Unravelling the Phenolic Composition and in Vitro Activities of Australian Native Mints

Home , Backhousia citriodora, Chrysanthemin, Mentha australis, Mentha diemenica, Mentha satureioides, Prostanthera incisa

Unravelling the Phenolic Composition and In vitro Activities of Australian Native Mints

A thesis in fulfilment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

Kitty Sze Chiu Tang

School of Chemical Engineering

Faculty of Engineering

March 2016

THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: Tang

First name: Kitty Sze Chiu Other name/s:

Abbreviation for degree as given in the University calendar:

School: Chemical Engineering Faculty: Engineering

Title: Unravelling the phenolic composition and in vitro biological activities of Australian native mints.

Abstract

Australia native flora has served as food and medicine for the indigenous people for thousands of years. However, systematic scientific investigation into their health properties is still in the early stage and rather limited in scope. In this thesis, the phenolic composition and in vitro biological activities of two native Australian mints – river mint (Mentha australis R. Br.) and mint bush (Prostanthera rotundifolia R. Br.) were systematically investigated for the first time. The phenolic compounds were extracted using 80% (v/v) aqueous methanol and purified by adsorbent material XAD-7 Amberlite® resin. The mint extracts were investigated for their phenolic composition, antioxidant capacity and inhibition towards key enzymes related to diabetes, obesity and inflammation (α-glucosidase, α-amylase, pancreatic lipase and hyaluroniase). The purified extracts were also analysed using state-of-the-arts analytical instruments, including high performance liquid chromatography in conjunction with various mass spectrometric methods and nuclear magnetic resonance spectroscopy, to identify and quantify the phenolic compounds in the mints. Overall, river mint possessed high total phenolic contents, free-radical scavenging and ferric reducing ability, and exhibited strong inhibitory effects on peroxyl radical-induced oxidation superior or comparable to most common herbs such as spearmint. Mint bush exhibited lower total phenolic contents and antioxidant capacities than the Mentha species, but the results were superior to many common herbs. The crude extracts of the Australian mints were stronger inhibitors of pancreatic lipase than the other enzymes. Furthermore, the polyphenolic-rich extracts of the mint herbs were also effective inhibitors of α-glucosidase. Major compounds identified in river mint were rosmarinic acid (30.4%), neoponcirin (29.5%), chlorogenic acid (7.7%), narirutin (5.6%), and biochanin A (1.9%) and trace amount of caffeic acid, apigenin, naringenin and hesperetin. Major phenolic compounds identified in mint bush were verbascoside (48.8%), 4-methoxycinnamic acid (36.4%), glucose ester of p-coumaric acid (9.2%) and 1-O-β-ᴅ-glucopyranosyl sinapate (5.6%), while minor compounds were caffeic acid, p-coumaric acid, hesperidin and naringenin. Significantly, neoponcirin and biochanin A were identified for the first time in the Mentha genus. Furthermore, 4-methoxycinnamic acid, glucose ester of p-coumaric acid and 1-O-β-ᴅ-glucopyranosyl sinapate were identified for the first time in the genus of Prostanthera.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

………………………………………………… …………………………..……………… ……….……………………...…….… Signature Witness Signature Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

i ORIGINALITY STATEMENT

‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project’s design and conception or in style, presentation and linguistic expression is acknowledged.’

Signed ………………………………………………..

Date ………………………………………………..

‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

Signed ……………………………………………......

Date ……………………………………………......

AUTHENTICITY STATEMENT

‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Signed ……………………………………………......

Date ……………………………………………......

iii

ACKNOWLEDGEMENT

I would like to express my sincere gratitude to my supervisor Associate Prof. Jian Zhao for his guidance, patience and immense knowledge. At times when my project is not going as smoothly as expected, he has always shown me the right direction. To Dr. Izabela Konczak, I would like to send my thanks for her encouragement and support. Thank you so much to both of them for their supervision and invaluable advice.

I am grateful to Mr. Camillo Taraborreli for his continued technical assistance in the laboratory and Dr. Robert Chan for his IT support throughout my PhD.

To Mr. Lewis Alder and Dr. Martin Bucknall from the Bioanalytical Mass Spectrometry Facility, and Dr. James Hook from the Nuclear Magnetic Resonance Facility – many thanks for providing expert advice with using analytical instruments.

Lastly, I would like to thank to my family and friends, who have been with me throughout this journey, especially Ka Wai Fan for his endless support.

LIST OF PUBLICATIONS AND PRESENTATIONS

Tang, K. S., Konczak, I. & Zhao, J. (2016). Identification and quantification of phenolics in Australian native mint (Mentha australis R. Br.). Food Chemistry, 192, 698-705.

Tang, K. S., Konczak, I. & Zhao, J. (2014). Phenolic identification of Australian native herbs and their antioxidant capacity. Presented at the XXVIIth International Conference on Polyphenols & the 8th Tannin Conference (ICP2014), Nagoya, Japan, 2-6 September 2014 – poster presentation.

Tang, K. S. & Zhao, J. (2013). Antioxidant capacity and total phenolic content of three australian native plants. Presented at the Institute of Food Science and Technology 13th Annual Meeting & Food Expo, Chicago, USA, 13-16 July 2013 – poster presentation.

TABLE OF CONTENTS

TABLE OF CONTENTS ...... 1

LIST OF TABLES ...... 5

LIST OF FIGURES ...... 7

ABBREVIATIONS ...... 10

ABSTRACT ...... 12

CHAPTER 1 INTRODUCTION ...... 13

CHAPTER 2 LITERATURE REVIEW ...... 18

2.1. Introduction ...... 18

2.2. Phytochemicals identified from Australian edible native plants ...... 21 2.2.1. Phenolic compounds ...... 25 2.2.1.1. Phenolic acids ...... 25 2.2.1.2. Flavonoids ...... 25 2.2.2. Plant essential oils ...... 27 2.2.3. Carotenoids and other lipophilic compounds ...... 31

2.3. Health related biological properties of edible Australian native plants...... 36 2.3.1. Antioxidant properties of polyphenolics in native fruits, herbs and spices ...... 39 2.3.1.1. Antioxidation mechanisms of phenolic antioxidants ...... 39 2.3.1.2. Phenolic compounds and their antioxidative effects in native plants ...... 41 2.3.2. Antimicrobial properties of Australian native plants ...... 45 2.3.3. Inhibitory properties of Australian native plants on digestive enzymes ...... 48 2.3.3.1. α-Glucosidase and α-amylase inhibitory effects of Australian native plants ...... 50 2.3.3.2. Lipase inhibitory effect of Australian native plants ...... 51 2.3.4. Angiotensin I-converting enzyme inhibitory effect of Australian native plants...... 52 2.3.5. In vitro studies on anti-inflammatory properties and related cancer chemopreventive properties of Australian native plants ...... 53

2.4. Australian native mints from the Lamiaceae family selected for study in this thesis ..... 55 2.4.1. Australian native mints from the Mentha genus...... 55 2.4.2. Phytochemicals identified from the Mentha genus and their potential biological activities ... 57 2.4.3. Australian native mints from the Prostanthera genus ...... 62 2.4.4. Phytochemicals identified from the Prostanthera genus and their biological activities ...... 63

2.5. Methods for identification of polyphenolic compounds in plants...... 65 2.5.1. Sample preparation for chromatographic techniques ...... 65

2.5.2. Chromatographic techniques for qualitative and quantitative analysis of essential oils and phenolic compounds ...... 66 2.5.2.1. Gas chromatography coupled with mass spectrometry in native plant studies ...... 66 2.5.2.2. Reverse Phase-High Performance Liquid Chromatography ...... 68 2.5.2.3. Liquid chromatographic techniques coupled with mass spectrometry ...... 71 2.5.3. Nuclear Magnetic Resonance Spectroscopy in structural elucidation of phenolic compounds ...... 74

2.6. Conclusion ...... 75

CHAPTER 3 MATERIALS AND METHODS ...... 77

3.1 Plant materials ...... 77

3.2 Chemicals and reagents ...... 77 3.2.1 Chemicals used for sample preparation...... 77 3.2.2 Chemicals and reagents used for antioxidant capacity assays ...... 77 3.2.3 Chemicals and reagents used for enzyme-based assays ...... 79 3.2.4 Chromatographic and Nuclear Magnetic Resonance analyses ...... 79

3.3 Preparation of phenolic extracts ...... 79 3.3.1 Preparation of crude extracts ...... 79 3.3.2 Preparation of purified extracts ...... 80

3.4 Proximate analysis of Australian native mints ...... 82 3.4.1 Protein content ...... 82 3.4.2 Lipid analysis ...... 82 3.4.3 Moisture determination ...... 83 3.4.4 Ash determination ...... 83 3.4.5 Carbohydrate content ...... 83

3.5 Assays of antioxidant capacity of Australian native mints ...... 84 3.5.1 Total phenolic content (Folin-Ciocalteu) assay ...... 84 3.5.2 Determination of total flavonoids ...... 84 3.5.3 ABTS radical scavenging capacity assay ...... 85 3.5.4 DPPH radical scavenging capacity assay ...... 86 3.5.5 Ferric Reducing Antioxidant Power (FRAP) assay ...... 86 3.5.6 Oxygen Radicals Absorbance Capacity (ORAC) assay ...... 87

3.6 Assays of enzyme inhibition activities by extracts of Australian native mints ...... 87 3.6.1 α-Glucosidase inhibitory activity assay ...... 87 3.6.2 α-Amylase inhibitory activity assay ...... 88 3.6.3 Pancreatic lipase inhibitory assay ...... 89 3.6.4 Hyaluronidase inhibitory activity assay ...... 90

3.7 Identification and quantification of phenolic compounds in native mints...... 91 3.7.1 High Performance Liquid Chromatography-Photodiode Array Detector analysis ...... 91 3.7.2 Isolation of phenolic compounds by semi-preparative HPLC ...... 92 3.7.3 Liquid Chromatography-High Resolution Mass Spectrometry analysis ...... 92 3.7.4 Liquid Chromatography-Tandem Mass Spectrometry analysis ...... 93 3.7.5 Identification of phenolic compounds by Gas Chromatography-Mass Spectrometry ...... 94 3.7.5.1 Derivatisation of phenolic compounds ...... 94 2

3.7.6 Compound identification by GC-MS ...... 94 3.7.7 Structural assignment of phenolic compounds by Nuclear Magnetic Resonance Spectroscopy ...... 95 3.7.8 Quantification of phenolic compounds by HPLC-PDA ...... 95

3.8 Statistical Analysis ...... 96

CHAPTER 4 PHENOLIC COMPOSITION AND IN VITRO BIOLOGICAL ACTIVITIES OF MENTHA AUSTRALIS AND PROSTANTHERA ROTUNDIFOLIA ...... 97

