h and 24 h shows that even though the remaining amount of starting material (determined spectrophotometrically) is the same, the lag time is less for the 24 h reaction. This indicates the variability of the composition of the product mixture between the two time points and suggests the need to stop the enzymatic reaction before the depletion of the more potent antioxidant takes place, even though there is no further change in the amount of the starting flavonoid material. Notably, the amount of phenolic monomer remaining after reaction with laccase was quantified using spectrophotometry. It was later found out that these values are overestimates, when compared to quantifying the phenolic monomer in laccase reaction samples (5 min, 3 h, and 24 h reactions) using the intensity of the HPLC-MS peak of the monomer (Supplemental Fig. A3.1.). This overestimation can be due to the absorbance of some of the reaction products, at the same wavelength as the substrate. The overestimation is not very pronounced when there is over 50 % substrate remaining (for example, after 5 min reaction with enzyme there is 79 % ± 1 % or 68 % ± 7 % remaining quercetin based on absorbance or HPLC- MS data, respectively (Supplemental Fig. A3.1.)). The overestimation is most pronounced for the 24 h reactions, since absorbance readings suggest approximately 10 % to 15 % remaining

111 phenolic monomer while the HPLC-MS data indicates essentially no detectable levels of remaining phenolic monomer. Despite this overestimation, the observation can still be made that 3 h myricetin reaction samples without enzyme have no significant remaining phenolic monomer material and neither do 24 h reaction samples without enzyme, yet the lag time TRAP antioxidant activity from the 24 h reaction samples is less, indicating the formation and subsequent depletion of new antioxidant compounds.

112

A) 120 Lag time Remaining starting phenolic substrate

100

40 ) 20 0

% of initial %

-enz -enz -enz -enz -enz -enz -enz

(

+enz +enz +enz +enz +enz +enz +enz initial 5min 20min 40min 1h 3h 5h 24h 120

B) 100

remaining amount of and 80

-enz -enz -enz -enz -enz -enz -enz

+enz +enz +enz +enz +enz +enz +enz initial 5min 20min 40min 1h 3h 5h 24h

120

C) 100

on lag time activity based Antioxidant 80 on spectrophotometry based substrate phenolic starting 60

40 20

-enz -enz -enz -enz -enz -enz -enz

+enz +enz +enz +enz +enz +enz +enz initial 5min 20min 40min 1h 3h 5h 24h

113

Figure 5.8. TRAP antioxidant activity of laccase reaction-mixes from A) quercetin, B) morin, and C) myricetin reactions. Antioxidant activity is based on lag time until rapid rates of DCFH oxidation. -enz and +enz indicate reactions without and with enzyme, respectively. n=3; errors indicate standard deviation.

114

Another interesting observation was the difference in trend for antioxidant results based on lag time as opposed to antioxidant results based on end time at which all DCFH is oxidized. In this case, the lag time for the 40 min quercetin reaction with enzyme is 21 % of the lag time of initial quercetin (Fig. 5.8). However, the end point (at which all DCFH is oxidized) is 88 % that of the initial quercetin (Fig. 5.9). This suggests that the antioxidant activity of some of the products from enzyme treatment is manifested as a reduced rate of DCFH oxidation throughout the entire process of DCFH oxidation rather than as only a lag time at the beginning of the antioxidant assay. This phenomenon was also present with morin and myricetin but was most pronounced for quercetin. In the ferric reducing antioxidant power (FRAP) antioxidant assay, the Fe(III)- 2,4,6- tris(2-pyridyl)-s-triazine (TPTZ) complex reacts with antioxidants to reduce Fe(III), giving Fe(II)-TPTZ that has absorbance at 593 nm at acidic pH (Benzie & Strain, 1996). The FRAP assay determines the ability of the antioxidant to quench radicals by an electron transfer mechanism, as opposed to the TRAP antioxidant assay, which measures ability of antioxidants to quench radicals by a hydrogen atom transfer mechanism (Prior et al., 2005). The overall trend of results with the FRAP assay were the same as those of the TRAP assay. As the starting flavonoid was consumed, there was a decrease in antioxidant capacity of the reaction mixture (Fig. 5.10). Regardless of the mechanism of radical quenching (hydrogen atom transfer or electron transfer) the mix of products is not able to compensate for the lost starting flavonol. However, as seen with the end time TRAP activity values, there was not a direct correlation between loss in antioxidant activity and loss of starting phenolic monomer substrate for quercetin and myricetin. For example, 24 h reaction solutions, in which the starting phenolic material had been depleted below detectable levels, still showed FRAP antioxidant activity (Fig. 5.10), which has motivated the isolation of new bioactive products (by collaborators at NRCan; M. Abou-Zaid). Comparing 5 h reactions without enzyme and 20 min reactions with enzyme, which both have essentially half the remaining amount substrate (50 % ± 2 % and 47 % ± 2 %, respectively, measured spectrophotometrically), the residual FRAP antioxidant activity was comparable at 75 % ± 6 % and 71 % ± 1 %, respectively (approximately 1.06 times more activity for the non-enzymatic reaction) (Fig. 5.10). Residual TRAP antioxidant activities based on end time were also comparable for these reaction samples, giving 91 % ± 0 % and 89 % ± 2 % end time TRAP antioxidant activity for 5 h non-enzymatic

115 and 20 min enzymatic reactions, respectively (Fig. 5.9). This comparability breaks down slightly when looking at residual TRAP values based on lag time with values of 38 % ± 4 % and 29 % ± 4 % lag time TRAP antioxidant activity for 5 h non-enzymatic and 20 min enzymatic reactions, respectively (approximately 1.31 times more activity for the non-enzymatic reaction) (Fig. 5.8). A more apparent difference in the ratio of substrate depletion to antioxidant activity can be seen when comparing 1 h reactions without enzyme to 5 min reactions with enzyme, which have similar levels of remaining substrate (values of 85 % ± 1 % and 80 % ± 1 %, respectively), as determined spectrophotometrically. In this case, the lag time TRAP antioxidant values are 77 % ± 7 % and 45 % ± 6 % for 1 h non-enzymatic and 5 min enzymatic reactions, respectively (approximately 1.71 times more activity for the non-enzymatic reaction) (Fig. 5.8). However, this difference in antioxidant activity is not as great when looking at TRAP activity by end time (Fig. 5.9) (102 % ± 1 % and 85 % ± 4 %, respectively; approximately 1.20 times more activity for the non-enzymatic reaction) and FRAP activity (Fig. 5.10) (97 % ± 1 % and 89 % ± 7 %, respectively; approximately 1.09 times more activity for the non-enzymatic reaction). The almost two times difference seen for lag time TRAP values for 1 h quercetin reaction without enzyme to 5 min reaction with enzyme, despite similar levels of remaining substrate, indicates a variability in the product composition even before total substrate depletion has taken place. From the flavonols treated for 24 h with enzyme, the quercetin reaction sample retains the greatest percentage of initial antioxidant capacity (50 % FRAP value of initial quercetin) compared to morin and myricetin (8.6 % and 21 % FRAP values of initial flavonol, respectively). Enzyme-treated quercetin samples might therefore be the best candidates, among the three flavonols, for searching for new antioxidant products that have altered chemistry.

116

120 End time Remaining starting phenolic substrate 100

A) 80

40 ) 20 0

% of initial %

-enz -enz -enz -enz -enz -enz -enz

(

+enz +enz +enz +enz +enz +enz +enz initial 5min 20min 40min 1h 3h 5h 24h

120

100

remaining amount of and

B) 80

-enz -enz -enz -enz -enz -enz -enz

+enz +enz +enz +enz +enz +enz

+enz initial 5min 20min 40min 1h 3h 5h 24h 120 100

Antioxidant activity based on end time on end activity based Antioxidant

starting phenolic substrate based on spectrophotometry based substrate phenolic starting C) 80

-enz -enz -enz -enz -enz -enz -enz

+enz +enz +enz +enz +enz +enz +enz initial 5min 20min 40min 1h 3h 5h 24h

117

Figure 5.9. TRAP antioxidant activity of laccase reaction-mixes from A) quercetin, B) morin, and C) myricetin reactions. Antioxidant activity is based on end time at which all DCFH is oxidized. -enz and +enz indicate reactions without and with enzyme, respectively. n=3; errors indicate standard deviation.

118

FRAP Remaining starting phenolic substrate 120 A) 100 80

)

% of initial %

-enz -enz -enz -enz -enz -enz

-enz

+enz +enz +enz +enz +enz +enz +enz

( initial 5min 20min 40min 1h 3h 5h 24h 120 B) 100

remaining amount of and

spectrophotometry 60

-enz -enz -enz -enz -enz -enz -enz

+enz +enz +enz +enz +enz +enz +enz initial 5min 20min 40min 1h 3h 5h 24h

120 C) 100

Antioxidant activity based on FRAP value activity based Antioxidant

on based substrate phenolic starting 80

-enz -enz -enz -enz -enz -enz -enz

+enz +enz +enz +enz +enz +enz +enz initial 5min 20min 40min 1h 3h 5h 24h

119

Figure 5.10. FRAP antioxidant activity of laccase reaction-mixes from A) quercetin, B) morin, and C) myricetin reactions. Antioxidant activity is based on FRAP value. -enz and +enz indicate reactions without and with enzyme, respectively. n=3; errors indicate standard deviation.

120

Aside from antioxidant activity, the new dimer compounds may have other potential health benefits. For example, naturally occurring apigenin dimers have been shown to specifically inhibit fibrillogenesis and cytotoxicity of amyloid-β peptides, which are involved in Alzheimer’s disease (Thapa et al., 2011). This effect was proposed to arise from specific binding of the biflavonoids to the amyloid-β peptide due to the appropriate size of the biflavonoid, and was contrasted with the non-specific and less potent cytoprotective effect of the apigenin monomer (Thapa et al., 2011). Laccase enzymes, such as SCO6712, may be used as tools to generate novel biflavonoid compounds not known to be present from natural sources. This may generate further insight into the structural features of dimeric flavonoids that facilitate their inhibition of amyloid-β peptide fibrillogenesis.

5.3.3. HPLC-MS/MS analysis of quercetin dimer and myricetin dimer

Previous research groups have identified characteristic fragmentation patterns of the monomeric flavonols including quercetin and myricetin (Fabre et al., 2001; Ma et al., 1997). Among the fragment ions are those that result from breaking two bonds in the C ring resulting in A-ring and B-ring fragments (Fig. 5.11). I took advantage of fragmentations at the C ring to elucidate the rings of quercetin and myricetin that are taking part in forming their respective dimer linkages. This represents a first step in determining the complete structure of the dimer and can complement other subsequent techniques such as nuclear magnetic resonance (NMR) spectroscopy. Knowing the chemical structure of the dimer is useful for future work because the structure can provide insight into chemical properties such as antioxidant activity. For example, a dimer linkage might be formed through a covalent bond between a carbon of one monomer and phenolic hydroxyl oxygen of a second monomer to form an ether linkage. However, since phenolic hydroxyl groups are more antioxidant-active than ether groups (Van Acker et al., 1996), this would result in reduced antioxidant activity of the dimer per unit monomer, at least in aqueous solutions. The structure of the dimer might also provide understanding of the site of the monomer that is acted on by the laccase enzyme, although this can be complicated by the fact that the initial radical formed can undergo electron delocalization.

121

1,3 + OH A OH

OH 0,2 + OH A HO 0 O 1 B HO O B A C 2 A

4 3 OH OH 0,2B+ OH O OH O 1,3B+ 1,4A+ Figure 5.11. Fragmentation patterns of flavonols in positive ion mode tandem mass spectrometry. Figure adapted from Ma et al., 1997. Left image shows numbering system of bonds in the central ring. Right image shows previously observed fragments resulting from breaking two C-ring bonds. For example, 1,3A+ indicates a fragment ion containing the A ring after breaking bonds 1 and 3.

Upon breaking the C ring of quercetin or myricetin, some of the resulting A-ring fragments have the same number of carbons, hydrogens, and oxygens as some of the B-ring fragments. Hence, these fragments have identical m/z values, even though the structures of the fragments are not the same. For example, for myricetin the 1,3A+ and 0,2B+ fragments both have an m/z of 153 (Ma et al., 1997). Such ions are therefore not informative in elucidating whether the fragment contains the A ring or B ring. However, certain other fragmentations lead to a unique m/z value and these are informative fragment ions for elucidating the rings of the monomers that are taking part in the dimer linkage. Using the bond numbering system shown in Figure 5.11 (adapted from Ma et al., 1997), informative fragment ions were observed for the quercetin dimer isomer #1 (Table 5.1) that could be attributed to breaking bonds 1,3 of the C ring. These informative fragment ions are those with m/z values 451 and 153 (Supplemental Table A3.2). The m/z of 451 can be attributed to a dimer remnant containing the 1,3B+ fragment bonded to an intact monomer, while m/z 153 can be attributed to the 1,3A+ fragment (Fig. 5.12). Taken together, these fragments suggest that the quercetin dimer #1 is not formed by linkage of two A rings. Similarly for the myricetin dimer isomer #2 (Table 5.1), fragment ions with m/z 345 and 329 (Supplemental Table A3.3) can be attributed to dimer remnants containing 1,3B+

122 fragment linked to 0,3B fragment and 1,3B+ fragment linked to 1,3B fragment, respectively (Fig. 5.13). Taken together, these suggest that the myricetin dimer is also not formed by linkage of two A rings and is more specifically formed by linkage of two B rings or a B ring with a C ring. The tandem mass spectrum of non-enzymatically produced quercetin dimer (Supplemental Table A3.4) shows the informative fragment m/z values 451 and 153 as for enzymatically produced quercetin dimer. Similarly, the tandem mass spectrum for non-enzymatically produced myricetin dimer (Supplemental Table A3.5) shows the informative fragment m/z values 345 and 329 as for enzymatically produced myricetin dimer. This suggests the same dimer is formed with or without enzyme.

-2H A) OH B) OH OH OH HO O HO O 1,3B+

OH 1,3A+ OH OH O OH O OH OH HO O

OH OH O

Figure 5.12. Proposed fragment structures of quercetin dimer #1 (refer Table 5.1) after tandem mass spectrometry. Proposed fragments corresponding to m/z A) 451, and B) 153 are shown in the black outline. Brackets are shown in A) to indicate less two hydrogen atoms from the structure shown in the black outline (which arises from the fact that the dimer was formed with accompanying loss of two hydrogens).

123

-4H A) B) -4H

OH OH HO O OH 1,3 + OH B 1,3B+ OH O OH OH HO O OH

OH 0,3B 1,3B OH O

Figure 5.13. Proposed fragment structures of myricetin dimer #2 (refer Table 5.1) after tandem mass spectrometry. Proposed fragments corresponding to m/z A) 345, and B) 329 are shown in the black outline. Brackets are shown to indicate less four hydrogen atoms from the structure shown in the black outline (which arises from the fact that the dimer was formed with accompanying loss of four hydrogens).

124

Almost all previous reports characterizing the structure of an oxidatively produced and purified quercetin dimer use 1H NMR data to suggest that the dimer linkage occurs between the B ring oxygens of one monomer and the C2=C3 double bond of the C ring of a second monomer (Schreier & Miller, 1985; Hirose et al., 1999; Krishnamachari et al., 2004; Tram et al., 2005; Veverka et al., 2013). Slightly different linkages are proposed in the different works and quercetin was oxidized using different methods (in some cases with an enzyme and in other cases with non-enzymatic oxidation) but all these studies agree that the quercetin dimer linkage does not involve the A ring. To the best of my knowledge, there is only one study (Ramos et al., 2006) that reports a quercetin dimer produced by dimer linkage that is between the A ring of one monomer with the C ring of a second monomer. The dimer studied by Ramos et al. (2006) was isolated from the skin of yellow onion as was the dimer studied by Tram et al. (2005) even though the latter group reports the dimer linkage to not involve the A ring. The differences in the extraction procedure between the two research works might be the cause for isolating different dimers.