4.1. Introduction ...... 97

4.2. Results and discussion ...... 99 4.1.1 Proximate composition of Australian native mints ...... 99 4.1.2 Purification yield of Australian native mints ...... 100 4.1.2.1 Recovery of phenolic components during purification ...... 101 4.1.2.2 General discussion on the extraction and purification of phenolic compounds from the mint samples ...... 101 4.1.3 Phenolic composition of Mentha australis and Prostanthera rotundifolia ...... 102 4.1.4 Antioxidant capacity of Mentha australis and Prostanthera rotundifolia ...... 106 4.1.4.1 ABTS free radical scavenging assay ...... 106 4.1.4.2 DPPH free radical scavenging assay ...... 108 4.1.4.3 Ferric Reducing Antioxidant Power (FRAP) assay ...... 108 4.1.4.4 Oxygen Radicals Absorbance Capacity (ORAC) assay ...... 109 4.1.4.5 General discussion on the antioxidant capacity of Australian native mints ...... 110 4.1.4.6 Correlation analysis ...... 111 4.1.5 Inhibitory effects of Mentha australis and Prostanthera rotundifolia against health related enzymes ...... 113 4.1.5.1 Inhibition against α-glucosidase ...... 113 4.1.5.2 Inhibition against α-amylase ...... 114 4.1.5.3 Inhibition against pancreatic lipase...... 116 4.1.5.4 Inhibition against hyaluronidase ...... 119

4.3. Conclusion ...... 121

CHAPTER 5 IDENTIFICATION AND QUANTIFICATION OF PHENOLIC COMPOUNDS IN MENTHA AUSTRALIS R. BR...... 122

5.1. Introduction ...... 122

5.2. Results and discussion ...... 124 5.2.1. Determination of phenolic compounds in Mentha australis by High Performance Liquid Chromatography-Photodiode Array Detector ...... 124 5.2.1.1. Optimisation of HPLC-PDA method ...... 124 5.2.1.2. Preliminary identification of the phenolic compounds with HPLC-PDA...... 126 5.2.2. Determination of phenolic compounds by Gas Chromatography-Mass Spectrometry ...... 127 5.2.3. Identification of phenolic compounds in M. australis by Liquid Chromatography-High Resolution Mass Spectrometry ...... 133 5.2.3.1. Optimisation of LC-HRMS method ...... 133

5.2.3.2. Identification of phenolic compounds in Mentha australis by Liquid Chromatography- High Resolution Mass Spectrometry and Tandem Mass Spectrometry ...... 136 5.2.4. Confirmation by Nuclear Magnetic Resonance ...... 143 5.2.4.1. 1H and 13C NMR analysis ...... 143 5.2.4.2. Two-dimensional NMR spectroscopy analysis ...... 151 5.2.4.3. Heteronuclear Single Quantum Coherence analysis ...... 151 5.2.4.4. Heteronuclear Multiple Bond Coherence analysis ...... 155 5.2.4.5. Correlation Spectroscopy analysis ...... 155 5.2.5. Quantification of phenolic compounds by High Performance Liquid Chromatography- Photodiode Array Detector analysis ...... 160

5.3. Conclusion ...... 162

CHAPTER 6 IDENTIFICATION AND QUANTIFICATION OF PHENOLIC COMPOUNDS IN PROSTANTHERA ROTUNDIFOLIA R. BR...... 164

6.1. Introduction ...... 164

6.2. Results and discussion ...... 166 6.2.1. Determination of phenolic compounds in Prostanthera rotundifolia by HPLC-PDA analysis ...... 166 6.2.2. Identification and confirmation of phenolic compounds in P. rotundifolia by LC-HRMS and LC-MS/MS ...... 168 6.2.3. Determination of phenolic compounds by Gas Chromatography-Mass Spectrometry ...... 173 6.2.4. Confirmation by Nuclear Magnetic Resonance Spectroscopy ...... 177 6.2.4.1. 1H and 13C NMR analysis ...... 177 6.2.4.2. Heteronuclear Single Quantum Coherence analysis ...... 183 6.2.4.3. Heteronuclear Multiple Bond Coherence (HMBC) analysis ...... 187 6.2.4.4. Correlation Spectroscopy analysis ...... 189 6.2.5. Quantification of phenolic compounds by HPLC-PDA ...... 193

6.3. Conclusion ...... 194

CHAPTER 7 CONCLUSIONS AND RECOMMENDATIONS ...... 195

REFERENCES ...... 200

APPENDICES ...... 216

LIST OF TABLES

Table 1. Usage of investigated edible Australian native fruits, herbs and spices1 ...... 20 Table 2. Phenolic compounds identified in commercially available edible Australian native fruits, herbs and spices, and methods of their identification ...... 23 Table 3. Main volatile components of essential oils in edible Australian native bushfoods ...... 30 Table 4. Lipophilic compounds identified in edible Australian native fruits, herbs and spices ...... 35 Table 5. Commercially available edible Australian native plants exhibiting health- benefiting properties...... 36 Table 6. Antioxidant capacity in crude methanolic extracts of commercially edible Australian native fruits (fresh weight basis) ...... 44 Table 7. Antioxidant capacity in crude methanolic extracts of commercially edible Australian native herbs, spices and seeds (dry weight basis)...... 45 Table 8. Description of three Australian native Mentha species used by indigenous people for medicinal purposes...... 56 Table 9. Major phenolic compounds and essential oil constituents identified in common Mentha species ...... 61 Table 10. Distribution and usages of three Australian native Prostanthera species for medicinal and culinary purposes in indigenous community ...... 62 Table 11. Phenolic compounds and essential oil constituents identified in Prostanthera species and their biological functions ...... 64 Table 12. Characteristic ultraviolet spectral properties of common phenolic compounds ...... 70 Table 13. Proximate analysis of native Australian mints (w/w, fresh weight)1 ...... 99 Table 14. Extraction yield of crude and purified native Australian mints (%, DW)1 .. 100 Table 15. Phenolic content in crude and purified extracts obtained from native Australian mints and reference sample of spearmint1 ...... 105 Table 16. Antioxidant capacity of crude and purified samples (per gram dry plant material) obtained from native Australian mints and reference sample of spearmint1 . 107 Table 17. Pearson correlation coefficient (r) and level (p) for relationships between total phenolic content, total flavonoid content and antioxidant assays of native Australian mints1...... 112 Table 18. Lipase inhibitory activity of crude and purified extracts obtained from native mints with spearmint as reference sample ...... 118 Table 19. Molecular ions present in the mass spectra of silylated phenolic compounds1 in the methanolic crude extract of Mentha australis identified by Gas Chromatography- Mass spectrometry ...... 131 Table 20. Phenolic compounds identified in methanol-based lyophilised extract of Mentha australis ...... 141

Table 21. Comparison of 1H NMR data of peak 4 with rosmarinic acid reference standard and literature ...... 145 Table 22. Comparison of 13C NMR data of peak 4 with rosmarinic acid reference standard and literature ...... 146 Table 23. Comparison of 1H NMR data of purified fraction of peak 6 with poncirin, neoponcirin and hesperidin ...... 153 Table 24. Comparison of 13C NMR data of purified fraction of peak 6 with neoponcirin and hesperidin ...... 154 Table 25. Heteronuclear Single Quantum Coherence and Heteronuclear Multiple-Bond Correlation data of purified fraction of peak 4 (i.e., rosmarinic acid) ...... 157 Table 26. Heteronuclear Single Quantum Coherence and Heteronuclear Multiple-Bond Correlation data of purified fraction of peak 6 (i.e., neoponcirin) ...... 158 Table 27. Quantification of phenolic compounds identified in methanol-based lyophilised extract of Mentha australis by HPLC-PDA ...... 160 Table 28. Phenolic compounds identified in methanol-based lyophilised extract of Prostanthera rotundifolia ...... 172 Table 29. Molecular ions present in the mass spectra of silylated phenolic compounds1 identified by Gas Chromatography-Mass Spectrometry of P. rotundifolia ...... 176 Table 30. Comparison of 1H NMR data of peak 4 with verbascoside and isoverbascoside ...... 179 Table 31. Comparison of 13C NMR data of peak 4 with verbascoside ...... 180 Table 32. Comparison of 1H and 13C NMR data of peak 6 with 4-methyoxycinnamic acid ...... 181 Table 33. Comparison of 1H-NMR and 13C-NMR data of peak 7 with glucose ester of p- coumaric acid ...... 182 Table 34. Heteronuclear Single Quantum Coherence and Heteronuclear Multiple-Bond Correlation data of purified fraction of peak 4 (i.e., verbascoside)1 ...... 188 Table 35. Heteronuclear Single Quantum Coherence and Heteronuclear Multiple-Bond Correlation data of purified fraction of peak 6 (i.e., 4-methoxycinnamic acid)1 ...... 189 Table 36. COSY data of peak 4 and 6 in purified P. rotundifolia extract ...... 190 Table 37. Quantification of phenolic compounds identified in methanol-based extract of Prostanthera rotundifolia by HPLC-PDA ...... 194

LIST OF FIGURES

Figure 1. Classification of phytochemicals identified from edible Australian native plants...... 22 Figure 2. The triple-ring backbone of flavonoids...... 26 Figure 3. Chemical structures of major essential oil constituents found in the leaves of Australian native plants...... 29 Figure 4. Chemical structures of lipophilic compounds identified in Australian native plants...... 34 Figure 5. Mechanism of autoxidation – involving three steps: initiation, propagation, and termination...... 39 Figure 6. Damage to lipids, proteins and DNA induced by reactive oxygen species. ... 40 Figure 7. Influence of polyphenols on the management of blood glucose in type 2 diabetes...... 49 Figure 8. Renin-angiotensin system – the role of renin and ACE in the increase of blood pressure and potentially lead to hypertension...... 53 Figure 9. Classification of Australian native mints Mentha australis R. Br. and Prostanthera rotundifolia R. Br...... 57 Figure 10. Chemical structures of common essential oil constituents and phenolic compounds identified in the Mentha and Prostanthera species...... 60 Figure 11. Outline of the experimental approach for the thesis...... 78 Figure 12. The preparation steps of the crude and purified extracts of native mint samples...... 81 Figure 13. α-Glucosidase inhibitory activity of crude (C) and purified (P) extracts obtained from native Australian mints (Mentha australis and Prostanthera rotundifolia) and reference sample of spearmint (Mentha spicata). The data represent the mean ± standard deviation of at least three independent experiments...... 114 Figure 14. α-Amylase inhibitory activity of purified extracts obtained from native Australian mints (Mentha australis and Prostanthera rotundifolia) and reference sample of spearmint (Mentha spicata). The data represent the mean ± standard deviation of at least three independent experiments...... 116 Figure 15. Pancreatic lipase inhibitory activity of crude (C) and purified (P) extracts obtained from native Australian mints (Mentha australis and Prostanthera rotundifolia) and reference sample of spearmint (Mentha spicata). The data represent the mean ± standard deviation of at least three independent experiments...... 118 Figure 16. Hyaluronidase inhibitory activity of crude (C) and purified (P) extracts obtained from native Australian mints (Mentha australis and Prostanthera rotundifolia) and reference sample of spearmint (Mentha spicata). The data represent the mean ± standard deviation of at least three independent experiments...... 120 Figure 17. HPLC chromatogram at 320 nm of purified M. australis extract with a run time of 60 min...... 125