5.4. Conclusions

In this study, dimers of three out of four tested flavonols (quercetin, morin, and myricetin; but not kaempferol) were produced by treating the flavonol with laccase enzyme. Enzyme-free reactions were also able to produce the same dimer products as those of enzymatic reactions but at reduced rate. HPLC-MS/MS results suggested that the quercetin and myricetin dimer linkages involve at least one B ring. The earlier-formed dimer products may display increased lipid miscibility as suggested from their increased retention time on a C18 reversed- phase HPLC column. This work is the first to report oxidative modification of morin to produce a compound with m/z value consistent with being a dimer of morin with minor fragmentation in the process (i.e. loss of CO group). There was an overall reduction in radical scavenging capacity due to laccase treatment of the flavonol as measured by the TRAP and FRAP antioxidant assays. At the same time, reaction samples completely depleted of starting phenolic monomer still contained antioxidant activity, so there are new useful bioactive products. Future experiments should look into separation of the reaction mixture into different chromatographic fractions for identification of most active components.

125

CHAPTER 6. DISCUSSION

Phytochemical phenolics have a variety of bioactive properties that could be tweaked by changing the molecular structure of the starting phenolic compound. For example, the chemistry of phenolic phytochemicals could be modified by addition of alkyl groups or formation of larger molecular weight oligomers. This may lead to new effects that involve the well-known antioxidant activites of these compounds. It is possible that other bioactivities, besides antioxidant activity, are also altered after making such chemical modifications. As such, two enzyme types were investigated for their potential to modify the chemistry and activity of phenolics.

The first enzyme class, esterase, was used to try to esterify alkyl chains onto phenolic hydroxyl groups. Water has to be minimized in the reaction medium to avoid hydrolytic breakdown of the desired ester product; hence the requirement to use non-aqueous solvents such as organic solvents. At the same time, enzymes rely on water for maintaining optimal flexibility and functionality. Therefore, I carried out experiments to assess the effect of organic solvent- water mixtures during the catalytic process of the enzyme. One bacterial esterase from Pseudomonas putida (PP3645) and one archaeal esterase from Archaeoglobus fulgidus (AF1753) were used for these experiments. Despite the fact that the esterases were still active in organic solvent-water mixtures, they did not catalyze synthetic transesterification reaction between vinyl acetate and 4’- hydroxyacetophenone. On the other hand, a commercial lipase (lipase PS from Amano) produced both hydrolysis and transesterification products, demonstrating that both products could be formed in the same reaction mixture. This lead to the hypothesis that the inability of the esterases to carry out transesterification was due to comparatively lower accessibility of the 4’-hydroxyacetophenone substrate into the active site of the esterases. The lower accessibility of the esterase active sites may be due to naturally more restricted active sites compounded by increased rigidity of the enzymes in the presence of organic solvent (t-amyl alcohol/cyclohexane (1:9, v/v)). Increased enzyme rigidity may be a result of the organic solvent stripping off water from the enzyme, leading to less shielding of electrostatic interactions between active site residues of the enzyme (Klibanov, 2001). Therefore, alanine point mutants of AF1753 were

126 produced in an effort to identify variants with increased relative activity on more bulky long- chain pNP esters, which would presumably imply more accessible enzyme active sites. Assays using partially purified mutants identified mutant variants (Arg37Ala and Glu244Ala) with improved relative activity on longer-chain pNP esters compared to the relative activity of the wild type. Examination of the modeled structure of the mutants identified the predicted formation of a new tunnel leading to the active site. This new tunnel may provide better access of longer alkyl chains to the active site, leading to the observed higher relative activity of these mutants. Additionally, previous studies have found that enzyme activity in organic solvent depends on the pH of the solution in which the enzyme was lyophilized because the protonatable residues retain their last ionization state (for example, the ionization state in the lyophilization buffer) when put into organic solvents (Klibanov, 2001). As mentioned above, these charged residues are less shielded from each other in organic solvents. Therfore, an added benefit of converting the protonatable residues Arg37 and Glu244 of AF1753 to the non-polar alanine may be removal of ionic and polar interactions that were originally responsible for the rigid state of the enzyme in the organic solvent. Subsequently, these mutants might display detectable transesterification activity in t-amyl alcohol/cyclohexane (1:9, v/v). Finally, the lack of transesterification with the arylesterases in this study, despite their maintained hydrolytic activity in organic solvent, indicates that tests for hydrolysis in organic solvent are not adequate on their own to assess transesterification potential of hydrolases.

Lack of successful synthetic reactions with the esterases prompted the use of another enzyme class. Laccase was the second enzyme class that was utilized to try to modify the chemistry of phenolic compounds. In this case, the laccase oxidizes the phenolic to produce radicals, which can then react with each other to form dimers and higher molecular weight oligomers. I carried out initial experiments to characterize the activity of a bacterial laccase (SCO6712, also known as SLAC, from Streptomyces coelicolor) on bioactive phenolic compounds to determine the range of phenolic classes acted on by the enzyme. The enzyme showed activity on syringic acid, caffeic acid, ferulic acid, resveratrol, quercetin, morin, kaempferol and myricetin, suggesting that SCO6712 could be applied in organic syntheses involving bioactive phytochemicals from a wide range of phenolic classes,

127 whereby corresponding compounds could be coupled to oligomeric forms. Mutation studies indicate that saturating mutagenesis of Ser292, as well as Tyr230, is a feasible approach to extending the catalytic efficiency of SCO6712 on bioactive compounds. In particular, Ser292 may be involved in favourable interactions with the ortho hydroxyl groups of phenolic substrates and unfavourable steric clashes with ortho methoxyl groups. Notably, this may give insight into the natural substrate of this enzyme. Bacterial laccases have been proposed to play a role in lignin degradation (Bugg et al., 2010; Majumder et al., 2014). The affinity of Ser292 of SCO6712 towards ortho hydroxyl groups may be an indication that the enzyme does not directly target lignin, whose monomeric building blocks lack ortho hydroxyls. Rather, initial stages of lignin degradation by other enzymes such as peroxidases can produce phenolics with ortho hydroxyls. These can then be acted on by SCO6712 laccase to become radicalized phenolics. These radicalized phenolics would then attack lignin, in this way acting as mediators of SCO6712 activity on the lignin. While the Ser292 position may have unfavourable steric interactions with ortho methoxylated substituents, mutation of this residue to alanine resulted in decreased activity on most tested substrates (Fig. 4.1). This may be an indication that Ser292 residue influences the redox potential of the enzyme. The Ser292 residue is followed by the His293 residue, which is one of the residues that chelates the type 1 copper ion. The side chain hydroxyl hydrogen of Ser292 may hydrogen bond with the side chain nitrogen of His293, resulting in less sharing of electrons from His293 with the type 1 copper. This reduced electron density in the vicinity of the type 1 copper could result in greater ability of the copper to abstract an electron from the enzyme substrate. Upon mutation of Ser292 to an alanine, the alanine side chain cannot hydrogen bond with His293 and there may be greater electron density at the type 1 copper, resulting in lowered ability of the copper to abstract electrons from the substrates of the enzyme (meaning lower redox potential of the enzyme). An effect of hydrogen bonding residues adjacent to residues that directly interact with type 1 copper has been previously reported, as well (Marshall et al., 2009). In that report, a non-hydrogen bonding phenylalanine within the enzyme azurin was mutated to the hydrogen bonding asparagine and the mutant had increased redox potential compared to the wild type. It was proposed that the asparagine can hydrogen bond with a carbonyl oxygen of the protein backbone. This carbonyl oxygen has ionic interactions with the type 1 copper of azurin, and the authors suggested that the reduced electron

128 density resulting from the hydrogen bonding is what lead to the increased redox potential of the enzyme (Marshall et al., 2009).

The products of reaction from laccase treatment of flavonols can vary widely because of electron delocalization. Oligomers of the original flavonol can be expected but it is not straightforward to know where the oligomer linkage will be. In the course of forming oligomers, other side-products may also be formed, such as quinones. It has also been reported that oxidative cleavage of flavonols may occur in the same reaction as oligomer formation, as is the case for quercetin treated with chemical and enzymatic oxidizers (Zhou & Sadik, 2008). Therefore, the end result can be a complicated reaction mixture. It was thus necessary to evaluate the product identities and antioxidant activities from the reaction mixtures of the laccase-treated flavonols. I used HPLC-MS to confirm the production of dimers of three out of four tested flavonols (quercetin, morin, and myricetin; but not kaempferol). HPLC-MS/MS gave initial clues to the site of dimer linkage on quercetin and myricetin. The data suggested an important role for the B ring of quercetin and myricetin in dimer formation. In addition, the lack of dimer formed from kaempferol may indicate that dimer linkage for flavonols occurs through forming a dioxane-type linkage involving two oxygens in the B ring. Such a linkage is found in naturally occurring quercetin dimers in which the B ring oxygens of one monomer are bonded to the C2 and C3 carbons of another monomer (Fig. 5.7) (Tram et al., 2005). The two isomeric forms of each of the dimers of quercetin and myricetin (Table 5.1) may be regioisomers formed due to the fact that a B ring oxygen can react either with C2 to form one regioisomer or with C3 to form the other regioisomer. Despite the radical mediated process, only one dimer of m/z 603 was the predominant product observed from the laccase reaction samples (Fig. 5.1). This may be a reflection of the fact that the initially produced radical is stabilized by extensive electron delocalization, such that only a limited set of reaction pathways involving the most reactive resonance forms lead to product formation. While studying the formation of quercetin dimer, Krishnamachari et al. (2004) concluded that the dimer formation mechanism likely involves reaction of two quercetin radicals rather than reaction of a quercetin ortho quinone with quercetin (Fig. 6.1). The formation of morin dimers in this study could also be explained by this proposed mechanism, since the meta substitution in morin’s B ring means morin cannot form an

129 ortho quinone and a meta quinone would be an extremely unstable species with no previously reported such structure (Dewar & Gleicher, 1966).

130

OH A) OH HO HO HO HO O O O C HO HO O O O OH O OH O O O OH O HO O H OH O OH OH

OH HO

HO O O OH O O OH OH O O OH

B) O O O O HO O HO O OH OH O OH OH O

OH HO

HO O O OH O O X OH OH O O OH

OH Figure 6.1. Two possible mechanisms of quercetin dimer formation. A) Reaction of two radicals to form a dimer. B) Formation of an ortho quinone from a diradical followed by concerted cyclic reaction to form a dimer. Adapted from Krishnamachari et al. (2004).

131

The enzymatic and non-enzymatic quercetin reactions appear to give slightly different product profiles after 24 h, when all the substrate has been depleted in both cases. Now this may simply be due to the fact that initially formed products are produced sooner in the enzymatic reaction and so have had more time to undergo further reactions while the non-enzymatic counterparts have not yet had enough time to react to the same extent. On the other hand, it may also suggest that the initially formed radicals are produced at a faster rate in the enzymatic reaction, and so achieve a higher concentration such that they are more able to react with each other to form coupling products rather than undergoing decomposition reactions. Furthermore, the non-enzymatic process may be producing hydrogen peroxide (Canada et al., 1990) that may be taking part in reactions with quercetin and its oxidized products. Finally, comparison of the TRAP antioxidant activities based on lag time for quercetin for 1 h non-enzymatic and 5 min enzymatic reactions supports the idea that different secondary reactions occur (or different relative extents of the same secondary reactions) between the enzymatic and non-enzymatic processes. To gain an overall view of the effect of laccase treatment on the inherent radical scavenging capacity of the flavonols, TRAP and FRAP antioxidant assays were carried out with whole laccase reaction-mixes. In this way, all products are assayed at once without the very laborious process of isolating each product. The results showed a decrease in total antioxidant activity with decrease in original starting phenolic substrate compound. The decrease was least pronounced with quercetin reaction-mix both with and without enzyme, (particularly for FRAP antioxidant activity) suggesting this compound is most promising for future characterization work of its products. The eventual total decrease in antioxidant activity (that seen for 24 h reaction-mixes) was slightly more for the reactions with enzyme (59 % ± 3 % and 50 % ± 1 % residual FRAP value for 24 h quercetin reactions without and with enzyme, respectively), which may be a reflection of the slightly different product profiles and the fact that secondary reactions have had more time to proceed in the enzymatic case than non-enzymatically, leading to depletion of the more active antioxidants. In almost all cases, as expected, the percent total antioxidant activity was at least equal to (and in some cases greater than) the percent remaining amount of starting phenolic substrate (i.e. percent remaining amount of quercetin, morin, or myricetin) from the laccase reaction samples. However, this was not the case for antioxidant activity (as measured by lag time) of

132 quercetin reaction samples after 5 min with enzyme (Fig. 5.8), in which the antioxidant activity was 40 % of initial while the amount of quercetin was still 80 % of initial; indicating that the mixture of products inhibited the antioxidant effect of the starting phenolic substrate during the lag phase of the antioxidant assay. However, as can be seen from figure 5.9, the expected minimum 80 % of antioxidant activity was present for the 5 min laccase reaction sample when taking into account antioxidant activity during both the lag phase and propogation phase of DCFH oxidation by AAPH (as measured by end time of DCFH oxidation). Therefore, the antagonistic interaction between quercetin and its laccase reaction products is only during the lag phase. Antagonism between antioxidants in mixtures has been previously reported and the proposed mechanism is that the more effective antioxidant (in this case quercetin) regenerates the less effective antioxidant (in this case the laccase generated quercetin products) rather than reacting with the radical of the antioxidant assay (in this case AAPH radicals) (Fig. 6.2) (Peyrat- Maillard et al., 2003). A modified form of this mechanism can be proposed to explain the results observed in this thesis work, as follows. In the sample containing only quercetin, quercetin reacts with AAPH radicals to form a quercetin radical and the initially formed quercetin radicals undergo a further reaction with AAPH radicals to produce quinones of quercetin (Fig. 6.3A equations (1) and (2)). However, for the 5 min laccase reaction sample, quercetin reacts with AAPH radicals to form a quercetin radical as usual, but the quercetin radicals then react with the laccase-generated products rather than with another round of AAPH radicals (Fig. 6.3A equations (3), (4), and (5); and B), thus explaining why only half the expected minimum antioxidant activity is observed in the lag phase. These new products then show their antioxidant effect during the propagation phase of DCFH oxidation. The main difference between this mechanism and that proposed by Peyrat-Maillard et al. (2003) is that the less effective antioxidants (laccase-generated products) are not being regenerated because in the first place they never reacted with AAPH radicals during the lag phase.

133

. . A1H + ROO  A1 + ROOH (1) . . A1 + A2H  A1H + A2 (2)

Figure 6.2. Mechanism of antioxidant antagonism proposed by Peyrat-Maillard et al. (2003).

The more effective antioxidant (A2H) is used up to regenerate the less effective antioxidant

(A1H). ROOH represents a reactive oxygen species.

134

A) . . QH2 + ROO  QH + ROOH (1) .QH + ROO.  Q + ROOH (2)

. . QH2 + ROO  QH + ROOH (3) .QH + QP  HQ-QP. (4) .QH + HQ-QP.  Q + HQ-QPH (5)

B) OH O HO OH OH OH O O OH OH O O O O HO OH O O H O OH OH (4) OH HO O O O O C OH O O HO OH O HO O O OH HO O OH O OH HO HO O OH OH OH

O OH HO OH OH O O HO OH O O O OH OH (5) O O OH O HO O O OH O HO HO O OH OH

135

Figure 6.3. Mechanism of antioxidant antagonism that involves reaction of quercetin radicals with laccase-generated quercetin products. A) Shows the schematic sequence of reactions. Equations (1) and (2) are antioxidant assay reactions that would occur in the absence of laccase generated products; and equations (3), (4), and (5) are antioxidant assay reactions that would occur in the presence . . of laccase-generated products. QH2, QH, Q, QP, HQ-QPH, and ROO represent quercetin, quercetin radical, quercetin quinone, laccase generated product, laccase-generated product that is covalently bonded to quercetin, and peroxyl radical. B) Shows possible corresponding structures for reactions (4) and (5). In this case QP is a predicted structure of a quercetin dimer.