Figure 18. HPLC chromatogram at 320 nm of purified M. australis extract: chlorogenic acid (1), caffeic acid (2), narirutin (3), rosmarinic acid (4), biochanin A (5), neoponcirin (6), apigenin and naringenin (7), and hesperetin (8). Refer to Table 22 for the precise retention time...... 125 Figure 19. Processes used to identify phenolic compounds in the purified extract of M. australis...... 126 Figure 20. Naringenin (A) and the silylated naringenin treated with BSTFA reagent (B)...... 128 Figure 21. GC Mass spectrum confirmation of trimethylsilyl-derivatised caffeic acid using the NIST library (above) and the head to tail matching (below)...... 130 Figure 22. Total ion current (TIC) GC-MS chromatogram of non-hydrolysed (A) and hydrolysed crude extract of M. australis...... 132 Figure 23. LC-HRMS Total ion current (TIC) chromatogram (A) of M. australis using atmospheric pressure chemical ionisation with mass range between m/z 120-700. Mobile phase acetonitrile and acetic acid were used. The extracted ion chromatogram: [M-H]- m/z of 359 (B) and 593 (C) were found to correspond to the major peaks from the TIC chromatogram...... 134 Figure 24. LC-HRMS Total ion current (TIC) chromatogram (A) of M. australis using negative atmosphere pressure chemical ionisation with mass range between m/z 120- 700. Mobile phase ammonium formate in water and ammonium formate in 90% methanol were used. The extracted ion chromatogram: [M-H]- of 359 (B) and 593 (C) corresponded to peaks at retention time 2 and 13.4 min, respectively...... 135 Figure 25. Comparison of molecular ion [M-H]- (m/z 359) of rosmarinic acid

(C18H16O8) in M. australis (A) and the simulated data (C18H16O8 – H) generated from Xcalibur (B) by LC-HRMS...... 136 Figure 26. Chemical structure of compounds identified in purified M. australis extract...... 137 Figure 27. Glucose and rhamnose sugar attachment in neoponcirin (left) and poncirin (right)...... 138 Figure 28. Negative HESI spectrum of peak 6 (A) obtained from LC-HRMS Total ion current (TIC) chromatography and comparison with simulated spectra of molecular ion [M-H]- spectra of m/z 593 (B) and fragment ion [M-H]- of m/z 285 (C)...... 139 Figure 29. Chemical structure of rosmarinic acid indicating the protons at positions 3, 4, 3′, 4′, which cannot be detected by NMR (above). 1H NMR spectra showing the corresponding proton chemical shifts (below)...... 144 Figure 30. 1H NMR spectra comparing reference standard rosmarinic acid (A) against the purified fraction of peak 4 from M. australis (B)...... 149 Figure 31. Structural similarity of the flavonoid backbone and differences in position 2′′ and 5′′ of the sugar component in poncirin, naringin, neoponcirin and hesperidin...... 150

Figure 32. HSQC spectrum of peak 6 of M. australis in CD3OD showing carbon position 2′′, 6′′ and 1′′′...... 156

Figure 33. COSY spectrum of crude extract of M. australis in CD3OD showing the protons relationship in the double bond of rosmarinic acid (peak 4) and benzene ring of neoponcirin (peak 6)...... 159 Figure 34. Initial HPLC chromatogram at 280 nm of purified P. rotundifolia extract...... 167 Figure 35. HPLC chromatogram at 280 nm of purified P. rotundifolia extract: caffeic acid (1), p-coumaric acid (2), hesperidin (3), verbascoside (4), 1-O-β-ᴅ-glucopyranosyl sinapate (5), 4-methoxycinnamic acid (6), glucose ester of p-coumaric acid (7), and naringenin (8)...... 168 Figure 36. LC-HRMS total ion current (TIC) chromatogram (A) of P. rotundifolia using atmosphere pressure chemical ionisation with mass range between m/z 120-700 (mobile phase: acetonitrile and acetic acid). A [M-H]- mass of 623 (B) and 177 (C) corresponded to the major peaks from the TIC chromatogram...... 169 - Figure 37. Comparison of molecular ion [M-1] (m/z 623) of peak 4 (C29H36O15) in P. rotundifolia and the simulated data generated from Xcalibur by LC-HRMS...... 171 + Figure 38. Comparison of molecular ion [M-1] (m/z 179) of peak 6 (C10H10O3) in P. rotundifolia and the simulated data generated from Xcalibur by LC-HRMS...... 171 Figure 39. Total ion current (TIC) GC-MS chromatogram of non-hydrolysed (A) and hydrolysed (B) crude extract of P. rotundifolia ...... 175 Figure 40. Chemical shift differences as given in Table 32 between verbascoside (A) and its isomer isoverbascoside (B)...... 178 Figure 41. HSQC spectrum of peak 4 indicating position 6′ of verbascoside...... 185 Figure 42. HSQC spectrum of peak 6 indicating benzene ring positions 3 and 5, positions 2 and 6, position 7 and position 8...... 186

Figure 43. COSY spectrum of crude extract P. rotundifolia (CD3OD) showing the protons relationship in the double bond of verbascoside (peak 4) and the benzene ring of 4-methoxycinnamic acid (peak 6)...... 191 Figure 44. Chemical structure of compounds identified in purified P. rotundifolia extract...... 192

ABBREVIATIONS

4-MUO 4-methylumbelliferyl oleate AAPH 2, 2’-azobis(2-amidinopropane)dihydrochloride ABTS 2, 2’-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) diammonium salt

AC Absorbance of control

ACB Absorbance of control blank ACE Angiotensin I-converting enzyme ANOVA Analysis of variance APCI Atmospheric pressure chemical ionisation

AS Absorbance of sample

ASB Absorbance of sample blank BSTFA N,O-bis(trimethylsilyl)trifluoroacetamide

CD3OD Deuterated methanol 13C NMR Carbon-13 Nuclear Magnetic Resonance COSY Correlation spectroscopy DMAB 4-dimethylaminobenzaldehyde DMSO Dimethyl sulfoxide DPPH Di(phenyl)-(2,4,6-trinitrophenyl) iminoazanium DW Dry weight EI Electron impact ESI Electrospray FRAP Ferric reducing antioxidant power FW Fresh weight GC-MS Gas Chromatography-Mass Spectrometry GE Gallic acid equivalent HESI Heated Electrospray Ionisation HPLC High Performance Liquid Chromatography HPLC-DAD High Performance Liquid Chromatography-Diode Array Detector HPLC-PDA High Performance Liquid Chromatography-Photodiode Array Detector HMBC Heteronuclear Multiple-Bond Correlation 1H NMR Proton Nuclear Magnetic Resonance HSQC Heteronuclear Single Quantum Coherence

IC50 Half maximal inhibitory concentration i.d. Internal diameter 10

LC-HRMS Liquid Chromatography-High Resolution Mass Spectrometry LC-MS Liquid Chromatography-Mass Spectrometry LC-MS/MS Liquid Chromatography-Tandem Mass Spectrometry LOD Limit of detection

MCwb Moisture content on a wet basis MRM Multiple reaction monitoring m/z Mass-to-charge ratio NMR Nuclear Magnetic Resonance ORAC Oxygen radicals absorbance capacity P-NPG p-nitrophenyl-β-ᴅ-glucopyranoside PTFE polytetrafluoroethylene RE Rutin hydrate equivalent ROS Reactive oxygen species SD Standard deviation SIM Selective ion monitoring SRM Selective reaction monitoring TE Trolox equivalent TEAC Trolox equivalent antioxidant capacity TIC Total ion current TLC Thin layer chromatography TMCS Trimethylchlorosilane TPTZ 2,4,6-tripyridin-2-yl-1,3,5-triazine v/v Volume per volume w/w Weight per weight

Wi Weight of initial sample

Wf Weight of final sample

ABSTRACT

Australia native flora has served as food and medicine for the indigenous people for thousands of years. However, systematic scientific investigation into their health properties is still in the early stage and rather limited in scope. In this thesis, the phenolic composition and in vitro biological activities of two native Australian mints – river mint (Mentha australis R. Br.) and mint bush (Prostanthera rotundifolia R. Br.) were systematically investigated for the first time. The phenolic compounds were extracted using 80% (v/v) aqueous methanol and purified by XAD-7 Amberlite® resin. The mint extracts were investigated for their phenolic composition, antioxidant capacity and inhibition towards key enzymes related to diabetes, obesity and inflammation (α- glucosidase, α-amylase, pancreatic lipase and hyaluroniase). The purified extracts were also analysed using state-of-the-arts analytical instruments, including high performance liquid chromatography in conjunction with various mass spectrometric methods and nuclear magnetic resonance spectroscopy, to identify and quantify the phenolic compounds in the mints. Overall, river mint possessed high total phenolic contents, free-radical scavenging and ferric reducing ability, and exhibited strong inhibitory effects on peroxyl radical-induced oxidation superior or comparable to most common herbs such as spearmint. Mint bush exhibited lower total phenolic contents and antioxidant capacities than the Mentha species, but the results were superior to many common herbs. The crude extracts of the Australian mints were stronger inhibitors of pancreatic lipase than the other enzymes. Furthermore, the polyphenolic-rich extracts of the mint herbs were also effective inhibitors of α-glucosidase. Major compounds identified in river mint were rosmarinic acid (30.4%), neoponcirin (29.5%), chlorogenic acid (7.7%), narirutin (5.6%), and biochanin A (1.9%) and trace amount of caffeic acid, apigenin, naringenin and hesperetin. Major phenolic compounds identified in mint bush were verbascoside (48.8%), 4-methoxycinnamic acid (36.4%), glucose ester of p- coumaric acid (9.2%) and 1-O-β-ᴅ-glucopyranosyl sinapate (5.6%), while minor compounds were caffeic acid, p-coumaric acid, hesperidin and naringenin. Significantly, neoponcirin and biochanin A were identified for the first time in the Mentha genus. Furthermore, 4-methoxycinnamic acid, glucose ester of p-coumaric acid and 1-O-β-ᴅ-glucopyranosyl sinapate were identified for the first time in the genus of Prostanthera. 12

CHAPTER 1 Introduction

Australia is home to a diverse range of unique native plants, which have served as a main source of food and medicine for the indigenous people for thousands of years. Prior to European settlement, the lifestyle of indigenous people was based upon the use of native flora and fauna for most of life’s essentials – from obtaining the essential nutrients to healing wounds and fighting infections (Gott, 2008, Hodgson and Wahlqvist, 1993). The indigenous people have developed an extensive knowledge of native plants in terms of their nutrition as well as using them for treatment of various illnesses and diseases (Barr et al., 1988). Over 500 Australian native plants have been used in traditional medicine long before the development of synthetic drugs and the arrival of modern medical science (Lassak and McCarthy, 2001). With the establishment of modern medical services since the European settlement, the medicinal role of native plants has gradually diminished and, for a long time, even their uses as food have largely been confined to the exotic food category, locally known as “bush tucker”, and overlooked by the general public. Since the 1980s, however, there has been a renewed and growing interest in native bushfoods, partly because of the nation’s desire for a local cuisine with a distinctly Australian flavour, but also due to the discovery that some native plants are extremely rich in certain nutrients. The exceedingly high vitamin C content in the Kakadu plant, over 50 times higher than orange (Brand et al., 1982), is a typical example of such discoveries. Since then, the Australian bushfood industry has grown tremendously and now is a sizable industry made of indigenous pickers, plantation farmers, processors and retailers (Phelps, 1997, Sultanbawa and Sultanbawa, 2016).

The Australian bushfoods cover a wide range of products from fruits, nuts and seeds to leaves, flowers, tubers, fungi and seaweeds. In recent years, consumers of native edible plants are expanding rapidly. According to the Rural Industries Research and Development Corporation report by Cleary et al. (2009), the bushfood industry, including essential oils from native plants, has an estimated annual value of $10 million, while the value for native food alone is $5 million. A number of endemic edible Australian native plants, especially fruits, vegetables, herbs and spices, are commercially produced as fresh and dried products (Ahmed and Johnson, 2000). Furthermore, the general public are becoming more aware of the health benefits of 13 native plants. Mainstream food manufacturers and retailers have shown an increasing interest in native food and many restaurants are also promoting and incorporating native bushfoods in their cuisine (Bannerman, 2006). Due to their unique tastes and flavours, native Australian fruits, herbs and spices are used for seasoning and garnish as well as being incorporated into a number of food, personal care and household products. The vegetables are consumed fresh as salads or cooked while the seeds are used as flavourings and flour.

Due to the harsh climate conditions of the Australian continent, such as low rainfall and high UV radiation, native plants here are exposed to a high degree of environmental stress. During the evolutionary process, a spectrum of native flora has transformed into unique species that are able to robustly protect themselves from the stresses (Cock, 2011). Part of the protective mechanisms developed by the plants is their high levels of unique phytochemicals, such as phenolic compounds, which can help them to mitigate the damages caused by oxidative species, among others things, generated during photosynthesis and other physiological activities (Liu, 2003). It is these phytochemicals that are responsible for the disease preventing and curing properties of the native plants that have been utilised by the indigenous people since time immemorial (Cock, 2011). However, much of the knowledge on nutritional and medical values of native plants has been passed down from generation to generation by word of mouth and, as a result, much of which has been forgotten (Barr et al., 1988, Lassak and McCarthy, 2001). In recent decades, a number of scientific studies have examined the biological activities of Australian native plants which are found to possess substantial antimicrobial, antioxidant and other health-promoting properties (Zhao and Agboola, 2007, Konczak, 2009). However, there are still a large number of Australian native plant species that are yet to be explored.

Earlier research in this area has been directed towards identifying the composition of essential oils and plant extracts of native trees, bushfoods, wildflowers and truffles as well as their association to health beneficial properties. Three iconic Australian native trees from the Eucalyptus, Acacia and Melaleuca genus, were a significant part of the Australia’s essential oil industry. Their oil extracts have substantial antibacterial, antifungal, antiviral and anti-inflammatory properties (Noumi et al., 2011). Early studies have also revealed that many Australian native bushfoods contain significant levels of

14 phenolic compounds, many of which have been found to confer a number of benefits to human health (Netzel et al., 2006, Zhao and Agboola, 2007, Konczak, 2009).

Phenolic compounds belong to one of the largest class of phytochemicals and they occur abundantly in the plant kingdom as secondary metabolites (Dai and Mumper, 2010). The fundamental structure of phenolic compounds consists of a hydroxyl group attached to an aromatic ring. However, phenolic compounds found in plants display a large diversity of structures, from simple phenolic acids, through polyphenols such as flavonoids, to polymeric compounds formed from these different classes (Khoddami et al., 2013). Initially, phenolic compounds are mainly known for their role in the quality of plant based food. For example, they are responsible for the colour of many fruits and vegetables and flavour properties such as astringency in tea, and are the substrates for enzymatic browning (Garcia-Salas et al., 2010). In recent decades, however, the possible contribution of phenolic compounds to health has received considerable attention in the research community (Balasundram et al., 2006, Ramassamy, 2006, Ignat et al., 2011) . Plant phenolic compounds and their health-promoting properties have been widely studied and were found to possess a broad range of bioactivities associated with health (Huang et al., 2009). Dietary intake of these compounds through the consumption of fruits, vegetables, herbs and spices has been linked with reduced risk of a number of major chronic diseases such as diabetes, obesity, inflammation, cardiovascular disorder and cancer via several different mechanisms and pathways (Liu, 2003, Khoddami et al., 2013, Tucker and Robards, 2008). For instance, phenolic compounds have the ability to scavenge free radicals in the body, thereby mitigating oxidative stress and related disorders (Jacob et al., 2012). They also exhibit potential to inhibit key digestive enzymes, such as α-glucosidase and α-amylase, which are involved in the production of excessive sugars in diabetic patients (Tadera et al., 2006). In addition to the above mentioned bioactivities, phenolic compounds have been reported to exhibit other activities including anti-microbial, anti-proliferative, apoptosis, hydrogen peroxide induced cell damage and antibiotic effects (Tucker and Robards, 2008).

The effects of polyphenolic extracts derived from some Australian native bushfoods have been well investigated for its free radical scavenging effects, ferric reducing power and inhibitory activities against food pathogenic and spoilage bacteria (Konczak et al.,

2010b, Netzel et al., 2006, Weerakkody et al., 2010, Zhao and Agboola, 2007). Recently, studies of native plants have also been directed towards discovering other potential bioactivities, particularly for those related to prevention of chronic diseases, for instance, the capacity to inhibit key enzymes involved in diabetes, obesity, hypertension (Sakulnarmrat and Konczak, 2012, Sakulnarmrat et al., 2014), inflammation (Guo et al., 2014, Tan et al., 2011a) and proliferation of human cancer cells (Sakulnarmrat et al., 2013, Tan et al., 2011b, Vuong et al., 2014). Notable findings include studies on Kakadu plum and quandong, which possessed substantial antioxidant capacity in comparison to that of blueberries (Konczak et al., 2010a); while Tasmannia pepper, anise myrtle and lemon myrtle were found to inhibit digestive enzymes including α-glucosidase and pancreatic lipase, which are major contributors to chronic illnesses such as diabetes and obesity (Sakulnarmrat and Konczak, 2012). Most of these studies, however, have not identified the key components, such as phenolic acids and flavonoids, which are responsible for these health benefitting properties in the plants.

In summary, although there has been considerable amount of research conducted on health related biological activities of a number of Australian native bushfoods, systematic scientific investigations into this bountiful source of food and medicinal plants and phytochemicals are still in the early stage and there are many more plants yet to be explored. Significant gap of knowledge also exists in the identification of the phytochemicals in the plants that are responsible for their health benefitting properties. This PhD project is thus conceived to partly fill these important areas of knowledge gap.

Two native Australian herbs, Mentha australis and Prostanthera rotundifolia, both from the Lamiaceae family, were selected for study in this PhD project. The Lamiaceae family comprises a number of common culinary herbs such as rosemary, mints, oregano and basil, which have been reported to exhibit a range of health-promoting properties from antioxidant activity to antidiabetic and anti-inflammatory effects that are largely due to their polyphenolic compounds (Kratchanova et al., 2010, Park, 2011, Shan et al., 2005). Mentha and Prostanthera are two of the largest genera of the Lamiaceae family that are used in food and herbal medicine. Mentha australis and Prostanthera rotundifolia are currently available commercially both through collection of wild plants by indigenous pickers and cultivation in small scale farms. Traditionally, M. australis was used for headaches, cough and colds, while P. rotundifolia was used to treat sores,

16 skin diseases, aches and pains (Lassak and McCarthy, 2001). Zhao and Agboola (2007) have reported that both M. australis and P. rotundifolia exhibit antioxidant and antimicrobial activities, while Fulton (2000) has investigated the essential oil component of P. rotundifolia. However, other health-related biological activities of these two native herbs have not been studied and no study, to date, has attempted to identify the phenolic compounds in them.

The overall aim of this PhD project was to expand the knowledge on the phenolic composition and health-related biological activities of Mentha australis R. Br. and Prostanthera rotundifolia R. Br. The specific objectives were:

1. To investigate the key health-related bioactivities of Mentha australis and Prostanthera rotundifolia, including antioxidant capacity, inhibitory activities against digestive enzymes (α-glucosidase and α-amylase and obesity pancreatic lipase) and the inflammation related hyaluronidase, using in vitro methods; and 2. To identify and quantify the main phenolic components of M. australis and P. rotundifolia using a number of modern techniques, including High Performance Liquid Chromatography-Photodiode Array Detector (HPLC-PDA), Gas Chromatography-Mass Spectrometry (GC-MS), Liquid Chromatography-High Resolution Mass Spectrometry (LC-HRMS), Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS) and Nuclear Magnetic Resonance (NMR) Spectroscopy.

The thesis begins with a general introduction, followed by a review of recent literatures on the subject, focusing on the health beneficial bioactivities of Australian native plants and the identification of the phytochemicals that are responsible for the activities. Chapter 3 outlines the research approach and describes the experimental methods used for the entire project. Chapter 4 describes phenolic contents of the two native mints and the health-related bioactivities of their phenolic extracts. Chapter 5 and Chapter 6 give a detailed account of the identification of phenolic compounds in Mentha australis and Prostanthera rotundifolia, respectively, using HPLC-PDA, GC-MS, LC-HRMS, LC- MS/MS and NMR techniques. Finally, Chapter 7 summarises the major conclusions of the thesis and provides some recommendations for future research in this field.

CHAPTER 2 Literature Review

2.1. Introduction

Australian native flora has been used for nutritional and medicinal purposes since long before the introduction of supplements and pharmaceutical drugs. Barr (1993) has described the medicinal usage of various plants by the Australian indigenous people from the Northern Territory. Lassak and McCarthy (2001) have compiled a collection of medicinal native plants from across Australia including methods of preparation and suspected active ingredients. However, the use of native plants as herbal remedies was gradually replaced by pharmaceutical drugs and knowledge of their traditional usage faded as time passed.