136

In conclusion, the feasibility of using esterases and laccases to modify the chemistry of phenolics was evaluated in this thesis work. The esterases did not yield desired product while the laccase produced new phenolic dimers. The inherent antioxidant activity was overall gradually reduced upon oxidative modification of the phenolics with or without laccase enzyme. In this respect, an enzymatic approach may be advantageous compared to a chemically- mediated oxidative modification by making it possible to more easily stop the oxidative modification reaction before depletion of the more active new antioxidant products. This may be achieved by immobilizing the enzyme onto a water-insoluble matrix such that it can easily be removed from the reaction-mix by simple filtration. Furthermore, the enzymatic approach may be more selective than a chemical oxidizing agent in so far as the initially formed higher molecular weight products may not be accessible to the laccase active site and cannot be directly oxidized by the enzyme. Finally, it may be possible to carry out the laccase reaction in organic solvent-water mixtures in order to vary the product profile towards more lipophilic products due to preferential partitioning of the products into the organic solvent phase. Such products might find application as antioxidant preservatives of emulsified food lipids. In this case, while the esterases did not yield synthetic products, lessons learned from the esterase reactions in organic solvent-water mixtures may be used to optimize the laccase reactions.

137

CHAPTER 7. FUTURE RESEARCH

The following recommendations are made for further research to continue the investigations started in this project:

7.1. Further characterization of esterases for transesterification potential, continuing work of Chapter 3

- Determine the hydrolytic capability of wild type esterases AF1753 and PP3645 in additional water-miscible organic solvents DMSO (log P = -1.22), methanol (log P = -0.76), acetonitrile (log P = -0.34), acetone (log P = -0.23), tetrahydrofuran (log P = 0.49), and 1- butanol (log P = 0.8), and compare to t-amyl alcohol (log P = 0.89), while maintaining the same level of water in the reaction medium for all solvents. This can determine if there is a direct correlation between hydrolysis activity and solvent log P for these enzymes. - Examine the transesterification activity of wild type esterases AF1753 and PP3645 in (co-solvent)/cyclohexane (1:9, v/v) using vinyl acetate and 4’-hydroxyacetophenone substrates, where the co-solvent is a water-mimicking hydrophilic solvent that is able to engage in hydrogen bonds. The use of such co-solvents (for example, acetone (log P = -0.23) and any other solvents found to be miscible with cyclohexane) may help make the enzymes less rigid (Klibanov, 2001) to more easily accommodate the 4’-hydroxyacetophenone substrate for transesterification. - Characterize the ability of esterase AF1753 mutants Arg37Ala and Glu244Ala to carry out transesterification of vinyl acetate and 4’-hydroxyacetophenone. The reaction can be carried out in t-amyl alcohol/cyclohexane (1:9, v/v) using lyophilized enzymes. Ultimately, it may be found that the esterases are not a feasible approach to add alkyl groups via transesterification reactions. For this reason, it may be advisable to assess other enzymes in parallel to the esterases. Lipases would be a good candidate for this purpose since these enzymes have been used many times to carry out transesterification in organic solvent (as was seen for the lipase used in this thesis work). Acyltransferases are another good candidate group of enzymes, with the advantage that the reaction can be carried out in water. Aside from transferring an alkyl group via an acyl linkage, prenylation would involve attaching an alkyl

138 group in the form of the branched 5-carbon prenyl group. For this purpose, prenyltransferases would be appropriate enzymes.

7.2. Further assessment of biochemical potential of laccase SCO6712, continuing work of

Chapter 4

- Use recombinant cell culture conditions with a range of copper amounts added to the medium to further improve copper content of the laccase enzyme. Currently, even with the use of microaerobic cultivation, the enzyme used in this study still has less than the theoretical four coppers per protein monomer. - Carry out laccase SCO6712 assays on two to three members of the flavonoid subclasses isoflavones, flavanones, anthocyanidins, and flavanols to further investigate the substrate limits of the enzyme. This may help identify molecular substitution patterns that are crucial for action of SCO6712 laccase on flavonoids. In addition to the flavonoids, phenolics from other classes could be used to try to form naturally occurring phenolic oligomers and isomers thereof. For example, the monolignols could be used to try to produce lignans, which are investigated for their antioxidant and anticancer activities (reviewed in Marcotullio et al., 2014). Another option is to combine monolignols with flavonoids to produce the heterocoupled falvonolignans. These compounds, most famously found in milk thistle plants, are being investigated for a range of medicinal applications (reviewed in Gazak et al., 2007). Finally, stilbenes (such as resveratrol) could be oxidatively oligomerized to form compounds that may be potential anticancer agents (reviewed in Xue et al., 2014). - Perform saturation mutagenesis of Ser292 residue of laccase SCO6712 to potentially identify mutations that lead to improved enzyme activity on phenolics with ortho-dihydroxy moieties such as quercetin. Additionally, mutation of residues (such as Ser292 and others) that are adjacent to the type I copper-binding residues may result in increases in redox potential of the enzyme. For example, if the mutation leads to greater hydrogen bonding with the copper- binding histidines this may increase redox potential of the enzyme. As with the esterases, it may be advisable to explore the use of other enzyme types. In this case, peroxidase enzymes can catalyze oxidation of phenolic substrates. Using peroxidases would allow controlling the concentration of the oxidant by gradual addition of hydrogen

139 peroxide to the reaction mixture, which may influence the composition of the resultant products (Saake et al., 1996). The oxygen concentration (for reaction with laccases) can also be controlled, but it may not be as technically easy as controlling hydrogen peroxide concentrations (for reactions with peroxidases) since all solutions would have to first be purged of oxygen and handled anaerobically before addition of desired oxygen amounts.

7.3. Further examination of bioactivity of flavonol dimers, continuing work of Chapter 5

- Use preparatory HPLC to purify the dimers of quercetin, morin, and myricetin and characterize their structures using NMR analysis. This may provide insight into the mechanism of reaction leading to dimer formation. It may also help elucidate correlations between molecular structure and function. - Carry out TRAP and FRAP antioxidant assays on the purified quercetin dimer and compare to the monomer. This will quantitatively show the difference in inherent antioxidant activity of the dimer compared to the monomer as measured in a single-phase aqueous system (where lipid solubility is not an influence), since potential mixture effects influencing the dimer activity will have been removed. Changes in antioxidant activity could then be related to changes in structure resulting from dimerization. - Carry out a two-phase antioxidant assay using emulsified lipid as the target, and AAPH or AMVN as radical generators (see Appendix 4 for challenges encountered with initial attempts of using linoleic acid as emulsified lipid and some potential ways to overcome those challenges; and Appendix 5 for attempts at directly evaluating solubility in aqueous and lipid solvents). This assay can be used to compare the purified dimer to the monomer. These results can be compared to the antioxidant results in single-phase aqueous TRAP assays to see the advantages gained towards lipids by formation of flavonol dimers. In addition to antioxidant activity, it will also be of interest to examine the broader physiological health benefits of the flavonol dimer products. For example, dimers of apigenin can inhibit amyloid β toxicity in the treatment of Alzheimer’s disease (Thapa et al., 2011). The new flavonol dimers produced in this study may also have improved inhibition activity in this regard, or may be advantageous in terms of reduced detrimental side effects. Furthermore, flavonoid oligomers can have antimicrobial applications. For example, proanthocyanidins are oligomers of catechin and epicatechin and these oligomeric compounds can disrupt adhesion

140 and motility of pathogenic microorganisms (Eydelnant & Tufenkji, 2008; O’May & Tufenkji, 2011). Accordingly, an interesting future research direction would be to investigate the effect of laccase-generated dimers produced in this study on the adhesion and motility of bacterial pathogens.

141

REFERENCES

Abou-zaid, M. M., & Nozzolillo, C. (1991). Flavonol glycosides from needles of Pinus banksiana. Biochemical Systematics and Ecology, 19(3), 237-240.

Abou-Zaid, M. M., & Nozzolillo, C. (1999). 1-O-galloyl-α-L-rhamnose from Acer rubrum. Phytochemistry, 52(8), 1629-1631.

Abou-Zaid, M. M., Lombardo, D. A., & Nozzolillo, C. (2009). Methyl gallate is a natural constituent of maple (Genus Acer) leaves. Natural Product Research, 23(15), 1373- 1377.

Addis, P. B. & Warner, G. J. (1991). The potential health aspects of lipid oxidation products in foods. In O. I. Aruoma & B. Halliwell (Eds.). Free Radicals and Food Additives (pp. 25-28). London: Taylor & Francis.

Adelakun, O. E., Kudanga, T., Green, I. R., le Roes-Hill, M., & Burton, S. G. (2012). Enzymatic modification of 2,6-dimethoxyphenol for the synthesis of dimers with high antioxidant capacity. Process Biochemistry, 47(12), 1926-1932.

Adman, E. T., Stenkamp, R. E., Sieker, L. C., & Jensen, L. H. (1978). A crystallographic model for azurin at 3 Å resolution. Journal of Molecular Biology, 123(1), 35-47.

Adman, E. T., Turley, S., Bramson, R., Petratos, K., Banner, D., Tsernoglou, D., Beppu, T., & Watanabe, H. (1989). A 2.0-Å structure of the blue copper protein (cupredoxin) from Alcaligenes faecalis S-6. Journal of Biological Chemistry, 264(1), 87-99.

Aida, Y., Tamogami, S., Kodama, O., & Tsukiboshi, T. (1996). Synthesis of 7- Methoxyapigeninidin and Its Fungicidal Activity against Gloeocercospora sorghi. Bioscience, Biotechnology and Biochemistry, 60(9), 1495-1496.

Alemán, M., Bou, R., Guardiola, F., Durand, E., Villeneuve, P., Jacobsen, C., & Sørensen, A. (2015). Antioxidative effect of lipophilized caffeic acid in fish oil enriched mayonnaise and milk. Food Chemistry, 167, 236-244.

Allevi, P., Ciuffreda, P., Longo, A., & Anastasia, M. (1998). Lipase-catalysed chemoselective monoacetylation of hydroxyalkylphenols and chemoselective removal of a single acetyl group from their diacetates. Tetrahedron Asymmetry, 9(16), 2915-2924.

Balasundram, N., Sundram, K., & Samman, S. (2006). Phenolic compounds in plants and agri-industrial by-products: Antioxidant activity, occurrence, and potential uses. Food Chemistry, 99(1), 191-203.

Banerjee, A., Kaul, P., Sharma, R., & Banerjee, U. C. (2003). A high-throughput amenable colorimetric assay for enantioselective screening of nitrilase-producing

142 microorganisms using pH sensitive indicators. Journal of Biomolecular Screening, 8(5), 559- 565.

Benzie, I. F. F., & Strain, J. J. (1996). The ferric reducing ability of plasma (FRAP) as a measure of 'antioxidant power': The FRAP assay. Analytical Biochemistry, 239(1), 70-76.

Bhatnagar, B. S., Bogner, R. H., & Pikal, M. J. (2007). Protein stability during freezing: Separation of stresses and mechanisms of protein stabilization. Pharmaceutical Development and Technology, 12(5), 505-523.

Boerjan, W., Ralph, J., & Baucher, M. (2003). Lignin Biosynthesis. Annual Review of Plant Biology, 54, 519-546.

Bornscheuer, U. T. (2002). Microbial carboxyl esterases: Classification, properties and application in biocatalysis. FEMS Microbiology Reviews, 26(1), 73-81.

Bourbonnais, R., & Paice, M. G. (1992). Demethylation and delignification of kraft pulp by Trametes versicolor laccase in the presence of 2,2′-azinobis-(3-ethylbenzthiazoline-6- sulphonate). Applied Microbiology and Biotechnology, 36(6), 823-827.

Brady, L., Brzozowski, A. M., Derewenda, Z. S., Dodson, E., Dodson, G., Tolley, S., Turkenburg, J. P., Christiansen, L., Huge-Jensen, B., Norskov, L., Thim, L., & Menge, U. (1990). A serine protease triad forms the catalytic centre of a triacylglycerol lipase. Nature, 343(6260), 767-770.

Brzozowski, A. M., Derewenda, U., Derewenda, Z. S., Dodson, G. G., Lawson, D. M., Turkenburg, J. P., Bjorkling, F., Huge-Jensen, B., Patkar, S. A., & Thim, L. (1991). A model for interfacial activation in lipases from the structure of a fungal lipase-inhibitor complex. Nature, 351(6326), 491-494.

Brzozowski, A. M., Savage, H., Verma, C. S., Turkenburg, J. P., Lawson, D. M., Svendsen, A., & Patkar, S. (2000). Structural origins of the interfacial activation in Thermomyces (Humicola) lanuginosa lipase. Biochemistry, 39(49), 15071-15082.

Bugg, T. D. H., Ahmad, M., Hardiman, E. M., & Singh, R. (2010). The emerging role for bacteria in lignin degradation and bio-product formation. Current Opinion in Biotechnology, 22(3), 394-400.

Bukh, C., & Bjerrum, M. J. (2010). The reversible depletion and reconstitution of a copper ion in Coprinus cinereus laccase followed by spectroscopic techniques. Journal of Inorganic Biochemistry, 104(10), 1029-1037.

Cai, X., Fang, Z., Dou, J., Yu, A., & Zhai, G. (2013). Bioavailability of quercetin: Problems and promises. Current Medicinal Chemistry, 20(20), 2572-2582.

143

Camarero, S., Ibarra, D., Martínez, A. T., Romero, J., Gutiérrez, A., & del Río, J. C. (2007). Paper pulp delignification using laccase and natural mediators. Enzyme and Microbial Technology, 40(5), 1264-1271.

Canada, A. T., Giannella, E., Nguyen, T. D., & Mason, R. P. (1990). The production of reactive oxygen species by dietary flavonols. Free Radical Biology and Medicine, 9(5), 441- 449.

Cañas, A. I., & Camarero, S. (2010). Laccases and their natural mediators: Biotechnological tools for sustainable eco-friendly processes. Biotechnology Advances, 28(6), 694-705.

Carunchio, F., Crescenzi, C., Girelli, A. M., Messina, A., & Tarola, A. M. (2001). Oxidation of ferulic acid by laccase: Identification of the products and inhibitory effects of some dipeptides. Talanta, 55(1), 189-200.

Chaiyasit, W., Elias, R. J., McClements, D. J., & Decker, E. A. (2007). Role of physical structures in bulk oils on lipid oxidation. Critical Reviews in Food Science and Nutrition, 47(3), 299-317.

Chalas, J., Claise, C., Edeas, M., Messaoudi, C., Vergnes, L., Abella, A., & Lindenbaum, A. (2001). Effect of ethyl esterification of phenolic acids on low-density lipoprotein oxidation. Biomedicine and Pharmacotherapy, 55(1), 54-60.

Chebil, L., Anthoni, J., Humeau, C., Gerardin, C., Engasser, J. M., & Ghoul, M. (2007). Enzymatic acylation of flavonoids: Effect of the nature of the substrate, origin of lipase, and operating conditions on conversion yield and regioselectivity. Journal of Agricultural and Food Chemistry, 55(23), 9496-9502.

Chen, Y., Xiao, L., Liu, Y., Li, X., Zhang, J., & Shu, Y. (2014). A lipase-based electrochemical biosensor for target DNA. Microchimica Acta, 181(5-6), 615-621.

Cho, M. H., Park, H. L., Park, J. H., Lee, S. W., Bhoo, S. H., & Hahn, T. R. (2012). Characterization of regiospecific flavonoid 3′/5′-O-methyltransferase from tomato and its application in flavonoid biotransformation. Journal of the Korean Society for Applied Biological Chemistry, 55(6), 749-755.

Claus, H. (2004). Laccases: Structure, reactions, distribution. Micron, 35(1-2), 93-96.

Colman, P. M., Freeman, H. C., Guss, J. M., Murata, M., Norris, V. A., Ramshaw, J. A. M., & Venkatappa, M. P. (1978). X-ray crystal structure analysis of plastocyanin at 2.7 Å resolution. Nature, 272(5651), 319-324.

Coleman, M. H., & Macrae, A. R. (1981). Fat process and composition. US Patent 4,275,081.

144

Coplen, T. B. (1996). Atomic weights of the elements 1995 (Technical Report). Pure and Applied Chemistry, 68(12), 2339-2359.

Cuvelier, M., Bondet, V., & Berset, C. (2000). Behavior of phenolic antioxidants in a partitioned medium: Structure-activity relationship. Journal of the American Oil Chemists' Society, 77(8), 819-823.

Dangles, O., Fargeix, G., & Dufour, C. (1999). One-electron oxidation of quercetin and quercetin derivatives in protic and non protic media. Journal of the Chemical Society. Perkin Transactions 2(7), 1387-1395.

Dao, T. T. H., Linthorst, H. J. M., & Verpoorte, R. (2011). Chalcone synthase and its functions in plant resistance. Phytochemistry Reviews, 10(3), 397-412.

De Laeter, J. R., Böhlke, J. K., De Bièvre, P., Hidaka, H., Peiser, H. S., Rosman, K. J. R., & Taylor, P. D. P. (2003). Atomic weights of the elements: Review 2000 (IUPAC Technical Report). Pure and Applied Chemistry, 75(6), 683-800.

Dewar, M. J. S., & Gerald Gleicher, J. A. Y. (1966). Ground states of conjugated molecules. VII.* compounds containing nitrogen and oxygen. The Journal of Chemical Physics, 44(2), 759-773.

Dewick, P. M. (1995). The biosynthesis of shikimate metabolites. Natural Product Reports, 12(2), 101-133.

Dordick, J. S. (1992). Designing enzymes for use in organic solvents. Biotechnology Progress, 8(4), 259-263.

Dundas, J., Ouyang, Z., Tseng, J., Binkowski, A., Turpaz, Y., & Liang, J. (2006). CASTp: Computed atlas of surface topography of proteins with structural and topographical mapping of functionally annotated residues. Nucleic Acids Research, 34(WEB. SERV. ISS.), W116-W118.

Duran, N., & Duran, M. (2000). Enzyme applications in the textile industry. Review of Progress in Coloration and Related Topics, 30, 41-44.

Durao, P., Bento, I., Fernandes, A.T., Melo, E.P., Lindley, P.F., & Martins, L.O. (2006). Perturbations of the T1 copper site in the CotA laccase from Bacillus subtilis: Structural, biochemical, enzymatic and stability studies. Journal of Biological Inorganic Chemistry, 11(4), 514-526.

Durao, P., Chen, Z., Fernandes, A. T., Hildebrandt, P., Murgida, D. H., Todorovic, S., Pereira, M. M., Melo, E. P., & Martins, L. O. (2008a). Copper incorporation into recombinant CotA laccase from Bacillus subtilis: Characterization of fully copper loaded enzymes. Journal of Biological Inorganic Chemistry, 13(2), 183-193.

145

Durao, P., Chen, Z., Silva, C. S., Soares, C. M., Pereira, M. M., Todorovic, S., Hildebrandt, P., Bento, I., Lindley, P. F., & Martins, L. O. (2008b). Proximal mutations at the type 1 copper site of CotA laccase: Spectroscopic, redox, kinetic and structural characterization of I494A and L386A mutants. Biochemical Journal, 412(2), 339-346.

EFSA Panel on Food Additives and Nutrient Sources added to Food (ANS). (2012). Scientific opinion on the re-evaluation of butylated hydroxytoluene BHT (E 321) as a food additive. EFSA Journal, 10(3), 2588-2630.

Eydelnant, I. A., & Tufenkji, N. (2008). Cranberry derived proanthocyanidins reduce bacterial adhesion to selected biomaterials. Langmuir : the ACS Journal of Surfaces and Colloids, 24(18), 10273-10281.

Fabre, N., Rustan, I., De Hoffmann, E., & Quetin-Leclercq, J. (2001). Determination of flavone, flavonol, and flavanone aglycones by negative ion liquid chromatography electrospray ion trap mass spectrometry. Journal of the American Society for Mass Spectrometry, 12(6), 707- 715.

Faulds, C. B. (2010). What can feruloyl esterases do for us? Phytochemistry Reviews, 9(1), 121-132.

Fenster, K. M., Parkin, K. L., & Steele, J. L. (2003). Nucleotide sequencing, purification, and biochemical properties of an arylesterase from Lactobacillus casei LILA. Journal of Dairy Science, 86(8), 2547-2557.

Figueroa-Espinoza, M. C., & Villeneuve, P. (2005). Phenolic acids enzymatic lipophilization. Journal of Agricultural and Food Chemistry, 53(8), 2779-2787.

Fitzpatrick, P. A., & Klibanov, A. M. (1991). How can the solvent affect enzyme enantioselectivity? Journal of the American Chemical Society, 113(8), 3166-3171.

Fitzpatrick, P. A., Ringe, D., & Klibanov, A. M. (1992). Computer-assisted modeling of subtilisin enantioselectivity in organic solvents. Biotechnology and Bioengineering, 40(6), 735- 742.

Frankel, E. N., Huang, S., Kanner, J., & Bruce German, J. (1994). Interfacial phenomena in the evaluation of antioxidants: Bulk oils vs emulsions. Journal of Agricultural and Food Chemistry, 42(5), 1054-1059.

Galli, I., Musci, G., & Bonaccorsi Di Patti, M. C. (2004). Sequential reconstitution of copper sites in the multicopper oxidase CueO. Journal of Biological Inorganic Chemistry, 9(1), 90-95.

Garavaglia, S., Cambria, M. T., Miglio, M., Ragusa, S., Iacobazzi, V., Palmieri, F., D'Ambrosio, C., Scaloni, A., Rizzi, M. (2004). The structure of Rigidoporus lignosus laccase

146 containing a full complement of copper ions, reveals an asymmetrical arrangement for the T3 copper pair. Journal of Molecular Biology, 342(5), 1519-1531.

Garrote, G., Cruz, J. M., Moure, A., Domínguez, H., & Parajó, J. C. (2004). Antioxidant activity of byproducts from the hydrolytic processing of selected lignocellulosic materials. Trends in Food Science and Technology, 15(3-4), 191-200.

Gavezzotti, P., Vavříková, E., Valentová, K., Fronza, G., Kudanga, T., Kuzma, M., Riva, S., Biedermann, D., & Křen, V. (2014). Enzymatic oxidative dimerization of silymarin flavonolignans. Journal of Molecular Catalysis B: Enzymatic, 109, 24-30.

Gazak, R., Walterová, D., & Křen, V. (2007). Silybin and silymarin - New and emerging applications in medicine. Current Medicinal Chemistry, 14(3), 315-338.

Giovanelli, G., & Ravasini, G. (1993). Apple Juice Stabilization by Combined Enzyme- Membrane Filtration Process. LWT - Food Science and Technology, 26(1), 1-7.

Giordani, M. A., Collicchio, T. C. M., Ascêncio, S. D., De Oliveira Martins, D. T., Balogun, S. O., Bieski, I. G. C., Da Silva, L. A., Colodel, E. M., De Souza, R. L., De Souza, D. L. P., De França, S. A., Andrade, C. M. B., & Kawashita, N. H. (2015). Hydroethanolic extract of the inner stem bark of Cedrela odorata has low toxicity and reduces hyperglycemia induced by an overload of sucrose and glucose. Journal of Ethnopharmacology, 162, 352-361.

Gorham, J. (1989). Stilbenes and Phenanthrenes. In J. B. Harborne (Ed.). Methods in plant biochemistry: Volume 1 Plant Phenolics (pp. 159-196). London: Academic Press.

Gorman, L. A. S., & Dordick, J. S. (1992). Organic solvents strip water off enzymes. Biotechnology and Bioengineering, 39(4), 392-397.

Grochulski, P., Li, Y., Schrag, J. D., & Cygler, M. (1994). Two conformational states of Candida rugosa lipase. Protein Science, 3(1), 82-91.

Gross, G. G. (1992). Enzymes in the Biosynthesis of Hydrolyzable Tannins. In R. W. Hemingway & P. E. Laks (Eds.). Plant Polyphenols: Synthesis, Properties, Significance (pp. 43- 49). New York: Plenum Press.

Gulsen, A., Makris, D. P., & Kefalas, P. (2007). Biomimetic oxidation of quercetin: Isolation of a naturally occurring quercetin heterodimer and evaluation of its in vitro antioxidant properties. Food Research International, 40(1), 7-14.

Gunne, M., Al-Sultani, D., & Urlacher, V. B. (2013). Enhancement of copper content and specific activity of CotA laccase from Bacillus licheniformis by coexpression with CopZ copper chaperone in E. coli. Journal of Biotechnology, 168(3), 252-255.

147

Gunne, M., Hoppner, A., Hagedoorn, P. L., & Urlacher, V. B. (2013). Structural and redox properties of the small laccase Ssl1 from Streptomyces sviceus. FEBS Journal, 281(18), 4307-4318.

Hajji, H. E., Nkhili, E., Tomao, V., & Dangles, O. (2006). Interactions of quercetin with iron and copper ions: Complexation and autoxidation. Free Radical Research, 40(3), 303-320.

Hall, J. F., Kanbi, L. D., Strange, R. W., & Hasnain, S. S. (1999). Role of the axial ligand in type 1 Cu centers studied by point mutations of Met148 in rusticyanin. Biochemistry, 38(39), 12675-12680.

Harborne, J. B. (1989). General Procedures. In J. B. Harborne (Ed.). Methods in plant biochemistry: Volume 1 Plant Phenolics (pp. 1-28). London: Academic Press.

Harborne, J. B., Baxter, H., & Moss, G. P. (Eds.). (1990). Phytochemical dictionary: Handbook of bioactive compounds from plants (2nd ed.). London: Taylor & Francis.

Heller, W., & Forkmann, G. (1994). Biosynthesis of flavonoids. In J. B. Harborne (Ed.). The Flavonoids: Advances in Research since 1986 (pp. 499-501). London: Chapman & Hall.

Hirose, Y., Fujita, T., & Nakayama, M. (1999). Structure of doubly-linked oxidative product of quercetin in lipid peroxidation. Chemistry Letters(8), 775-776.

Huang, D., Boxin, O. U., & Prior, R. L. (2005). The chemistry behind antioxidant capacity assays. Journal of Agricultural and Food Chemistry, 53(6), 1841-1856.

Ishikawa, T., Suzukawa, M., Ito, T., Yoshida, H., Ayaori, M., Nishiwaki, M., Yonemura, A., Hara, Y., & Nakamura, H. (1997). Effect of tea flavonoid supplementation on the susceptibility of low- density lipoprotein to oxidative modification. American Journal of Clinical Nutrition, 66(2), 261-266.

Jaeger, K. E., Dijkstra, B. W., & Reetz, M. T. (1999). Bacterial biocatalysts: Molecular biology, three-dimensional structures, and biotechnological applications of lipases. Annual Review of Microbiology, 53, 315-351.

Jala, R. C. R., Hu, P., Yang, T., Jiang, Y., Zheng, Y., & Xu, X. (2012). Lipases as biocatalysts for the synthesis of structured lipids. In G. Sandoval (Ed.). Lipases and Phospholipases Methods and Protocols (pp. 403-408). New York: Springer

Johannes, C., & Majcherczyk, A. (2000). Laccase activity tests and laccase inhibitors. Journal of Biotechnology, 78(2), 193-199.

Jørgensen, L. V., & Skibsted, L. H. (1998). Flavonoid deactivation of ferrylmyoglobin in relation to ease of oxidation as determined by cyclic voltammetry. Free Radical Research, 28(3), 335-351.

148

Kale, R., Saraf, M., Juvekar, A., & Tayade, P. (2006). Decreased B16F10 melanoma growth and impaired tumour vascularization in BDF1 mice with quercetin-cyclodextrin binary system. Journal of Pharmacy and Pharmacology, 58(10), 1351-1358.

Kamat, S., Beckman, E. J., & Russell, A. J. (1992). Role of diffusion in nonaqueous enzymology. I. Theory. Enzyme and Microbial Technology, 14(4), 265-271.

Kaplan, M., Hayek, T., Raz, A., Coleman, R., Dornfeld, L., Vaya, J., & Aviram, M. (2001). Pomegranate juice supplementation to atherosclerotic mice reduces macrophage lipid peroxidation, cellular cholesterol accumulation and development of atherosclerosis. Journal of Nutrition, 131(8), 2082-2089.

Kaur, H. & Perkins, J. (1991). The free-radical chemistry of food additives. In O. I. Aruoma & B. Halliwell (Eds.). Free Radicals and Food Additives (pp. 25-28). London: Taylor & Francis.

Kawai, S., Umezawa, T., & Higuchi, T. (1988). Degradation mechanisms of phenolic β- 1 lignin substructure model compounds by laccase of Coriolus versicolor. Archives of Biochemistry and Biophysics, 262(1), 99-110.

Keck, C. M., & Müller, R. H. (2006). Drug nanocrystals of poorly soluble drugs produced by high pressure homogenisation. European Journal of Pharmaceutics and Biopharmaceutics, 62(1), 3-16.

Kelley, L. A., & Sternberg, M. J. (2009). Protein structure prediction on the Web: a case study using the Phyre server. Nature Protocols, 4(3), 363-371.

Kim, K. K., Song, H. K., Shin, D. H., Hwang, K. Y., Choe, S., Yoo, O. J., & Suh, S. W. (1997a). Crystal structure of carboxylesterase from Pseudomonas fluorescens, an α/β hydrolase with broad substrate specificity. Structure, 5(12), 1571-1584.

Kim, K. K., Song, H. K., Shin, D. H., Hwang, K. Y., & Sun, S. W. (1997b). The crystal structure of a triacylglycerol lipase from Pseudomonas cepacia reveals a highly open conformation in the absence of a bound inhibitor. Structure, 5(2), 173-185.

Kim, C., Lorenz, W. W., Hoopes, J. T., & Dean, J. F. D. (2001). Oxidation of phenolate siderophores by the multicopper oxidase encoded by the Escherichia coli yacK gene. Journal of Bacteriology, 183(16), 4866-4875.

Kim, S., López, C., Güebitz, G., & Cavaco-Paulo, A. (2008). Biological coloration of flax fabrics with flavonoids using laccase from Trametes hirsuta. Engineering in Life Sciences, 8(3), 324-330.

Klibanov, A. (1997). Why are enzymes less active in organic solvents than in water? Trends in Biotechnology, 15(3), 97-101.

149

Klibanov, A. M. (2001). Improving enzymes by using them in organic solvents. Nature, 409(6817), 241-246.

Krishnamachari, V., Levine, L. H., Zhou, C., & Paré, P. W. (2004). In vitro flavon-3-ol oxidation mediated by a B ring hydroxylation pattern. Chemical Research in Toxicology, 17(6), 795-804.