Since its establishment in the 1980s, the Australian bushfood industry has grown markedly (Zhao and Agboola, 2007). Several hundreds of native edible plant species have been reported in Australia (Ahmed and Johnson, 2000), and at least 208 of these are located in the Sydney region (Cherikoff and Isaacs, 1989). These native fruits, herbs and spices are increasingly used at a commercial scale as ingredients in many food products (Zhao and Agboola, 2007). A number of products have incorporated fresh or dried native plants or their essential oils to enhance flavour. Native herbs have been used as alternatives to common culinary herbs to improve nutritional value and taste, while native fruits such as native finger limes have replaced commercial lemons and limes or made into sauces and jams. Australian native plants have also been used in cheese making. Agboola and Radovanovic-Tesic (2002) have studied the effects of dried lemon myrtle, cut leaf and bush tomato on cheese maturation and reported that bush tomato can significantly influence the properties of the cheese, improve the rate of rennet action and affect lipid and protein hydrolysis.

In the last decade, an increasing number of studies have been directed towards unravelling the potential health properties of edible native plants and the chemical constituents that are responsible for these health benefits (Konczak, 2009). Around 25 species have been investigated for their bioactive properties, including fruits, seeds, nuts, herbs and spices (Table 1). Other investigated aspects of native bushfoods include their influence on cheese maturation (Agboola and Radovanovic-Tesic, 2002),

18 emulsifying properties of wattle seed (Agboola et al., 2007, Agboola et al., 2012), effect of thermal processing on commercial bushfoods and its products (Sommano et al., 2011, Ee et al., 2009a), and effects of post-harvest handling (McDonald et al., 2006) on food safety and quality (Forbes-Smith and Paton, 2002, Hegarty et al., 2001). However, studies on their phytochemical composition are still limited and there are a large number of species that are yet to be explored. In addition to providing an overview of the research that has been conducted on Australian bushfoods, this review will focus on examining the main phytochemicals (i.e. phenolic compounds, carotenoids and essential oils) identified and their association with the reported biological functions. Furthermore, the advantages and disadvantages of the analytical techniques used in the identification of phenolic compounds will be discussed.

Table 1. Usage of investigated edible Australian native fruits, herbs and spices1

Common name Family Scientific name Habitat2 Usage3

Fruits Australian desert lime Rutaceae Citrus glauca Semi-arid/arid Fruit Brush cherry Myrtaceae Syzygium australe Subtropical Fruit Burdekin plum Anacardiaceae Pleiogynium timorense Rainforest Fruit, wine, jam, jelly Bush tomato Solanaceae Solanum centrale Arid Dried for flavour, spice Cedar bay cherry Myrtaceae Eugenia reinwardtiana Tropical/subtropical Fruit, jam Davidson’s plum Cunoniaceae Davidsonia pruriens Rainforest Fruit, jam Finger lime Rutaceae Citrus australasica Rainforest Fruit Illawarra plum Podocarpaceae Podocarpus elatus Rainforest Fruit, jam Kakadu plum Combretaceae Terminalia ferdinandiana Tropical Fruit, jam, sauce, juice Lemon aspen Rutaceae Acronychia acidula Tropical Sweet and savoury dish Magenta lilly pilly Myrtaceae Syzygium paniculatum Littoral/subtropical Fruit Molucca raspberry Rosaceae Rubus moluccanus Tropical/rainforest Fruit Mountain pepper Winteraceae Tasmannia lanceolate Cool temperate Herbs, oil Muntries Myrtaceae Kunzea pomifera Semi-arid/coast Fruit Native currant Ericaceae Acrotriche depressa Woodland Fruit, jam, jelly Quandong Santalaceae Santalum acuminatum Semi-arid Fruit, jam Riberry Myrtaceae Syzygium luehmannii Rainforest Fruit, jam

Herbs and spices Anise myrtle Myrtaceae Syzgium anisatum Rainforest Herb and spice Cut leaf Lamiaceae Prostanthera incisa Coastline Herbs, oil Lemon ironbark Myrtaceae Eucalyptus staigeriana Terrestrial Spice, herbal tea Lemon myrtle Myrtaceae Backhousia citriodora Coastal rainforest Herbs, oil Mint bush Lamiaceae Prostanthera rotundifolia Cool temperate Herbs, oil Mountain pepper Winteraceae Tasmannia lanceolate Cool temperate Herbs, oil River/Native mint Lamiaceae Mentha australis Rivers/creeks Herb, oil Strawberry gum Myrtaceae Eucalyptus olida Woodland Oil, spice, herbal tea Vegetables and seeds Warrigal greens Aizoaceae Tetragonia tetragonioides Coastal/ inland salt Salad, quiche marshes Wattle seed Fabaceae Acacia spp. (Acacia Inland/coastal Flavour, victoriae) baking, seed, coffee substitute 1 Updated from Ahmed and Johnson (2000). Refer to Table 5 for their bioactive functions. 2Obtained from Australian Native Plants Society (Australia), anpsa.org.au (ANPS, 2015) 3Oil: essential oil distilled from leaves for flavour

2.2. Phytochemicals identified from Australian edible native plants

Phytochemicals are secondary metabolites naturally synthesised in plants, which are not essential nutrients like vitamins, yet contain a variety of biologically active constituents or bioactive compounds that can confer significant health benefits to human (Dai and Mumper, 2010). From a botanic point of view, phytochemicals perform a number of functions of biological significance to the plant. In addition to providing fruits and vegetables with colour and flavour, these compounds are toxic to many lower life forms to an extent that protects plants from predators and also diseases caused by fungi, bacteria or viruses (Liu, 2003). As many phytochemicals are natural antioxidants, they also protect plants from the harmful effects of light radiation and reactive molecules generated from photosynthesis (Johnson and Williamson, 2003). Thus, phytochemicals play a crucial role in plant growth and development (Khoddami et al., 2013).

Plants contain a diverse range of phytochemicals, which are generally classified into six main groups – phenolic compounds, essential oils, carotenoids, alkaloids, nitrogen- containing compounds and organosulfur compounds (Bellik et al., 2012, Raaman, 2006). For indigenous Australian edible fruits, herbs and spices, phenolic compounds and essential oils are the two most studied groups of phytochemicals in contrast to the others (Figure 1). Information on phenolic compounds identified in commercially cultivated edible Australian native plants is compiled in Table 2.

Other (tannins, stilbenes, coumarin) Hydroxybenzoic acids Gallic, vanillic

Caffeic, ferulic, Phenolic acids Hydroxycinnamic acid coumaric

Quercetin, kaempherol, Phenolic compounds Flavonol myricetin Flavonoids Carotenoids Flavones Apiginen, luteolin (carotene, lycopene)

Hesperetin, hesperdin, Flavonones naringenin Phytochemicals in Australian native plants Flavanol Catechin, epicatechin

Terpenoids

Essential oils Anthocyanidin Cyanidin, pelagonidin

Terpenes Others (alkaloids, nitrogen-containing, Isoflavones Genistein organosulfur compounds)

Figure 1. Classification of phytochemicals identified from edible Australian native plants (Bellik et al., 2012). Groups that are more

22 extensively studied are presented in the box highlighted in red.

Table 2. Phenolic compounds identified in commercially available edible Australian native fruits, herbs and spices, and methods of their identification

Common name Instrumental Phenolic compounds2 References (Scientific name) analysis1 Fruits Brush cherry LC-MS/MS Malvidin 3,5-diglucoside, delphinidin 3,5- (Netzel et al., 2007) (Syzygium australe) (ESI) diglucoside, peonidin 3,5-diglucoside, petunidin 3,5-diglucoside Burdekin plum LC-MS/MS Chrysanthemin (Netzel et al., 2006, (Pleiogynium (ESI) Netzel et al., 2007) timorense) Cedar Bay cherry LC-MS/MS Chrysanthemin (Netzel et al., 2006, (Eugenia (ESI) Netzel et al., 2007) reinwardtiana) Davidson’s Plum LC-MS/MS Cyanidin 3-sambubioside, delphinidin 3- (Netzel et al., 2006, (Davidsonia pruriens) (ESI) sambubioside, peonidin 3-sambubioside, Netzel et al., 2007, petunidin 3-sambubioside, (malvidin 3- Konczak et al., 2010a, sambubioside, rutin, quercetin hexoside, Konczak, 2009) myricetin, pelargonidin sambubiose) (Davidsonia jerseyana) HPLC-DAD, Delphinidin sambubioside, cyanidin (Konczak, 2009) LC-MS/MS sambubioside, petunidin sambubioside, peonidin sambubioside Finger Lime LC-MS/MS Chrysanthemin, peonidin 3-glucoside (Netzel et al., 2007) (Citrus australasica) (ESI) Illawarra plum LC-MS/MS Chrysanthemin, pelargonidin 3-glucoside, (Netzel et al., 2006, (Podocarpus elatus) (ESI) delphinidin 3-glucoside Netzel et al., 2007, Tan et al., 2011b) Kakadu plum HPLC-DAD, Quercetin/hesperitin glucoside*, (Konczak, 2009) (Terminalia LC-MS/MS kaempherol/luteolin glycoside* ferdinandiana) Lemon aspen HPLC-DAD, kaempherol/luteolin glycoside*, quercetin (Konczak, 2009) (Acronychia acidula) LC-MS/MS hexoside, rutin*, chlorogenic acid*, caffeic acid*, coumaric acid*, ferulic acid* Magenta Lilly Pilly HPLC Chlorogenic acid, epicatechin, catechin, (Vuong et al., 2014) (Syzygium gallic acid paniculatum) Molucca raspberry LC-MS/MS Cyanidin 3-rutinoside, chrysanthemin, (Netzel et al., 2006, (Rubus moluccanus) (ESI) pelargonidin 3-rutinoside Netzel et al., 2007) Mountain Pepper HPLC-DAD, Chlorogenic acid, caffeic acid, quercetin, (Netzel et al., 2006, (berry) LC-MS/MS rutin, chrysanthemin, cyanidin 3- Netzel et al., 2007, (Tasmannia (ESI) rutinoside, cyanidin 3-glucoside, Konczak et al., 2010b, lanceolata) (kaempherol/ luteolin hexoside) Konczak, 2009) Muntries LC-MS/MS Delphinidin 3-glucoside, chrysanthemin (Netzel et al., 2006, (Kunzea pomifera) (ESI) Netzel et al., 2007) Quandong LC-MS/MS Chrysanthemin, pelargonidin 3-glucoside, (Konczak et al., 2010a, (Santalum (ESI) rutin, kaempherol, quercetin rutinoside, Konczak, 2009) acuminatum) (cyanidin 3-rutinoside, chlorogenic acid, kaempherol/luteolin rutinoside)

Riberry LC-MS/MS Cyanidin 3,5-diglucoside, cyanidin 3- (Netzel et al., 2007, (Syzygium luehmannii) (ESI) glucoside, chrysanthemin, cyanidin 3- Konczak et al., 2010a, galactoside, rutin, quercetin hexoside, Konczak, 2009) quercetin rhamnoside, kaempherol, (myricetin hexoside, kaempherol/luteolin rutinoside) 1HPLC: High Performance Liquid Chromatography; DAD: Diode Array Detector; MS/MS: Tandem Mass Spectrometry; ESI: Electrospray Ionisation. 2Minor compound are in brackets; possible compounds (i.e., confirmation required) are indicated by *.

Table 2. (Continued) Phenolic compounds identified in commercially available edible Australian native fruits, herbs and spices, and methods of their identification

Common name Instrumental Phenolic compounds1 References (Scientific name) analysis Herbs and spices

Anise myrtle HPLC-DAD; Chlorogenic acid, quercetin (Guo et al., 2014, (Syzgium anisatum) LC-MS/MS pentoside, myricetin, ellagic acid Sakulnarmrat and (ESI) and derivatives, catechin, Konczak, 2012, quercetin, myricetin, hesperetin, Konczak, 2009, quercetin hexoside* Konczak et al., 2010b)

Bush tomato (spice) HPLC-DAD, Hydroxybenzoic acid, ferulic (Konczak, 2009, (Solanum centrale) LC-MS/MS acid, chlorogenic acid, (caffeic Konczak et al., 2010b) (ESI) acid, p-coumaric acid, quercetin hexoside, rutin, kaempherol), quercetin rutinoside/hexoside*, kaempherol/luteolin hexoside*

Lemon Myrtle HPLC-DAD, Myricetin, hesperetin rhamnoside, (Konczak et al., 2010b, (Backhousia LC-MS/MS hesperetin hexoside, (hesperetin Konczak, 2009, citriodora) (ESI) pentoside*), ellagic acid and Sakulnarmrat and derivative, quercetin Konczak, 2012, Guo et al., 2014) Mountain Pepper HPLC-DAD, p-coumaric acid, quercetin, (Konczak et al., 2010b, (leaf) LC-MS/MS chrysanthemin, antirrhinin, Sakulnarmrat and (Tasmannia (ESI) chlorogenic acid, quercetin, Konczak, 2012, Guo et lanceolata) quercetin glycosides, quercetin 3- al., 2014, Konczak, rutinoside, cyanidin 3-glucoside, 2009) cyanidin 3-rutinoside, rutin

Vegetables and seeds

Warrigal greens HPLC-PDA Quercetin, kaempherol (Yang et al., 2008) (tetragonia tetragonioides) Wattle seed HPLC-DAD, Rutin hexoside*, quercetin (Konczak, 2009) (Acacia spp.) LC-MS/MS hexoside*, kaempherol/luteolin hexoside* 1HPLC: High Performance Liquid Chromatography; DAD: Diode Array Detector; MS/MS: Tandem Mass Spectrometry; ESI: Electrospray Ionisation. 2Minor compound are in brackets; possible compounds (i.e., confirmation required) are indicated by *.

2.2.1. Phenolic compounds 2.2.1.1. Phenolic acids

The bioactivities of phenolic compounds are diverse, with several structures reported to exhibit properties that can reduce oxidative damage, inflammation and carcinogenicity (Liu, 2003, Giardi et al., 2011). Phenolic acids are common in fruits and herbs, as well as cereal grains. These organic compounds comprise of a phenolic ring attached to a carboxylic acid, and can be divided into two major classes, namely hydroxybenzoic and hydroxycinnamic acids with benzoic acid and cinnamic acid forming the backbone of compounds in the corresponding class. Despite its common occurrence in plants, the levels of hydroxybenzoic acids (e.g., gallic and vanillic acid) are found to be minimal in Australian native fruits, herbs and spices. Rather, hydroxycinnamic acids (caffeic, p- coumaric, ferulic and chlorogenic acid) are found to be present more abundantly in Australian native edible plants. Interestingly, ellagic acid, a dimeric derivative of gallic acid, and its derivatives are found to be dominant in the native herbs, anise myrtle and lemon myrtle (Sakulnarmrat and Konczak, 2012), and in Davidson’s plum (Sakulnarmrat et al., 2014).

2.2.1.2. Flavonoids

The triple-ring structure illustrated in Figure 2 is a distinctive feature which constitutes the backbone of almost all flavonoids (de la Rosa et al., 2010). Variations in the chemical structure of the triple-ring backbone lead to the division of flavonoids into six groups, namely flavonols, flavones, flavanones, flavanols, anthocyanidins and isoflavones (Pietta, 2000). Although the variations in structure may appear relatively minor, they can induce significant changes in chemical properties and bioactivities to the compounds. For instance, flavonols may have higher antioxidant capacity than flavones due to the presence of a hydroxyl group at position 3 of the C ring (Pietta, 2000). Relatively high levels of flavonoids and their glycosides are found in many native Australian food plants, with quercetin, myricetin, hesperetin, rutin and their derivatives being common in most herbs and spices, whilst anthocyanidin occurs more frequently in bush fruits.

Figure 2. The triple-ring backbone of flavonoids.

Anthocyanins (i.e., anthocyanidin with a sugar substituent) are a group of flavonoids that occur abundantly in many plants, including fruits such as blueberries and raspberries, and are largely responsible for the colours (i.e. red, purple and blue) of these plants (Dai and Mumper, 2010). They have been reported to exhibit many important biological activities such as antioxidant activity, and also inhibitory activities against cyclooxygenase and the growth of cancer cells (Bowen-Forbes et al., 2010, Huang et al., 2009). Several Australian native fruits are found to be rich in anthocyanins (Table 2), with Davidson’s plum containing the highest number of different anthocyanins compared to other fruits (Konczak et al., 2010a, Netzel et al., 2006). Chrysanthemin (cyanidin-3-O-glucoside) is the most abundant anthocyanin in native Australian plums, berries and limes (Netzel et al., 2007, Konczak et al., 2010a). Interestingly, no anthocyanin has been identified so far in Kakadu plum (Tan et al., 2011b), a fruit grown in the northern tropical regions with the reputation of containing the highest amount of vitamin C in all fruits (1183 g/ 100 g dry basis, McDonald et al. (2006)).

In addition to fruits, herbs and spices, other native plants, such as the leafy vegetable Warrigal greens (Tetragonia tetragonioides) and wattle seed (Acacia spp.) – two of the top 14 commercially significant edible native plants (Ahmed and Johnson, 2000) – are also reported to contain phenolic compounds. Warrigal greens are widely cultivated and their leaves are consumed fresh, often used as an alternative to spinach. This vegetable has been reported to contain quercetin and kaempherol (Yang et al., 2008). However,

26 only trace amounts of flavonoids (quercetin hexoside, rutin hexoside and kaempherol/luteolin hexoside) were identified in wattle seeds (Konczak et al., 2010b).

2.2.2. Plant essential oils

Essential oils are complex, volatile compounds with a characteristic strong odour which can be derived from leaves, stems, flowers and seeds. The composition of the essential oil characterises the distinct aroma of a particular plant. Their applications range from aromatherapy, perfumes and cosmetics to medicinal and pharmaceutical products (Burt, 2004). The utilisation of plant essential oils for medicinal purposes has a long history in many parts of the world, where they are used as traditional remedies against microbial infections and diseases (Sultanbawa and Sultanbawa, 2016, Rios and Recio, 2005). Plant essential oils have been well documented to contain chemical constituents that have the ability to reduce proliferation of infectious microorganisms, protect products against food pathogens and produce analgesic and anti-inflammatory effects (Burt, 2004, Gutierrez et al., 2008, Calo et al., 2015). In the food industry, essential oils from herbs, spices and fruits have been applied in various products as flavouring agents and preservatives. Similar to those from other sources, essential oils from Australian native plants are predominantly composed of mono- and sesquiterpenes (Figure 3), some of which contain a phenolic ring structure.

In comparison to the relatively large number of studies on phenolic compounds, there are only a few reports on the essential oils in edible Australian native bushfoods, especially the fruits. Table 3 displays the main volatiles identified in edible Australian fruits, herbs and spices. Southwell and Brophy (1992) first identified the chemical constituents in the oils of the Australian native Tasmannia genus. Among these species, oils from the leaf of Tasmannia lanceolata (mountain pepper) were found to contain α- pinene (14%), 1,8-cineole (14%), β-pinene (6%), limonene (4%). The oils extracted from the native Backhousia genus were examined in several studies (Brophy et al., 2007, Southwell et al., 2000, Brophy et al., 1995). Most of the hydrodistilled essential oils from this genus were found to contain high contents (up to 90%) of the lemon- scented citral mixture neral (α-citral) and geranial (β-citral). Citronellal, a compound structurally similar to citral which gives off an undesirable flavour, is also present in 27 some species. Lemon myrtle (Backhousia citriodora) was found to have two chemotypes: one with geranial (57.8%) and neral (39.5%) while the other with citronellal (63-80%) and isopulegol (13%) (Brophy et al., 1995). Blewitt and Southwell (2000) also identified the constituents in anise myrtle oil (Syzgium anisatum, previously named Backhousia anisata) from 10 different sites in New South Wales, Australia. Anise myrtle oil also exists as two chemotypes: one with (E)-anethole (>83%) as the main constituent and methyl chavicol as the minor one (>4%), while the other having higher methyl chavicol (>77%) and lower (E)-anethole (>19%) content (Blewitt and Southwell, 2000, Brophy et al., 1995, Brophy et al., 2007). Brophy et al. (2004) also studied the oil composition from the Acronychia genus, in which oils from lemon aspen leaves (Acronychia acidula) were found to contain δ-3-carene and terpinolene. Gilles et al. (2010) identified the main constituents in the oils of two Eucalyptus herbs, lemon ironbark (Eucalyptus staigeriana) and strawberry gum (Eucalyptus olida). The leaves of lemon ironbark contain 1,8-cineole (34.8%), α-neral (10.8%), geranial (10.8%) and phellandrene (8.8%), while (E)-methyl cinnamate (>99.4%) was the main constituent identified in the latter. Fulton (2000) studied the oil extracted from the herbs P. incisa and P. rotundifolia, and showed that both species contain the principal components cineole, α-pinene and β-pinene. With regards to essential oils in Australian native fruits, very few studies have been conducted to date. The only available report is on the peels of finger limes, which were found to contain limonene (73.5%) and isomenthone (7.5%) as the major constituents (Delort and Jaquier, 2009).