Kudanga, T., Nyanhongo, G. S., Guebitz, G. M., & Burton, S. (2011). Potential applications of laccase-mediated coupling and grafting reactions: A review. Enzyme and Microbial Technology, 48(3), 195-208.

Kunamneni, A., Camarero, S., García-Burgos, C., Plou, F. J., Ballesteros, A., & Alcalde, M. (2008). Engineering and Applications of fungal laccases for organic synthesis. Microbial Cell Factories, 7.

Lange, H., Decina, S., & Crestini, C. (2013). Oxidative upgrade of lignin - Recent routes reviewed. European Polymer Journal, 49(6), 1151-1173.

Latté, K. P., Kaloga, M., Schäfer, A., & Kolodziej, H. (2008). An ellagitannin, n-butyl gallate, two aryltetralin lignans, and an unprecedented diterpene ester from Pelargonium reniforme. Phytochemistry, 69(3), 820-826.

Laufer, Z., Beckett, R. P., & Minibayeva, F. V. (2006). Co-occurrence of the multicopper oxidases tyrosinase and laccase in lichens in sub-order peltigerineae. Annals of Botany, 98(5), 1035-1042.

Lee, S. J., Kim, J. C., Kim, M. J., Kitaoka, M., Park, C. S., Lee, S. Y., Ra, M. J., Moon, T. W., Robyt, J. F., & Park, K. H. (1999). Transglycosylation of naringin by Bacillus stearothermophilus maltogenic amylase to give glycosylated naringin. Journal of Agricultural and Food Chemistry, 47(9), 3669-3674.

Lenfant, N., Hotelier, T., Velluet, E., Bourne, Y., Marchot, P., & Chatonnet, A. (2013). ESTHER, the database of the α/β-hydrolase fold superfamily of proteins: Tools to explore diversity of functions. Nucleic Acids Research, 41(D1), D423-D429.

Leonowicz, A., Szklarz, G., & Wojtaś-Wasilewska, M. (1985). The effect of fungal laccase on fractionated lignosulphonates (peritan Na). Phytochemistry, 24(3), 393-396.

Levites, Y., Amit, T., Youdim, M. B. H., & Mandel, S. (2002). Involvement of protein kinase C activation and cell survival/cell cycle genes in green tea polyphenol (-)- epigallocatechin 3-gallate neuroprotective action. Journal of Biological Chemistry, 277(34), 30574-30580.

Li, C., Sun, J., Li, T., Liu, S. Q., & Huang, D. (2014). Chemical and enzymatic synthesis of a library of 2-phenethyl esters and their sensory attributes. Food Chemistry, 154, 205-210.

150

Lombard, V., Golaconda Ramulu, H., Drula, E., Coutinho, P. M., & Henrissat, B. (2013). The carbohydrate-active enzymes database (CAZy) in 2013. Nucleic Acids Research, 42(D1), D490-D495.

Lotito, S. B., Actis-Goretta, L., Renart, M. L., Caligiuri, M., Rein, D., Schmitz, H. H., Steinberg, F. M., Keen, C. L., & Fraga, C. G. (2000). Influence of oligomer chain length on the antioxidant activity of procyanidins. Biochemical and Biophysical Research Communications, 276(3), 945-951.

Lucas, R., Comelles, F., Alcántara, D., Maldonado, O. S., Curcuroze, M., Parra, J. L., & Morales, J. C. (2010). Surface-active properties of lipophilic antioxidants tyrosol and hydroxytyrosol fatty acid esters: A potential explanation for the nonlinear hypothesis of the antioxidant activity in oil-in-water emulsions. Journal of Agricultural and Food Chemistry, 58(13), 8021-8026.

Ma, Y. L., Li, Q. M., Van Den Heuvel, H., & Claeys, M. (1997). Characterization of flavone and flavonol aglycones by collision-induced dissociation tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 11(12), 1357-1364.

Maarit, L., Kristiina, K., Petri, H., & Jussi, S. (2009). On the reactions of two fungal laceases differing in their redox potential with lignin model compounds: products and their rate of formation. Journal of Agricultural and Food Chemistry, 57(18), 8357-8365.

Machczynski, M. C., Vijgenboom, E., Samyn, B., & Canters, G. W. (2004). Characterization of SLAC: A small laccase from Streptomyces coelicolor with unprecedented activity. Protein Science, 13(9), 2388-2397.

Majumdar, S., Lukk, T., Solbiati, J. O., Bauer, S., Nair, S. K., Cronan, J. E., & Gerlt, J. A. (2014). Roles of small laccases from Streptomyces in lignin degradation. Biochemistry, 53(24), 4047-4058.

Makris, D. P., & Rossiter, J. T. (2002). An investigation on structural aspects influencing product formation in enzymic and chemical oxidation of quercetin and related flavonols. Food Chemistry, 77(2), 177-185.

Manach, C., Scalbert, A., Morand, C., Rémésy, C., & Jiménez, L. (2004). Polyphenols: Food sources and bioavailability. American Journal of Clinical Nutrition, 79(5), 727-747.

Marcotullio, M. C., Pelosi, A., & Curini, M. (2014). Hinokinin, an emerging bioactive lignan. Molecules, 19(9), 14862-14878.

Markham, K. R. (1989). Flavones, Flavonols and their Glycosides. In J. B. Harborne (Ed.). Methods in plant biochemistry: Volume 1 Plant Phenolics (pp. 197-201). London: Academic Press.

151

Marshall, L. F., Camp, P. E., & Bowers, S. A. (1984). Dimethyl sulfoxide for the treatment of intracranial hypertension: A preliminary trial. Neurosurgery, 14(6), 659-663.

Marshall, N. M., Garner, D. K., Wilson, T. D., Gao, Y., Robinson, H., Nilges, M. J., & Lu, Y. (2009). Rationally tuning the reduction potential of a single cupredoxin beyond the natural range. Nature, 462(7269), 113-116.

Meotti, F. C., Senthilmohan, R., Harwood, D. T., Missau, F. C., Pizzolatti, M. G., & Kettle, A. J. (2008). Myricitrin as a substrate and inhibitor of myeloperoxidase: Implications for the pharmacological effects of flavonoids. Free Radical Biology and Medicine, 44(1), 109-120.

Messerli, S. M., Ahn, M. R., Kunimasa, K., Yanagihara, M., Tatefuji, T., Hashimoto, K., Mautner, V., Uto, Y., Hori, H., Kumazawa, S., Kaji, K., Ohta, T., & Maruta, H. (2009). Artepillin C (ARC) in Brazilian green propolis selectively blocks oncogenic PAK1 signaling and suppresses the growth of NF tumors in mice. Phytotherapy Research, 23(3), 423-427.

Miura, Y., Chiba, T., Tomita, I., Koizumi, H., Miura, S., Umegaki, K., Hara, Y., Ikeda, M., & Tomita, T. (2001). Tea catechins prevent the development of atherosclerosis in apoprotein E-deficient mice. Journal of Nutrition, 131(1), 27-32.

Mohammadian, M., Fathi-Roudsari, M., Mollania, N., Badoei-Dalfard, A., & Khajeh, K. (2010). Enhanced expression of a recombinant bacterial laccase at low temperature and microaerobic conditions: Purification and biochemical characterization. Journal of Industrial Microbiology and Biotechnology, 37(8), 863-869.

Mori, T. A. (2014). Omega-3 fatty acids and cardiovascular disease: Epidemiology and effects on cardiometabolic risk factors. Food and Function, 5(9), 2004-2019.

Morris, G. M., Goodsell, D. S., Halliday, R. S., Huey, R., Hart, W. E., Belew, R. K., & Olson, A. J. (1998). Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. Journal of Computational Chemistry, 19(14), 1639-1662.

Muther, R. S., & Bennett, W. M. (1980). Effects of dimethyl sulfoxide on renal function in man. Journal of the American Medical Association, 244(18), 2081-2083.

Natella, F., Nardini, M., Di Felice, M., & Scaccini, C. (1999). Benzoic and cinnamic acid derivatives as antioxidants: Structure- activity relation. Journal of Agricultural and Food Chemistry, 47(4), 1453-1459.

NC-IUBMB (2010a). The Enzyme List. Class 1 - Oxidoreductases. Nomenclature Committee of the International Union of Biochemistry and Molecular Biology. Generated from the ExplorEnz database, September 2010. (pp. 196).

NC-IUBMB (2010b). The Enzyme List. Class 3 - Hydrolases. Nomenclature Committee of the International Union of Biochemistry and Molecular Biology. Generated from the ExplorEnz database, September 2010. (pp. 3-4).

152

Neacsu, M., Vaughan, N., Raikos, V., Multari, S., Duncan, G. J., Duthie, G. G., & Russell, W. R. (2015). Phytochemical profile of commercially available food plant powders: Their potential role in healthier food reformulations. Food Chemistry, 179, 159-169.

Nicolosi, G., Piattelli, M., & Sanfilippo, C. (1993). Lipase-catalyzed regioselective protection of hydroxyl groups in aromatic dihydroxyaldehydes and ketones. Tetrahedron, 49(15), 3143-3148.

Nicotra, S., Cramarossa, M. R., Mucci, A., Pagnoni, U. M., Riva, S., & Forti, L. (2004). Biotransformation of resveratrol: Synthesis of trans-dehydrodimers catalyzed by laccases from Myceliophtora thermophyla and from Trametes pubescens. Tetrahedron, 60(3), 595-600.

Nieter, A., Haase-Aschoff, P., Linke, D., Nimtz, M., & Berger, R. G. (2014). A halotolerant type A feruloyl esterase from Pleurotus eryngii. Fungal Biology, 118(3), 348-357.

Nishimura, H., Nonaka, G. I., & Nishioka, I. (1984). Seven quinic acid gallates from Quercus stenophylla. Phytochemistry, 23(11), 2621-2623.

Nugroho Prasetyo, E., Nyanhongo, G. S., & Guebitz, G. M. Enzymatically enriching naringenin with hydroxylated and/or methoxylated phenolic compounds. Process Biochemistry, 46(4), 1019-1024.

O'May, C., & Tufenkji, N. (2011). The swarming motility of Pseudomonas aeruginosa is blocked by cranberry proanthocyanidins and other tannin-containing materials. Applied and Environmental Microbiology, 77(9), 3061-3067.

Ollis, D. L., Cheah, E., Cygler, M., Dijkstra, B., Frolow, F., Franken, S. M., Harel, M., Remington, S. J., Silman, I., Schrag, J., Sussman, J. L., Verschueren, K. H. G., & Goldman, A. (1992). The α/β hydrolase fold. Protein Engineering, 5(3), 197-211.

Onoue, S., Takahashi, H., Kawabata, Y., Seto, Y., Hatanaka, J., Timmermann, B., & Yamada, S. (2010). Formulation design and photochemical studies on nanocrystal solid dispersion of curcumin with improved oral bioavailability. Journal of Pharmaceutical Sciences, 99(4), 1871-1881.

Ozaki, T., Mishima, S., Nishiyama, M., & Kuzuyama, T. (2009). NovQ is a prenyltransferase capable of catalyzing the addition of a dimethylallyl group to both phenylpropanoids and flavonoids. Journal of Antibiotics, 62(7), 385-392.

Panya, A., Laguerre, M., Bayrasy, C., Lecomte, J., Villeneuve, P., McClements, D. J., & Decker, E. A. (2012). An investigation of the versatile antioxidant mechanisms of action of rosmarinate alkyl esters in oil-in-water emulsions. Journal of Agricultural and Food Chemistry, 60(10), 2692-2700.

153

Paulino, N., Abreu, S. R. L., Uto, Y., Koyama, D., Nagasawa, H., Hori, H., Dirsch, V. M., Vollmar, A. M., Scremin, A., & Bretz, W. A. (2008). Anti-inflammatory effects of a bioavailable compound, Artepillin C, in Brazilian propolis. European Journal of Pharmacology, 587(1-3), 296-301.

Park, Y. J., Yoon, S. J., & Lee, H. B. (2008). A novel thermostable arylesterase from the archaeon Sulfolobus solfataricus P1: Purification, characterization, and expression. Journal of Bacteriology, 190(24), 8086-8095.

Perron, N. R., Hodges, J. N., Jenkins, M., & Brumaghim, J. L. (2008). Predicting how polyphenol antioxidants prevent DNA damage by binding to iron. Inorganic Chemistry, 47(14), 6153-6161.

Peyrat-Maillard, M. N., Cuvelier, M. E., & Berset, C. (2003). Antioxidant Activity of Phenolic Compounds in 2,2'-Azobis (2-amidinopropane) Dihydrochloride (AAPH)-Induced Oxidation: Synergistic and Antagonistic Effects. JAOCS, Journal of the American Oil Chemists' Society, 80(10), 1007-1012.

Pikal-Cleland, K. A., & Carpenter, J. F. (2001). Lyophilization-induced protein denaturation in phosphate buffer systems: Monomeric and tetrameric β-galactosidase. Journal of Pharmaceutical Sciences, 90(9), 1255-1268.

Piontek, K., Antorini, M., & Choinowski, T. (2002). Crystal structure of a laccase from the fungus Trametes versicolor at 1.90-Å resolution containing a full complement of coppers. Journal of Biological Chemistry, 277(40), 37663-37669.

Plumb, G. W., De Pascual-Teresa, S., Santos-Buelga, C., Cheynier, V., & Williamson, G. (1998). Antioxidant properties of catechins and proanthocyanidins: Effect of polymerisation, galloylation and glycosylation. Free Radical Research, 29(4), 351-358.

Ponzoni, C., Beneventi, E., Cramarossa, M. R., Raimondi, S., Trevisi, G., Pagnoni, U. M., Riva, S., & Forti, L. (2007). Laccase-catalyzed dimerization of hydroxystilbenes. Advanced Synthesis and Catalysis, 349(8-9), 1497-1506.

Porter, W. L., Black, E. D., & Drolet, A. M. (1989). Use of polyamide oxidative fluorescence test on lipid emulsions: Contrast in relative effectiveness of antioxidants in bulk versus dispersed systems. Journal of Agricultural and Food Chemistry, 37(3), 615-624.

Pourcel, L., Routaboul, J. -., Kerhoas, L., Caboche, M., Lepiniec, L., & Debeaujon, I. (2005). TRANSPARENT TESTA10 encodes a laccase-like enzyme involved in oxidative polymerization of flavonoids in Arabidopsis seed coat. The Plant Cell, 17(11), 2966-2980.

Prior, R. L., Wu, X., & Schaich, K. (2005). Standardized methods for the determination of antioxidant capacity and phenolics in foods and dietary supplements. Journal of Agricultural and Food Chemistry, 53(10), 4290-4302.

154

Quideau, S., Deffieux, D., Douat-Casassus, C., & Pouységu, L. (2011). Plant polyphenols: Chemical properties, biological activities, and synthesis. Angewandte Chemie - International Edition, 50(3), 586-621.

Regoli, F., & Winston, G. W. (1999). Quantification of total oxidant scavenging capacity of antioxidants for peroxynitrite, peroxyl radicals, and hydroxyl radicals. Toxicology and Applied Pharmacology, 156(2), 96-105.

Rice-evans, C. A., Miller, N. J., Bolwell, P. G., Bramley, P. M., & Pridham, J. B. (1995). The relative antioxidant activities of plant-derived polyphenolic flavonoids. Free Radical Research, 22(4), 375-383.

Ryu, K., & Dordick, J. S. (1989). Free energy relationships of substrate and solvent hydrophobicities with enzymatic catalysis in organic media. Journal of the American Chemical Society, 111(20), 8026-8027.

Saake, B., Argyropoulos, D. S., Beinhoff, O., & Faix, O. (1996). A comparison of lignin polymer models (DHPs) and lignins by 31P NMR spectroscopy. Phytochemistry, 43(2), 499-507.