Anethole Methyl chavicol

Methyl cinnamate Limonene

Terpinolene 1,8-cineole α-pinene

β-pinene δ-3-carene

Geranial Neral

Figure 3. Chemical structures of major essential oil constituents found in the leaves of Australian native plants.

Table 3. Main volatile components of essential oils in edible Australian native bushfoods

Common name Plant Extraction Essential oils1 References (Scientific name) part techniques

Anise myrtle leaf Ethanol, steam Anethole (>83%), (Blewitt and (Syzgium anisatum) distillation methyl chavicol Southwell, 2000, (>4%); or anethole Brophy et al., (>19%), methyl 1995) chavicol (>77%)

Cut leaf leaf Dichloromethane 1,8-cineole, α-pinene, (Fulton, 2000) (Prostanthera incisa) β-pinene, globulol

Lemon aspen leaf Steam distillation δ-3-carene (32-40%), Brophy et al. (Acronychia acidula) terpinolene (13-46%) (2004)

Lemon ironbark leaf Steam distillation 1,8-cineole (34.8%), (Gilles et al., (Eucalyptus staigeriana) neral (10.8%), geranial 2010) (10.8), α-phellandrene (8.8%)

Lemon Myrtle leaf Steam distillation Geranial (57.8%), (Brophy et al., (Backhousia citriodora) neral (39.5%); or 1995, Southwell citronellal (63-80%), et al., 2000) isopulegol (13%)

Mint bush leaf Dichloromethane 1,8-cineole, α-pinene, (Fulton, 2000) (Prostanthera β-pinene rotundifolia)

Mountain pepper (leaf) leaf Ethanol, steam α-pinene (14%), 1,8- (Southwell and (Tasmannia lanceolata) distillation cineole (14%), β- Brophy, 1992) pinene (6%), limonene (4%)

Strawberry gum leaf Steam distillation (E)-methyl cinnamate (Gilles et al., (Eucalyptus olida) (99.4%) 2010)

Finger lime peel Dichloromethane Limonene (73.5%), (Delort and (Citrus australasica) isomenthone (7.5%) Jaquier, 2009)

1Approximate percentage compositions are in brackets.

2.2.3. Carotenoids and other lipophilic compounds

Carotenoids represent a large group of pigments with hundreds of naturally occurring variants identified (Azqueta and Collins, 2012). These lipid-soluble compounds can only be synthesised in plants and microorganisms and they are responsible for the red, orange and yellow colours of fruits and vegetables (Britton et al., 2009). The basic structure of carotenoids is made of eight isoprene units forming a tetraterpene, with two sets of 20 carbon units arranged in a way that both are joined from tail to tail (DeMan, 2013). Carotenoids have a major role in mitigating the harmful effects of byproducts formed from photosynthesis, which is partly related to their antioxidant activity to scavenge free radicals (Britton et al., 2009). The ability of carotenoids to scavenge free radicals is thought to be responsible for the numerous health benefits that many of them reportedly confer to the human body (Rao and Rao, 2007). β-Carotene, lycopene and lutein are examples of carotenoids that are widely reported to possess beneficial biological activities (Figure 4).

β-Carotene is a member of the carotenes and is distinguished by its beta rings at the tail of both ends of the molecule. Like other members of the carotenes, it is a precursor to vitamin A. Reported health benefits of β-carotene include antioxidant activity, source of vitamin A, enhancement of immune and reproductive system function, and potential anti-cancer properties (Mutanen and Pajari, 2010). Lycopene is a polyunsaturated hydrocarbon, which lacks vitamin A activity that is distinct in carotenes (Britton et al., 2009). As lycopene does not convert to vitamin A, its chemical structure is unchanged and as a result, the conjugated double bonds are available to respond to free radical species leading to antioxidant effect (Özben, 2013, Rao and Rao, 2007). Several studies on lycopene have also suggested that it may have an effect against cardiovascular disease, diabetes, osteoporosis and cancer (Rao and Rao, 2007, Britton et al., 2009), although the evidence is still inconclusive. Lutein is another carotenoid that does not convert to vitamin A in the body. Studies have shown that lutein exhibits antioxidant capacity against oxidative stress and plays a role in reducing the risk of cataract and macular degeneration, which are two of the most common eye diseases in elderly people (Mares-Perlman et al., 2002).

A few studies have investigated carotenoids in Australian bushfoods. The most common carotenoids identified in Australian native plants are β-carotene, lutein and lycopene, as shown in Table 4. Konczak (2009) studied the lipophilic antioxidants in Australian native bushfoods and identified lutein in Australian limes, Kakadu plum, Davidson’s plum, Tasmannia pepper leaf, anise myrtle and lemon myrtle. The presence of these carotenoids in the former three fruits was confirmed by another study from the same group (Konczak and Roulle, 2011), who also investigated green and pink finger limes. The green finger limes were reported to contain 0.4 mg lutein/100 g, fresh weight (FW), which is substantially higher than the lutein levels in lemon and grapefruit juice (~0.01 mg/100 g, FW) and comparable to the levels of kiwifruit (~0.3 mg/100 g, FW) (Konczak and Roulle, 2011).

Lycopene and β-carotene were also examined in Australian native limes, plums and berries, yet none were detected in these fruits (Konczak, 2009), while only β-carotene was found in Tasmannia pepper leaf. Lycopene and β-carotene were reported by McDonald et al. (2006) in ground bush tomato, which contained 3.3 µg/100 g, dry weight (DW) for both carotenoids. This study also observed the effect of post-harvest handling methods on bioactive compounds, and it was found that after sun-drying and oven-drying, lycopene levels in bush tomato products were not detectable while anthocyanin also suffered a substantial loss. Similarly, Sommano et al. (2011) investigated the stability of carotenoids in commercial native plant food products during thermal processing. Their findings showed that heat processing of acidic food, including bush tomato and Kakadu plum sauce, does not reduce lycopene, β-carotene and ascorbic acid contents.

In addition to carotenoids, two other groups of lipophilic compounds, vitamin E and chlorophyll, in Australian native plants have also been investigated. Konczak (2009) examined the vitamin E components (α-, β- and γ-tocopherol) and chlorophyll levels (chlorophyll a and b) in Australian native fruits, herbs and spices. Overall, anise myrtle was found to have the highest α-tocopherol content at 49.4 mg/100 g, DW, followed by lemon myrtle (20.2 mg/100 g, DW) and Tasmannia pepper leaf (17.4 mg/100 g, DW). Anise myrtle also contains higher levels of γ-tocopherol than the other bushfoods. Low levels of β- and γ-tocopherol (<1.6 mg/100 g, DW) were found in the other bushfoods as listed in Table 4. Furthermore, anise myrtle contains substantially higher chlorophyll

32 levels than other herbs and spices. The lipophilic phytochemicals (α-, β- and δ- tocopherol and chlorophyll) in Australian native fruits were also studied by Konczak and Roulle (2011), with similar findings to the previous study conducted by Konczak (2009).

Isoprene unit

Carotene

Lycopene

Lutein

α-tocopherol

β-tocopherol

δ-tocopherol

γ-tocopherol Figure 4. Chemical structures of lipophilic compounds identified in Australian native plants.

Table 4. Lipophilic compounds identified in edible Australian native fruits, herbs and spices

Common name Main lipophilic References (Scientific name) compounds Herbs and spice Bush tomato Lycopene, β-carotene, α- (Sommano et al., 2011, McDonald et (Solanum centrale) tocopherol, β-tocopherol, γ- al., 2006, Konczak, 2009) tocopherol.

Tasmannia pepper leaf Lutein, β-carotene, α- (Konczak, 2009) (Tasmannia lanceolate) tocopherol, β-tocopherol, chlorophyll a, chlorophyll b.

Anise myrtle Lutein, α-tocopherol, β- (Konczak, 2009) (Syzgium anisatum) tocopherol, γ-tocopherol, chlorophyll a, chlorophyll b.

Lemon myrtle Lutein, α-tocopherol, β- (Konczak, 2009) (Backhousia citriodora) tocopherol, γ-tocopherol, chlorophyll a, chlorophyll b.

Wattle seed α-tocopherol, β-tocopherol, γ- (Konczak, 2009) (Acacia spp.) tocopherol. Fruits Australian Desert Lime Lutein, α-tocopherol, β- (Konczak and Roulle, 2011, (Citrus glauca) tocopherol, chlorophyll b. Konczak, 2009) Tasmannia pepper berry α-tocopherol, β-tocopherol, γ- (Konczak, 2009) (Tasmannia lanceolate) tocopherol. Finger lime (green) Lutein, α-tocopherol, β- (Konczak and Roulle, 2011) (Citrus australasica) tocopherol. Finger lime (pink) Lutein, α-tocopherol, β- (Konczak and Roulle, 2011) (Citrus australasica) tocopherol. Kakadu plum Lutein, α-tocopherol, β- (Konczak, 2009, Konczak and (Terminalia ferdinandiana) tocopherol, chlorophyll a, Roulle, 2011) chlorophyll b.

Davidson’s plum Lutein, α-tocopherol, β- (Konczak, 2009, Konczak and (Davidsonia pruriens) tocopherol, δ-tocopherol, γ- Roulle, 2011) tocopherol, chlorophyll b.

Davidson’s plum Lutein, α-tocopherol, β- (Konczak, 2009) (Davidsonia jerseyana) tocopherol, γ-tocopherol.

Lemon aspen α-tocopherol, β-tocopherol. (Konczak and Roulle, 2011, (Acronychia acidula) Konczak, 2009) Riberry α-tocopherol, β-tocopherol. (Konczak and Roulle, 2011, (Syzygium luehmannii) Konczak, 2009) Quandong α-tocopherol, β-tocopherol, δ- (Konczak and Roulle, 2011, (Santalum acuminatum) tocopherol, γ-tocopherol. Konczak, 2009)

2.3. Health related biological properties of edible Australian native plants

Plant polyphenolic compounds have received much attention in recent decades largely due to their diverse biological functions that are beneficial to health (Balasundram et al., 2006). With the increasing popularity of native fruits, herbs and spice in Australia, their bioactive functions have also attracted considerable interest in recent years. As a result, extensive studies have been dedicated to unravelling the health beneficial biological activities of edible Australian native plants. Table 5 lists the Australian bushfoods and their health-benefiting properties that have been investigated to date.