Sabally, K., Karboune, S., Louis, R., & Kermasha, S. (2006). Lipase-catalyzed transesterification of dihydrocaffeic acid with flaxseed oil for the synthesis of phenolic lipids. Journal of Biotechnology, 127(1), 167-176.

Santhanam, N., Vivanco, J. M., Decker, S. R., & Reardon, K. F. (2011). Expression of industrially relevant laccases: Prokaryotic style. Trends in Biotechnology, 29(10), 480-489.

Santos, N. C., Figueira-Coelho, J., Martins-Silva, J., & Saldanha, C. (2003). Multidisciplinary utilization of dimethyl sulfoxide: Pharmacological, cellular, and molecular aspects. Biochemical Pharmacology, 65(7), 1035-1041.

Scalbert, A., Manach, C., Morand, C., Rémésy, C., & Jiménez, L. (2005). Dietary polyphenols and the prevention of diseases. Critical Reviews in Food Science and Nutrition, 45(4), 287-306.

Schreier, P., & Miller, E. (1985). Studies on flavonol degradation by peroxidase (donor: H2O2-oxidoreductase, EC 1.11.1.7): Part 2-Quercetin. Food Chemistry, 18(4), 301-317.

Selloum, L., Reichl, S., Müller, M., Sebihi, L., & Arnhold, J. (2001). Effects of flavonols on the generation of superoxide anion radicals by xanthine oxidase and stimulated neutrophils. Archives of Biochemistry and Biophysics, 395(1), 49-56.

Shahidi, F., & Zhong, Y. (2011). Revisiting the polar paradox theory: A critical overview. Journal of Agricultural and Food Chemistry, 59(8), 3499-3504.

Sharma, P., Goel, R., & Capalash, N. (2007). Bacterial laccases. World Journal of Microbiology and Biotechnology, 23(6), 823-832.

155

Sharma, S., & Kanwar, S. S. (2014). Organic solvent tolerant lipases and applications. The Scientific World Journal, 2014.

Shaw, J. F., Chang, R. C., Chuang, K. H., Yen, Y. T., Wang, Y. J., & Wang, F. G. (1994). Nucleotide sequence of a novel arylesterase gene from Vibrio mimicus and characterization of the enzyme expressed in Escherichia coli. Biochemical Journal, 298(3), 675- 680.

Skálová, T., Dohnálek, J., Østergaard, L. H., Østergaard, P. R., Kolenko, P., Dušková, J., Štěpánková, A., & Hašek, J. (2009). The Structure of the Small Laccase from Streptomyces coelicolor Reveals a Link between Laccases and Nitrite Reductases. Journal of Molecular Biology, 385(4), 1165-1178.

Solomon, E. I., Sundaram, U. M., & Machonkin, T. E. (1996). Multicopper oxidases and oxygenases. Chemical Reviews, 96(7), 2563-2605.

Stevanovic, T., Diouf, P.N., & Garcia-Perez, M.E. (2009). Bioactive polyphenols from healthy diets and forest biomass. Current Nutrition and Food Science, 5(4), 264-295.

Sun, C. M., Syu, W. J., Huang, Y. T., Chen, C. C., & Ou, J. C. (1997). Selective cytotoxicity of ginkgetin from Selaginella moellendorffii. Journal of Natural Products, 60(4), 382-384.

Suzuki, H., Nakayama, T., Yonekura-Sakakibara, K., Fukui, Y., Nakamura, N., Yamaguchi, M. A., Tanaka, Y., Kusumi, T., & Nishino, T. (2002). cDNA cloning, heterologous expressions, and functional characterization of malonyl-coenzyme A:anthocyanidin 3-O- glucoside-6′′-O-malonyltransferase from dahlia flowers. Plant Physiology, 130(4), 2142-2151.

Tahir, H. E., Xiaobo, Z., Zhihua, L., & Yaodi, Z. (2015). Comprehensive Evaluation of Antioxidant Properties and Volatile Compounds of Sudanese Honeys. Journal of Food Biochemistry. Advance online publication. doi:10.1111/jfbc.12135.

Thapa, A., Woo, E. R., Chi, E. Y., Sharoar, M. G., Jin, H. G., Shin, S. Y., & Park, I. S. (2011). Biflavonoids are superior to monoflavonoids in inhibiting amyloid-β toxicity and fibrillogenesis via accumulation of nontoxic oligomer-like structures. Biochemistry, 50(13), 2445-2455.

Topakas, E., Stamatis, H., Mastihubova, M., Biely, P., Kekos, D., Macris, B. J., & Christakopoulos, P. (2003). Purification and characterization of a Fusarium oxysporum feruloyl esterase (FoFAE-I) catalysing transesterification of phenolic acid esters. Enzyme and Microbial Technology, 33(5), 729-737.

Topakas, E., Vafiadi, C., & Christakopoulos, P. (2007). Microbial production, characterization and applications of feruloyl esterases. Process Biochemistry, 42(4), 497-509.

156

Torres, P., Poveda, A., Jimenez-Barbero, J., Ballesteros, A., & Plou, F. J. (2010). Regioselective lipase-catalyzed synthesis of 3-O-acyl derivatives of resveratrol and study of their antioxidant properties. Journal of Agricultural and Food Chemistry, 58(2), 807-813.

Tram, N. L., Hazama, C., Shimoyamada, M., Ando, H., Kato, K., & Yamauchi, R. (2005). Antioxidative compounds from the outer scales of onion. Journal of Agricultural and Food Chemistry, 53(21), 8183-8189.

US Food and Drug Administration. (2004). FDA announces qualified health claims for omega-3 fatty acids [Press Release]. Retrieved from http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/2004/ucm108351.htm

Uzan, E., Portet, B., Lubrano, C., Milesi, S., Favel, A., Lesage-Meessen, L., & Lomascolo, A. (2011). Pycnoporus laccase-mediated bioconversion of rutin to oligomers suitable for biotechnology applications. Applied Microbiology and Biotechnology, 90(1), 97- 105.

Vafiadi, C., Topakas, E., Alissandratos, A., Faulds, C. B., & Christakopoulos, P. (2008). Enzymatic synthesis of butyl hydroxycinnamates and their inhibitory effects on LDL-oxidation. Journal of Biotechnology, 133(4), 497-504.

Valkonen, M., & Kuusi, T. (1997). Spectrophotometric assay for total peroxyl radical- trapping antioxidant potential in human serum. Journal of Lipid Research, 38(4), 823-833.

Van Acker, S. A. B. E., Van Den Berg, D., Tromp, M. N. J. L., Griffioen, D. H., Van Bennekom, W. P., Van Der Vijgh, W. J. F., & Bast, A. (1996). Structural aspects of antioxidant activity of flavonoids. Free Radical Biology and Medicine, 20(3), 331-342.

Van het Hof, K. H., De Boer, H. S. M., Wiseman, S. A., Lien, N., Weststrate, J. A., & Tijburg, L. B. M. (1997). Consumption of green or black tea does not increase resistance of low- density lipoprotein to oxidation in humans. American Journal of Clinical Nutrition, 66(5), 1125- 1132.

Van Pouderoyen, G., Eggert, T., Jaeger, K. E., & Dijkstra, B. W. (2001). The crystal structure of Bacillus subtilis lipase: A minimal α/β hydrolase fold enzyme. Journal of Molecular Biology, 309(1), 215-226.

Van Sumere, C. F. (1989). Phenols and phenolic acids. In J. B. Harborne (Ed.). Methods in plant biochemistry: Volume 1 Plant Phenolics (pp. 29-73). London: Academic Press.

Verger, R. (1997). 'Interfacial activation' of lipases: Facts and artifacts. Trends in Biotechnology, 15(1), 32-38.

Veverka, M., Gallovic, J., Svajdlenka, E., Veverkova, E., Pronayova, N., Milackova, I., & Stefek, M. Novel quercetin derivatives: Synthesis and screening for anti-oxidant activity and aldose reductase inhibition. Chemical Papers, 67(1), 76-83.

157

Viji, P., Binsi, P. K., Visnuvinayagam, S., Bindu, J., Ravishankar, C. N., & Srinivasa Gopal, T. K. (2015). Efficacy of mint (Mentha arvensis) leaf and citrus (Citrus aurantium) peel extracts as natural preservatives for shelf life extension of chill stored Indian mackerel. Journal of Food Science and Technology. Advance online publication. doi: 10.1007/s13197-015-1788-1

Wang, L. (2009). Discovery and characterization of microbial esterases for fibre modification. Unpublished Master's Thesis, University of Toronto, Toronto, ON.

Wang, L. F., & Zhang, H. Y. (2005). A theoretical study of the different radical- scavenging activities of catechin, quercetin, and a rationally designed planar catechin. Bioorganic Chemistry, 33(2), 108-115.

Wangikar, P. P., Michels, P. C., Clark, D. S., & Dordick, J. S. (1997). Structure and function of subtilisin BPN' solubilized in organic solvents. Journal of the American Chemical Society, 119(1), 70-76.

Wescott, C. R., & Klibanov, A. M. (1993). Solvent variation inverts substrate specificity of an enzyme. Journal of the American Chemical Society, 115(5), 1629-1631.

Wescott, C. R., & Klibanov, A. M. (1994). The solvent dependence of enzyme specificity. Biochimica et Biophysica Acta (BBA)/Protein Structure and Molecular, 1206(1), 1- 9.

Williams, S. J., Thomas, C. J., Boujaoude, M., Gannon, C. T., Zanatta, S. D., Jarrott, B., May, C. N., & Woodman, O. L. (2011). Water soluble flavonol prodrugs that protect against ischaemia-reperfusion injury in rat hindlimb and sheep heart. MedChemComm, 2(4), 321-324.

Williamson, G., Kroon, P. A., & Faulds, C. B. (1998). Hairy plant polysaccharides: A close shave with microbial esterases. Microbiology, 144(8), 2011-2023.

Wright, J. S., Johnson, E. R., & DiLabio, G. A. (2001). Predicting the activity of phenolic antioxidants: Theoretical method, analysis of substituent effects, and application to major families of antioxidants. Journal of the American Chemical Society, 123(6), 1173-1183.

Wu, C., Liu, X., Li, Y., Du, X., Wang, X., & Xu, P. (2014). Lipase-nanoporous gold biocomposite modified electrode for reliable detection of triglycerides. Biosensors and Bioelectronics, 53, 26-30.

Xu, F. (1996). Oxidation of phenols, anilines, and benzenethiols by fungal laccases: Correlation between activity and redox potentials as well as halide inhibition. Biochemistry, 35(23), 7608-7614.

Xu, F. (1997). Effects of redox potential and hydroxide inhibition on the pH activity profile of fungal laccases. Journal of Biological Chemistry, 272(2), 924-928.

158

Xu, F., Palmer, A. E., Yaver, D. S., Berka, R. M., Gambetta, G. A., Brown, S. H., & Solomon, E. I. (1999). Targeted mutations in a Trametes villosa laccase: Axial perturbations of the T1 copper. Journal of Biological Chemistry, 274(18), 12372-12375.

Xue, Y., Di, J., Luo, Y., Cheng, K., Wei, X., & Shi, Z. (2014). Resveratrol oligomers for the prevention and treatment of cancers. Oxidative Medicine and Cellular Longevity, 2014, Article ID 765832, 9 pages. doi:10.1155/2014/765832.

Yang, Z., & Robb, D. A. (1991). Enzyme activity in organic media. Biochemical Society Transactions, 20(1), 13S.

Yang, Z., & Robb, D. A. (1994). Partition coefficients of substrates and products and solvent selection for biocatalysis under nearly anhydrous conditions. Biotechnology and Bioengineering, 43(5), 365-370.

Yu, X. W., Zhu, S. S., Xiao, R., & Xu, Y. (2014). Conversion of a Rhizopus chinensis lipase into an esterase by lid swapping. Journal of Lipid Research, 55(6), 1044-1051.

Zaks, A., & Klibanov, A. M. (1988). The effect of water on enzyme action in organic media. Journal of Biological Chemistry, 263(17), 8017-8021.

Zeller, E. A. (1956). Enzymology of the refractory media of the eye. IV. Direct photometric determination of cholinesterase and aliesterase of corpus vitreum. Archives of Biochemistry and Biophysics, 61(1), 231-240.

Zenkevich, I. G., Eshchenko, A. Y., Makarova, S. V., Vitenberg, A. G., Dobryakov, Y. G., & Utsal, V. A. (2007). Identification of the products of oxidation of quercetin by air oxygen at ambient temperature. Molecules, 12(3), 654-672.

Zhang, L., Kong, Y., Wu, D., Zhang, H., Wu, J., Chen, J., Ding, J., Hu, L., Jiang, H., & Shen, X. (2008). Three flavonoids targeting the β-hydroxyacyl-acyl carrier protein dehydratase from Helicobacter pylori: Crystal structure characterization with enzymatic inhibition assay. Protein Science, 17(11), 1971-1978.

Zhang, Y. X., Li, Q. Y., Yan, L. L., & Shi, Y. (2011). Structural characterization and identification of biflavones in Selaginella tamariscina by liquid chromatography-diode-array detection/electrospray ionization tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 25(15), 2173-2186.

Zhang, X., Xiong, Y. L., Chen, J., & Zhou, L. (2014). Synergy of licorice extract and pea protein hydrolysate for oxidative stability of soybean oil-in-water emulsions. Journal of Agricultural and Food Chemistry, 62(32), 8204-8213.

Zhou, A., & Sadik, O. A. (2008). Comparative analysis of quercetin oxidation by electrochemical, enzymatic, autoxidation, and free radical generation techniques: A mechanistic study. Journal of Agricultural and Food Chemistry, 56(24), 12081-12091.

159

APPENDIX 1. SUPPLEMENTAL INFORMATION FOR CHAPTER 3

Table A1.1. Sequences of primers used for construction of esterase AF1753 point mutants. Residue & Predicted Primers mutation function of mutated residue L31A alcohol binding GTCTTGTTCATGGTGCGGGGGAGCACTCTGG, pocket CCAGAGTGCTCCCCCGCACCATGAACAAGAC R37A acyl binding GGGGAGCACTCTGGAGCATACGAGCACGTTGC, pocket GCAACGTGCTCGTATGCTCCAGAGTGCTCCCC Y72A alcohol binding GGGGGCATGCTGAAGCTCAGCAGTTAATGGATG, pocket CATCCATTAACTGCTGAGCTTCAGCATGCCCCC S98A catalytic residue GATACTCTACGGCCACGCCATGGGCGGGAATCTC, GAGATTCCCGCCCATGGCGTGGCCGTAGAGTATC M99A alcohol binding CTCTACGGCCACAGCGCGGGCGGGAATCTCGC, pocket GCGAGATTCCCGCCCGCGCTGTGGCCGTAGAG F124A alcohol binding GGTATAATTTCCGCTCCCGCCCTTGCTCTGCCAAAG, pocket CTTTGGCAGAGCAAGGGCGGGAGCGGAAATTATACC K129A alcohol binding CTTCCTTGCTCTGCCAGCGGAGCTACCAAAACAC, pocket GTGTTTTGGTAGCTCCGCTGGCAGAGCAAGGAAG L189A alcohol binding GATTCATTCTGCAATCTGCTGAAGCGGGAAAATGG, pocket CCATTTTCCCGCTTCAGCAGATTGCAGAATGAATC E190A alcohol binding CATTCTGCAATCTCTTGCAGCGGGAAAATGGGCG, pocket CGCCCATTTTCCCGCTGCAAGAGATTGCAGAATG D214A catalytic residue GATACACGGAACTGCTGCCCAAATAACCTCCTAC, GTAGGAGGTTATTTGGGCAGCAGTTCCGTGTATC H243A catalytic residue CTACGAAGGCTTTTATGCCGAGCCCCACAACGAG, CTCGTTGTGGGGCTCGGCATAAAAGCCTTCGTAG E244A acyl binding GAAGGCTTTTATCACGCGCCCCACAACGAGCCC, pocket GGGCTCGTTGTGGGGCGCGTGATAAAAGCCTTC

160

Leu31Ala Tyr72Ala Ser98Ala Met99Ala Phe124Ala

ladder

P S

F F P S F P F P F

E E E

W W W W W 170 E 130 100 70 55 40 35 25 15

Lys129Ala Leu189Ala Glu190Ala Asp214Ala His243Ala

ladder

P S

P E P S P S P S F

F F E

F 170 W W F W W W 130 100 70 55 40 35 25 15 10

Glu244Ala Arg37Ala

ladder

S P

F E S P

E 170 170 130 130 100 100 70 70 55 55 40 40 35 35 25 25

15 15

10 10

161

Figure A1.1. Protein purification of AF1753 mutants. P, S, F, W, E, stand for pellet after sonication of cells, supernatant after sonication of cells, flowthrough after binding of supernatant to nickel resin, wash of resin, and elution after wash of resin, respectively. Molecular weights are indicated to the left of ladders. Arrow shows band corresponding to molecular weight of AF1753 (30 kD).

162

APPENDIX 2. SUPPLEMENTAL INFORMATION FOR CHAPTER 4

Table A2.1. Sequences of primers used for construction of laccase SCO6712 point mutants. Residue & Predicted Primers mutation function of mutated residue H231A type 1 Cu binding CACGGGGAGTACTACGCCACCTTCCACATGCAC, GTGCATGTGGAAGGTGGCGTAGTACTCCCCGTG C288A type 1 Cu binding GCGTGGATGTACCACGCCCACGTCCAGAGCCAC, GTGGCTCTGGACGTGGGCGTGGTACATCCACGC H293A type 1 Cu binding CTGCCACGTCCAGAGCGCCTCCGACATGGGCATG, CATGCCCATGTCGGAGGCGCTCTGGACGTGGCAG M298A fine-tuning type 1 CACTCCGACATGGGCGCGGTGGGGCTGTTCCTG, Cu redox potential CAGGAACAGCCCCACCGCGCCCATGTCGGAGTG H104A type 3 Cu binding GCCAGCCTGCACGTGGCCGGCCTGGACTACGAG, CTCGTAGTCCAGGCCGGCCACGTGCAGGCTGGC H156A type 3 Cu binding GGCTACTGGCACTACGCCGACCACGTCGTCGCC, GGCGACGACGTGGTCGGCGTAGTGCCAGTAGCC H289A type 3 Cu binding GTGGATGTACCACTGCGCCGTCCAGAGCCACTCC, GGAGTGGCTCTGGACGGCGCAGTGGTACATCCAC H158A type 3 Cu binding CTGGCACTACCACGACGCCGTCGTCGGCACCGAAC, GTTCGGTGCCGACGACGGCGTCGTGGTAGTGCCAG H287A type 3 Cu binding GGGGCGTGGATGTACGCCTGCCACGTCCAGAGC, GCTCTGGACGTGGCAGGCGTACATCCACGCCCC H236A type 3 Cu binding CACACCTTCCACATGGCCGGTCACCGCTGGGCG, CGCCCAGCGGTGACCGGCCATGTGGAAGGTGTG H102A type 2 Cu binding CGTGCGGGCCAGCCTGGCCGTGCACGGCCTGGAC, GTCCAGGCCGTGCACGGCCAGGCTGGCCCGCACG H234A type 2 Cu binding GTACTACCACACCTTCGCCATGCACGGTCACCG, GCGGTGACCGTGCATGGCGAAGGTGTGGTAGTAC M198A Substrate binding GATCGTCTTCAACGACGCGACCATCAACAACCG, GCGGTTGTTGATGGTCGCGTCGTTGAAGACGATC E228A Substrate binding CATGATCACGCACGGGGCGTACTACCACACCTTC, GAAGGTGTGGTAGTACGCCCCGTGCGTGATCATG Y229A Substrate binding GATCACGCACGGGGAGGCCTACCACACCTTCCAC, GTGGAAGGTGTGGTAGGCCTCCCCGTGCGTGATC Y230A Substrate binding CACGCACGGGGAGTACGCCCACACCTTCCACATG, CATGTGGAAGGTGTGGGCGTACTCCCCGTGCGTG S292A Substrate binding CACGTCCAGAGCCACGCCGACATGGGCATGGTG, CACCATGCCCATGTCGGCGTGGCTCTGGACGTG

163

APPENDIX 3. SUPPLEMENTAL INFORMATION FOR CHAPTER 5

Table A3.1. HPLC-MS data for kaempferol and product produced after enzymatic reaction. Mass spectra collected in positive ion mode. Compound m/z Retention Lost atoms from the monomers to form time (min) product (based on observed m/z value)a kaempferol 287.056 12.75 -- kaempferol product 271.060 12.61 1 O

Table A3.2. Fragment ions observed for quercetin dimer #1 after HPLC-MS/MS. Tandem mass spectra collected in positive ion mode using 35 eV collision energy. Fragment ions used for structure elucidation in bold font. m/z Relative m/z Relative m/z Relative intensity intensity intensity

(%) (%) (%) 603.0698 parent ion 423.0662 30.57 259.0206 1.31 586.0659 0.66 421.0506 0.71 249.0367 0.55 585.0605 4.20 407.0384 0.50 241.1014 5.28 557.0632 2.01 405.0546 1.55 231.0266 0.57 543.0510 3.19 363.0487 0.39 194.1722 0.30 539.0544 0.41 348.0391 11.27 172.0251 0.33 529.0724 0.58 347.0356 100.00 166.5065 0.31 525.0418 0.45 342.0289 0.68 163.8658 0.35 515.0562 3.43 341.0250 4.51 163.0372 0.49 511.0609 0.74 330.0281 1.80 154.0194 2.08

488.0647 0.49 329.0254 15.47 153.0165 37.31 487.0581 2.99 327.1378 1.36 138.0258 0.36 473.0484 0.39 323.0516 0.40 137.0217 13.74 467.0572 0.74 305.0262 2.99 113.0586 0.40 455.5465 0.31 303.0468 4.24 109.1997 0.26 452.0640 5.07 302.0377 0.72 93.3092 0.27 451.0606 27.86 301.0309 0.52 91.9900 0.48

450.0510 1.13 298.0416 0.46 89.0594 0.40 445.0451 0.35 297.0366 2.22 69.7212 0.26 433.0497 1.44 285.0356 1.65 63.7090 0.28 429.5676 0.38 274.0412 0.43

424.0690 4.07 273.0358 1.80

164

Table A3.3. Fragment ions observed for myricetin dimer #2 after HPLC-MS/MS. Tandem mass spectra collected in positive ion mode using 35 eV collision energy. Fragment ions used for structure elucidation in bold font. m/z Relative m/z Relative m/z Relative intensity intensity intensity (%) (%) (%) 633.0506 parent ion 419.0423 0.36 355.0068 0.99 577.0637 0.35 418.0280 0.26 354.0374 0.26 569.0279 0.20 417.0219 0.79 353.0289 0.34 559.0490 0.86 412.0405 0.67 352.0504 0.49 541.0418 0.25 411.0344 0.86 351.0497 3.19 539.2509 0.21 409.0526 0.41 351.0120 0.74 532.0568 0.24 408.0432 2.81 350.0399 0.22 531.0569 0.82 407.0395 11.11 349.0334 0.25 523.1160 0.21 406.0289 0.37 346.0251 0.20 513.0450 0.75 405.0246 1.13 345.0224 2.41 503.0630 0.65 397.0541 1.13 341.0653 0.25 485.0501 0.67 396.0439 1.72 341.0277 0.98 482.0424 0.24 395.0400 4.78 339.0490 0.34 481.0371 0.90 394.0305 0.35 337.0350 0.77 479.0257 0.29 393.0219 0.85 336.9966 0.60 475.0718 0.27 391.0440 0.33 330.0332 0.37 472.0377 0.19 389.0295 1.64 329.0285 1.84 464.0347 0.66 384.0447 0.26 327.0485 0.42 463.0290 0.36 383.0375 1.23 327.0132 0.81 455.0503 0.21 382.0318 0.37 324.0219 0.22 454.0479 2.89 381.0618 0.24 323.0564 0.62 453.0441 10.61 380.0474 2.97 323.0180 0.75 452.0351 3.42 379.0446 12.29 320.0479 0.25 451.0276 2.78 378.0364 0.26 319.0448 2.53 445.0195 0.86 377.0299 0.82 318.0359 1.43 441.0474 0.20 370.0272 0.30 317.0277 1.04 437.0450 0.24 369.0626 0.91 316.0208 1.33 436.0372 1.49 369.0226 1.75 315.0482 0.36 435.0342 5.93 368.0456 0.27 315.0136 3.39 433.0189 0.37 367.0458 1.04 313.0352 0.35 426.0544 1.62 366.0323 0.39 311.0176 0.35 425.0493 7.95 365.0294 1.14 302.0370 0.66 424.0387 0.86 363.0498 0.27 301.0334 2.21 423.0339 2.21 361.0352 0.22 300.0221 0.29 421.0191 0.73 355.0452 0.71 299.0195 1.34

165

Table A3.3. cont… m/z Relative m/z Relative intensity intensity (%) (%) 297.0857 0.18 89.0604 0.92 295.0596 0.24 79.0217 0.29 290.0399 0.35 68.9981 0.29 289.0337 0.96 52.5173 0.16 288.0250 0.35 287.0185 1.26 285.0376 0.18 283.0228 0.33 273.0393 0.33 272.0263 0.18 271.0241 1.48 261.0380 1.15 260.0296 1.45 259.0233 1.73 255.0290 0.23 247.0235 0.89 245.0455 0.21 243.0286 0.30 239.0165 0.18 235.0226 0.24 233.0442 0.20 231.0285 0.93 219.0301 0.19 196.0327 0.18 195.0287 1.92 193.0129 0.41 182.7762 0.18 182.0164 0.17 181.0129 4.45 171.0284 1.42 155.0222 0.35 154.0343 0.19 154.0215 6.72 153.0181 100.00 137.0231 0.32 127.0394 0.28 120.4652 0.19

166

Table A3.4. Fragment ions observed for non-enzymatically produced quercetin dimer after HPLC-MS/MS. Tandem mass spectra collected in positive ion mode using 35 eV collision energy. Fragment ions used for structure elucidation in bold font. m/z Relative m/z Relative m/z Relative intensity intensity intensity

(%) (%) (%) 603.0698 parent ion 395.0694 0.43 241.1018 16.29 586.0648 1.28 393.0570 0.49 231.0263 0.58 585.0588 5.08 385.1422 0.50 215.0188 0.39 558.0668 0.56 377.0626 0.35 213.0704 1.23 557.0648 2.21 375.0430 0.34 195.0612 1.08 544.0537 1.01 348.0391 10.74 175.0726 0.28

543.0507 4.10 347.0358 100.00 154.0196 1.12 539.0522 1.21 342.0308 0.56 153.0166 25.95 529.0670 1.13 341.0258 4.36 149.0213 0.28 516.0626 0.43 330.0273 0.57 140.6938 0.28 515.0540 2.59 329.0254 8.54 138.0250 0.58 511.0630 0.63 328.9886 0.54 137.0217 12.78 501.0385 0.45 327.1378 6.67 133.0843 0.49 499.2086 1.03 325.0302 0.45 113.0589 1.62 497.0437 0.45 323.0529 1.07 112.9073 0.28 488.0651 0.40 305.0254 2.64 111.0065 0.30 487.0611 2.95 304.0540 0.32 105.0326 0.48

483.0667 0.37 303.0467 2.88 89.0594 2.26 475.0627 0.35 302.0386 0.61 87.0775 0.29 473.0478 0.37 301.0336 0.33 87.0439 0.33 459.0625 0.31 299.1067 0.42 70.1348 0.24 452.0637 4.17 298.0432 0.56 53.1576 0.20 451.0607 22.76 297.0356 1.84 450.0534 1.12 285.0367 1.51 449.0481 0.40 281.0965 0.42 433.0497 0.54 277.0923 1.78 426.4293 0.35 274.0420 0.31 424.0686 4.12 273.0359 1.68

423.0662 24.85 259.0814 0.45 421.0484 0.52 259.0204 0.47 413.1745 2.25 251.0775 0.98 405.0551 1.40 249.0360 0.43 403.0418 0.35 242.1070 0.33

167

Table A3.5. Fragment ions observed for non-enzymatically produced myricetin dimer after HPLC-MS/MS. Tandem mass spectra collected in positive ion mode using 35 eV collision energy. Fragment ions used for structure elucidation in bold font. m/z Relative m/z Relative m/z Relative intensity intensity intensity (%) (%) (%) 633.0506 parent ion 411.0338 0.47 327.0490 0.46 577.0601 0.30 408.0420 2.55 327.0129 0.56 559.0457 0.38 407.0393 11.85 323.0534 0.36 531.0563 0.94 405.0237 2.07 323.0182 0.52 528.6182 0.27 397.0553 1.42 319.0437 3.21 504.6540 0.30 396.0433 1.15 318.0367 1.65 503.0600 0.38 395.0387 4.46 317.0287 1.58 489.0486 0.34 394.0314 0.37 316.0192 1.30 485.0461 0.36 393.0224 0.98 315.0489 0.38 483.0444 0.31 391.0424 0.46 315.0142 3.07 481.0401 0.99 389.0288 1.50 313.0348 0.43 479.0233 0.40 383.0399 1.35 311.0192 0.42 465.0481 0.48 380.0481 3.12 303.0490 0.36 464.0306 0.33 379.0445 13.58 302.0383 0.27 463.0282 0.93 378.0372 0.48 301.0333 2.04 455.0534 0.36 377.0282 0.95 299.0191 0.85 454.0472 2.89 369.0616 0.81 297.0807 0.27 453.0440 9.31 369.0231 1.86 289.0326 1.02 452.9801 0.29 367.0456 1.29 288.0264 0.34 452.0361 4.23 366.0352 0.36 287.0190 1.17 451.0282 3.28 365.0285 1.02 273.0381 0.38 445.0166 0.47 361.0345 0.50 271.0236 1.71 437.0487 0.52 356.0522 0.37 261.0384 0.93 436.0395 1.39 355.0461 0.36 260.0311 1.86 435.0331 6.14 355.0072 1.24 259.0229 1.52 432.4591 0.26 354.3288 0.27 255.0292 0.31 426.0523 1.76 352.0542 0.43 247.0244 0.45 425.0495 7.89 351.0496 2.75 231.0286 1.05 424.0390 1.11 351.0112 0.78 216.3581 0.31 423.0346 1.93 346.0259 0.25 195.0284 2.61 421.0176 0.43 345.0217 2.33 193.0126 0.48 419.0369 0.53 341.0294 0.50 182.0164 0.34 417.0244 1.70 339.0474 0.34 181.0131 4.79 413.0459 0.35 338.0408 0.41 177.1102 0.24 412.0401 0.44 329.0280 2.07 171.0286 2.00

168

Table A3.5. cont... m/z Relative intensity (%) 155.0225 0.50 154.0215 6.85 153.0181 100.00 137.0229 0.51 136.0756 0.43 133.0858 1.42 129.1024 0.98 127.0384 0.29 120.0808 1.03 113.0603 0.35 110.0714 0.29 107.0710 0.23 101.0600 0.25 89.0603 3.67 87.0448 0.50 86.0970 2.71 86.0605 0.40 84.9523 0.23 84.0814 1.38 84.0454 0.42 84.0083 0.21 80.0550 0.22 79.0219 1.01 73.0293 0.56 72.0816 1.48 70.0661 3.01 68.9984 0.25 61.2795 0.23 61.2394 0.21 54.7642 0.18

169

A) 120

100

) 0

UV-Vis UV-Vis UV-Vis UV-Vis UV-Vis UV-Vis UV-Vis

HPLC-MS HPLC-MS HPLC-MS HPLC-MS HPLC-MS HPLC-MS HPLC-MS -enz +enz -enz +enz -enz +enz

% of initial % ( initial 5min 3h 24h 120

B) strate 100

UV-Vis UV-Vis UV-Vis UV-Vis UV-Vis UV-Vis UV-Vis

HPLC-MS HPLC-MS HPLC-MS HPLC-MS HPLC-MS HPLC-MS HPLC-MS -enz +enz -enz +enz -enz +enz initial 5min 3h 24h 120

100 C) Remaining amount of starting phenolic sub phenolic amount of starting Remaining 80

UV-Vis UV-Vis UV-Vis UV-Vis UV-Vis UV-Vis UV-Vis

HPLC-MS HPLC-MS HPLC-MS HPLC-MS HPLC-MS HPLC-MS HPLC-MS -enz +enz -enz +enz -enz +enz initial 5min 3h 24h

170

Figure A3.2. Remaining phenolic monomer in laccase reactions, as measured by UV-Vis spectrophotometry and by intensity of HPLC-MS peak. A) Quercetin, B) morin, and C) myricetin reactions. -enz and +enz indicate reactions without and with enzyme, respectively. n=3; errors indicate standard deviation.

171

APPENDIX 4. TRAP ANTIOXIDANT ASSAY USING LINOLEIC ACID IN PLACE OF

DCFH

The TRAP antioxidant assay can be used with other probes besides 2′,7′- dichlorofluorescin (DCFH). This section includes my attempts at using the TRAP assay for laccase reaction-mix samples, with linoleic acid emulsion as a lipid-based probe. There was large variation between replicates and final results were not reproducible. This was possibly due to inconsistency in the emulsion formation between different samples.

A4.1. Method

TRAP antioxidant assay was carried out in 50 mM pH 7.4 sodium phosphate buffer with 2 mM 2,2′-azobis(2-methylpropionamidine) (AAPH) and 0.0045 % (w/v) linoleic acid at 37 oC (Liegeois et al., 2000). Linoleic acid was emulsified following the method of Bondet et al. (2000). Briefly, 1 volume of 1.35 % (w/v) linoleic acid in methanol (stored in -20 oC) was dried under nitrogen to remove the methanol (leaving behind the linoleic acid) and mixed with 10 volumes of 50 mM PIPES buffer pH 6.4 containing 0.85 % Tween 20 and 154.5 mM NaCl. This mixture was vortexed at 3000 rpm for 2 min and the resulting emulsion was used within 30 min. The antioxidant assay was prepared as follows. In a quartz microtitre plate 60 µL of 200 mM sodium phosphate buffer (pH 7.4) was mixed with 154 µL MilliQ water. To this was added 8 µL of the linoleic acid emulsion and 6 µL of diluted (25 x in MilliQ water) SCO6712 laccase reaction sample with quercetin. The mixture was mixed by pipetting up and down. Before the antioxidant assay was started by addition of 12 µL AAPH (40 mM stock), the microtitre plate was covered with parafilm and pre-incubated at 37 oC for 30 min. Absorbance of oxidized linoleic acid was followed at 234 nm every 1 min for 30 min and then every 30 min up to 300 min. For the first 30 min, the microtitre plate was kept in a platereader at 37 oC without any parafilm cover and absorbance readings were taken continuously every 1 min. Afterwards, the microtitre plate was covered in parafilm and kept in an incubator at 37 oC. When it was time to take absorbance readings the microtitre plate was taken out of the incubator, the parafilm cover was removed, absorbance measured, then the microtitre plate was recovered in parafilm and returned to the 37 oC incubator.

172

A4.2. Challenges

In the TRAP assay with linoleic acid, AAPH radicals oxidize linoleic acid to produce conjugated diene hydroperoxides from the linoleic acid, which absorb at 234 nm. However, AAPH also absorbs at this wavelength and the absorbance increases over time as AAPH breaks down to produce radicals. This makes it necessary to use low concentrations of AAPH in the antioxidant assay (2 mM AAPH in this assay as opposed to 50 mM AAPH in the assay with DCFH as the probe). The lower amount of AAPH means there is a lower rate of radical production, which increases the length of time of the assay (5 h in this assay as opposed to 2 h in the assay with DCFH as the probe). The low wavelength for measurement also requires the use of a quartz microtitre plate. The quartz microtitre plate does not have a quartz lid so parafilm was used to cover the microtitre plate and the plate was transferred to and from the platereader every 30 min for absorbance readings (removing and then returning the parafilm cover each time). Taken together, the long assay time and the required transferring of the microtitre plate to and from the platereader make the assay very time consuming and labour intensive; and not suitable for multiple reaction time points with reaction samples containing and lacking enzyme. Another challenge with this assay was that initial tests with a quercetin antioxidant standard concentration series showed irreproducibility between runs on different days, with up to 50 % difference in values. This may partly be caused by the lack of consistency of the linoleic acid emulsions. The emulsions in the microtitre plate are mixed by pipetting with a multichannel pipettor but this may not give reproducible mixing. In some cases, there was large variability even between replicates run on the same day. To circumvent some of the challenges of this antioxidant assay, the assay might be used with a limited number of samples using purified laccase reaction products, rather than multiple time points of the whole reaction-mix. Additionally, the antioxidant assay could be run in microtubes to allow for vigorous vortexing of the linoleic acid emulsion at the beginning of the assay. In this case, samples may be transferred at required time points to a microtitre plate to measure absorbance, and then returned to the microtubes to continue incubation. This will be easier if fewer samples (such as five concentrations of purified product material) are being analyzed. Nonetheless, the required total assay time of 300 min still makes the assay cumbersome and there may not be a way to avoid this.

173

A4.3. References

Bondet, V., Cuvelier, M., & Berset, C. (2000). Behavior of phenolic antioxidants in a partitioned medium: Focus on linoleic acid peroxidation induced by iron/ascorbic acid system. JAOCS, Journal of the American Oil Chemists' Society, 77(8), 813-818.

Liégeois, C., Lermusieau, G., & Collin, S. (2000). Measuring antioxidant efficiency of wort, malt, and hops against the 2,2'-azobis(2-amidinopropane) dihydrochloride-induced oxidation of an aqueous dispersion of linoleic acid. Journal of Agricultural and Food Chemistry, 48(4), 1129-1134.

174

APPENDIX 5. QUANTIFYING SOLUBILITY OF QUERCETIN MONOMER FOR

COMPARISON TO QUERCETIN DIMER

I attempted to test the solubility of the quercetin monomer to compare with the quercetin dimer once the dimer has been purified. The solubility assay was attempted in aqueous phosphate buffer (pH 7.4) with 0.1 M NaCl and also in 1-octanol to assess aqueous and lipid solubility, respectively. In anticipation of having only small quantities of purified dimer, I tried to establish the solubility test procedures for the monomer at small-scale. There were challenges that arose because of the low aqueous stability of the quercetin monomer.

A5.1. Method In brief, for assessing aqueous solubility quercetin was first dissolved in methanol, the methanol removed under reduced pressure using a speed vacuum, the required volume buffer added to the remaining quercetin solid and the solution mixed, and then the solution was centrifuged before measuring absorbance of the supernatant. The full procedure was as follows. A 6 mM quercetin solution in methanol was used to make a 0.6 mM solution in methanol. The 0.6 mM quercetin in methanol was used to make 0 mM, 0.15 mM, 0.3 mM, and 0.6 mM quercetin in methanol; while the 6 mM quercetin in methanol was used to make 0.9 mM, 1.5 mM, 1.8 mM, 2.25 mM, 3 mM, and 6 mM quercetin in methanol. Then 10 µL of the 0 mM, 0.15 mM, 0.3 mM, 0.6 mM, 0.9 mM, 1.5 mM, 1.8 mM, 2.25 mM, 3 mM, and 6 mM quercetin solutions were transferred to 1.5 mL microtubes in triplicates and speed vacuumed until the methanol evaporated away (10 min). To the quercetin (that remained after speed vacuuming) was added 0.3 mL of 50 mM sodium phosphate buffer with 0.1 M NaCl. The quercetin was manually grinded with a 10 µL pipette tip to dislodge it from the wall of the microtube and disperse it before subsequently autovortexing for 10 min at 3200 rpm (Vortex-Genie 2, Scientific Industries). The solutions were then ultracentrifuged at 14, 000 rpm and 25 oC for 10 min. Then, 200 µL of the supernatant was transferred to a microtitre plate to measure absorbance at 385 nm. After measuring absorbance, the 200 µL supernatant was returned to the 1.5 mL microtube and 60.1 µL DMSO was added to the 1.5 mL microtube (giving a final 16.7 % (v/v) DMSO) to completely solubilize any undissolved quercetin. The solution was briefly vortexed and then 200 µL transferred to a microtitre plate to measure absorbance at 385 nm. In

175 this way, the dissolved quercetin concentration (dissolved in buffer without DMSO) could be plotted against the actual total concentration (total dissolved after adding DMSO). The amount of quercetin was quantified using a quercetin standard curve made by starting with stock 0.2 M quercetin in DMSO and diluting in sodium phosphate buffer (pH 7.4) containing NaCl to get the desired final concentrations of quercetin, 50 mM buffer, 0.1 M NaCl, and 16.7 % (v/v) DMSO. The presence of DMSO was found to not cause change in absorbance characteristics of the quercetin at 385 nm. For assessing lipid solubility, a similar procedure was used as for aqueous solubility. A 10 mM quercetin solution in methanol was used to make a 0.2 mM solution in methanol. The 0.2 mM quercetin in methanol was used to make 0 mM, 0.02 mM, 0.05 mM 0.1 mM, and 0.2 mM quercetin in methanol; while the 10 mM quercetin in methanol was used to make 0.5 mM, 1 mM, 3 mM, 5 mM, and 10 mM quercetin in methanol. Then 0.3 mL of the 0 mM, 0.02 mM, 0.05 mM 0.1 mM, 0.2 mM, 0.5 mM, 1 mM, 3 mM, 5 mM, and 10 mM quercetin solutions were transferred to 1.5 mL microtubes in triplicates and speed vacuumed until the methanol evaporated away (40 min). To the quercetin (that remained after speed vacuuming) was added 0.3 mL of 1-octanol. The quercetin was briefly vortexed by hand to dislodge it from the wall of the microtube and disperse it before subsequently autovortexing for 10 min at 3200 rpm (Vortex-Genie 2, Scientific Industries). The solutions were then ultracentrifuged at 14, 000 rpm and 25 oC for 10 min. Then, 100 µL of the supernatant was transferred to a microtitre plate to measure absorbance at 385 nm (or the supernatant was diluted in 1-octanol and vortexed, as appropriate, before transferring 100 µL for measuring absorbance, to ensure absorbance readings were not too high). As the quercetin had not yet reached saturation at these concentrations in 1-octanol, and higher stock concentrations in methanol were not possible, higher concentrations in 1-octanol were assessed as follows. The required mass quercetin (1.8 mg, 2.7 mg, 4.5 mg, 6.3 mg, and 9.1 mg) was weighed in triplicate and transferred into 1.5 mL microtubes. Then 0.3 mL 1-octanol was added to get total concentrations (dissolved and undissolved) 20 mM, 30 mM, 50 mM, 70 mM, and 100 mM. The solutions were then vortexed and processed as above. The amount of quercetin was quantified using a quercetin standard curve made by starting with stock 0.2 M quercetin in DMSO and diluting in 1-octanol to get the desired final concentrations of quercetin and 5 % (v/v) DMSO. The presence of DMSO was found to not cause change in absorbance characteristics of the quercetin at 385 nm.

176

A5.2. Results and Challenges Aqueous solubility was assessed in 50 mM sodium phosphate buffer (pH 7.4) with 0.1 M NaCl, while lipid solubility was assessed in 1-octanol. As quercetin contains ionizable functional groups, it was important to test aqueous solubility in controlled buffer and electrolyte strength (Dearden & Bresnen, 1988). On the other hand, 1-octanol is a solvent considered as a good model for biological lipids (Sangster, 1989). The assays were attempted at small-scale because of the anticipation that there will be a limited quantity of purified quercetin dimer. Preliminary tests in the used buffer showed that the solubilized portion of quercetin is not stable and depletes over time at 25 oC. After 1 h approximately 80 % of the quercetin is remaining and there is only 20 % quercetin remaining by 22 h. This precluded the use of long incubation times to allow equilibration of the quercetin between soluble and insoluble states. In the end, 10 min was chosen for the autovortexing time and 10 min for the subsequent centrifugation time. It was possible to obtain a curve showing the dissolved amount of quercetin in buffer as a function of the total amount (dissolved and undissolved) present in solution and the results show that quercetin appears to reach saturation at approximately 16 µM concentration under the conditions used (Fig. A5.1A). However, it should be noted that even at the lowest tested concentration not all the quercetin was totally dissolved. At the lowest total concentration (dissolved and undissolved) of 3 µM quercetin, there was approximately 2 µM (66 %) in the dissolved portion. This may be an indication that the time needed to dissolve the maximum soluble amount of quercetin in buffer is longer than the 10 min mixing time that was used. In the case of 1-octanol, the same procedure could not be used for all concentrations of quercetin. Since it is not possible to weight out extremely low masses of quercetin for lower concentrations, quercetin was first dissolved in methanol, diluted, then the methanol evaporated away, as was done for solubility test in buffer. However, for higher concentrations, the solubility limit of quercetin in methanol had been reached so the required masses of quercetin were weighed out and 1-octanol was added directly to these. In 1-octanol, quercetin appeared to achieve saturation at approximately 15 mM concentration under the conditions used (Fig. A5.1B), which is three orders of magnitude higher saturation concentration than in buffer (Fig. A5.1A). Contrasted to the case of buffer, at the lowest tested concentration of quercetin in 1-

177 octanol (20 µM), all the quercetin was dissolved. In fact, even at a total concentration (dissolved and undissolved) of 10 mM quercetin, approximately 9 mM (90 %) was in the dissolved portion.

A) 18 16 14

) 12

µM

( 10 8 6 4

dissolved in buffer dissolved in buffer

quercetin of Concentration 2 0 0 20 40 60 80 100 120 140 Total concentration of quercetin (dissolved and undissolved) (µM)

18 B) 16

) 14

( 12

octanol

- 8

Concentration of quercetin of Concentration

dissolved in 1 2 0 0 20 40 60 80 100 Total concentration of quercetin (dissolved and undissolved) (mM) Figure A5.1. The maximum amount of quercetin that can be dissolved in A) 50 mM sodium phosphate buffer (pH 7.4) with 0.1 M NaCl and B) 1-octanol.

178

If a suitable spectrophotometric absorbance wavelength can be found for the quercetin dimer, then the above methods may be suitable for assessing its solubility. However, quite a lot of material may be needed for determining solubility in 1-octanol if the dimer has equal to or greater lipid solubility than the monomer, as suggested by the increased retention time of the dimer on reversed-phase HPLC compared to the monomer (Chapter 5). In the case of aqueous solubility, the stability of the dimer will first need to be determined in the chosen buffer.

A5.3. References

Dearden, J. C., & Bresnen, G. M. (1988). The measurement of partition coefficients. Quantitative Structure-Activity Relationships, 7(3), 133-144.

Sangster, J. (1989). Octanol-Water Partition Coefficients of Simple Organic Compounds. Journal of Physical and Chemical Reference Data, 18(3), 1111-1229.