Table 5. Commercially available edible Australian native plants exhibiting health- benefiting properties

Common name Bioactive References (Scientific name) functions Herbs and spices Anise myrtle Antioxidant (Konczak et al., 2010b, Sakulnarmrat et al., 2013, Zhao (Syzgium anisatum) and Agboola, 2007) Antibacterial (Dupont et al., 2006) Metabolic (Sakulnarmrat and Konczak, 2012) syndrome Cytoprotective, (Sakulnarmrat et al., 2013) proapoptotic Bush tomato Antioxidant (Konczak et al., 2010b, Zhao and Agboola, 2007) (Solanum centrale)

Cut leaf Antioxidant (Zhao and Agboola, 2007) (Prostanthera incisa) Antibacterial (Dupont et al., 2006)

Lemon ironbark Antioxidant (Zhao and Agboola, 2007) (Eucalyptus Antibacterial (Dupont et al., 2006, Weerakkody et al., 2010) staigeriana)

Lemon myrtle Antioxidant (Konczak et al., 2010b, Zhao and Agboola, 2007, (Backhousia Sakulnarmrat et al., 2013) citriodora) Antibacterial (Dupont et al., 2006) Metabolic (Sakulnarmrat and Konczak, 2012) syndrome Cytoprotective, (Sakulnarmrat et al., 2013) proapoptotic Mountain pepper Antioxidant (Konczak et al., 2010b, Zhao and Agboola, 2007, (leaf & berry) Sakulnarmrat et al., 2013, Netzel et al., 2006, Netzel et (Tasmannia al., 2007) lanceolata) Metabolic (Sakulnarmrat and Konczak, 2012) syndrome Cytoprotective, (Sakulnarmrat et al., 2013) proapoptotic Antimicrobial (Weerakkody et al., 2010) 36

Table 5 (Continued). Commercially available edible Australian native plants exhibiting health-benefiting properties

Common name Bioactive References (Scientific name) functions

Herbs and spices (Continued) Strawberry gum Antioxidant (Zhao and Agboola, 2007) (Eucalyptus olida) Antibacterial (Dupont et al., 2006)

River/Native mint Antioxidant (Zhao and Agboola, 2007) (Mentha australis)

Fruits Australian desert Antioxidant (Konczak et al., 2010a, Konczak and Roulle, 2011, Zhao lime and Agboola, 2007) (Citrus glauca)

Brush cherry Antioxidant (Netzel et al., 2007) (Syzygium australe)

Burdekin plum Antioxidant (Netzel et al., 2006, Netzel et al., 2007) (Pleiogynium timorense)

Cedar bay cherry Antioxidant (Netzel et al., 2006, Netzel et al., 2007) (Eugenia reinwardtiana)

Davidson’s plum Antioxidant (Konczak et al., 2010a, Konczak and Roulle, 2011, (Davidsonia Netzel et al., 2007, Netzel et al., 2006) pruriens) Metabolic (Sakulnarmrat et al., 2014) syndrome Cytoprotective, (Sakulnarmrat et al., 2015) proapoptotic Finger Lime Antioxidant (Konczak et al., 2010a, Konczak and Roulle, 2011, (Citrus australasica) Netzel et al., 2007)

Illawarra plum Antioxidant (Tan et al., 2011b, Tan et al., 2011d, Zhao and Agboola, (Podocarpus elatus) 2007, Netzel et al., 2007, Netzel et al., 2006) Cytoprotective (Tan et al., 2011b) Proapoptotic (Tan et al., 2011c, Tan et al., 2011d) iNOS & COX-2 (Tan et al., 2011a) inhibitory Kakadu plum Antioxidant (Konczak et al., 2010a, Konczak and Roulle, 2011, (Terminalia Netzel et al., 2007, Tan et al., 2011b, Tan et al., 2011d) ferdinandiana) Cytoprotective (Tan et al., 2011b) Proapoptotic (Tan et al., 2011c, Tan et al., 2011d) iNOS & COX-2 (Tan et al., 2011a) inhibitory Magenta lilly pilly Antioxidant, anti- (Vuong et al., 2014) (Syzygium proliferative paniculatum)

Table 5 (Continued). Commercially available edible Australian native plants exhibiting health-benefiting properties

Common name Bioactive References (Scientific name) functions

Fruits (Continued) Molucca raspberry Antioxidant (Netzel et al., 2006, Netzel et al., 2007) (Rubus moluccanus)

Muntries Antioxidant (Tan et al., 2011b, Netzel et al., 2007, Netzel et al., (Kunzea pomifera) 2006) Cytoprotective (Tan et al., 2011b) Proapoptotic (Tan et al., 2011c) iNOS & COX-2 (Tan et al., 2011a) inhibitory

Native currant Antioxidant, (Tan et al., 2011b) (Acrotriche cytoprotective depressa) Proapoptotic (Tan et al., 2011c) iNOS & COX-2 (Tan et al., 2011a) inhibitory Quandong Antioxidant (Konczak et al., 2010a, Zhao and Agboola, 2007) (Santalum Metabolic (Sakulnarmrat et al., 2014) acuminatum) syndrome Cytoprotective, (Sakulnarmrat et al., 2015) proapoptotic Riberry Antioxidant (Konczak et al., 2010a, Konczak and Roulle, 2011, (Syzygium Netzel et al., 2007, Zhao and Agboola, 2007) luehmannii)

Vegetables and seeds Warrigal greens Antibacterial (Zhao and Agboola, 2007) (Tetragonia Antioxidant (Yang et al., 2008) tetragonioides)

Wattle seed Antioxidant (Konczak et al., 2010b, Ee et al., 2008, Ee et al., 2011, (Acacia spp.) Ee et al., 2012, Ee and Yates, 2013) Antimicrobial (Zhao and Agboola, 2007) Trypsin, α- (Ee et al., 2009b, Ee et al., 2008) chymotrypsin inhibitor

2.3.1. Antioxidant properties of polyphenolics in native fruits, herbs and spices 2.3.1.1. Antioxidation mechanisms of phenolic antioxidants

Antioxidants serve a crucial role in preventing or retarding oxidation from occurring in both the body and food products (Xu and Howard, 2012). In food, oxidation-induced reactions, such as lipid peroxidation, can cause deterioration of quality by generating undesirable chemical and provoking sensory changes in the product. In the body, these reactions may cause damage to vital tissues (e.g., cell membrane and blood vessels) and also interfere with physiological functions and processes, leading to various diseases. A slow, steady accumulation of oxidative species in the body may contribute to the development of a number of chronic diseases – inflammation, autoimmune, cardiovascular and neurological disorder, diabetes and cancer (Madhavi et al., 1996, Uttara et al., 2009). The general mechanism of lipid peroxidation is illustrated in Figure 5 as an example. The process proceeds via a free radical-mediated chain reaction, which is often initiated by an input of energy (e.g., light and heat) and facilitated by the presence of chemical species such as metal ions and the haem group (Shahidi, 1997). The hydroperoxide compounds generated during lipid peroxidation are unstable and further decompose into secondary products (e.g., hydrocarbons, ketones, aldehydes, alcohols and esters), some of which can be toxic to the body or produce off-odours and off-flavours in food (Madhavi et al., 1996). Lipid peroxidation can produce oxidative end-products from cholesterol which can rupture cell membrane and blood vessel tissues interfering with normal cellular activities, hence raising the risk of chronic disease and cancer development (Madhavi et al., 1996).

  Radical Initiation: RH + O2  R + OH

  Radical Propagation: R + O2  ROO ROO + RH  R + ROOH

ROOH  RO + OH

RO + RH  R + ROH  Oxidation products

Radical Termination: R + R  RR

  R + ROO  ROOR  Figure 5 . Mechanism of autoxidation ROO – involving + ROO three  ROOR steps: + initiation,O2 propagation, and termination. Note: R = side chain;  = free radical/reactive species (Madhavi et al., 1996). 39

In a biological system, oxidative damage can also occur as a result of accumulation of reactive oxidative species (ROS) in the system, which can be caused by a range of environmental factors such as UV stress, oxygen deprivation, invasion of pathogens and herbicide action (Blokhina et al., 2003). Continual accumulation of ROS is harmful to the body damaging lipids, proteins and DNA, which can eventually develop into various diseases, as illustrated in Figure 6.

High concentration of reactive oxygen species

Oxidative damage

Lipid Protein DNA

- chain breakage - site-specific amino acid - deoxyribose oxidation - increase in membrane modification - strand breakage fluidity and permeability - fragmentation of the - removal of nucleotides peptide chain - modification of bases - aggregation of cross- linked reaction products - DNA-protein crosslinks - altered electric charge - enzyme inactivation - increased susceptibility of proteins to proteolysis

Disease development

Figure 6. Damage to lipids, proteins and DNA induced by reactive oxygen species (Sharma et al., 2012).

Antioxidants, which have the capacity to prevent, retard or terminate oxidative reactions, may play a crucial role in human health (Shahidi, 1997). Research has found that plant phytochemicals such as polyphenolic compounds, which possess excellent antioxidation properties, could have a key role in mitigating oxidative damage in the

40 body. Studies have revealed that the presence of phenol groups in phenolic compounds is the main cause for their antioxidative effects (Balasundram et al., 2006, Uttara et al., 2009). Polyphenolic antioxidants exert their antioxidative effects through their capacity to scavenge free radicals, donate hydrogen atoms and electrons, or chelate singlet oxygen and metals (Marschner, 2012, Hvattum and Ekeberg, 2003, Gulluce et al., 2007). This disables the harmful effects of free radicals on lipids, proteins and DNA materials by ensuring that no new unpaired electrons are formed, thereby terminating the free radical chain reactions (Netzel et al., 2007). Further initiation of the chain reaction will not occur due to resonance stability of the phenolic ring (Dai and Mumper, 2010). Hence, the consumption of polyphenols rich fruits, vegetables and herbs may induce an antioxidant effect in our body, and thereby reduce the risk of cancer, cardiovascular diseases, diabetes and a number of other disorders associated with oxidative stress.

2.3.1.2. Phenolic compounds and their antioxidative effects in native plants

A number of studies have investigated the link between the phenolic content of native plants and their antioxidant capacity. The total phenolic contents in Australian bushfoods are reported to be higher than those of blueberries (4.5 mg GE/g, FW) and peppermint leaf (13.2 mg GE/g, DW) (Konczak et al., 2010b). Among the fruits, Kakadu plum (27.2 mg GE/g, FW), Burdekin plum (17.1 mg GE/g, FW) and mountain pepper berry (14.0 mg GE/g, FW) were found to be the most abundant in phenolic components (Netzel et al., 2006, Netzel et al., 2007), while mountain pepper leaf (102.1 mg GE/g, DW) and anise myrtle (55.9 mg GE/g, DW) were the richest of herbs and spices (Konczak et al., 2010b). Sakulnarmrat and Konczak (2012) also investigated the polyphenolic-rich fractions of Australian native plants after purification by adsorbent chromatography using a glass column packed with the XAD-16 resins.

The antioxidation activities of the phenolic fractions of the native fruits and herbs have been investigated by several groups using a number of different assays including the free radical scavenging (ABTS cation and DPPH), oxygen (peroxyl) radicals absorbance capacity (ORAC) and the ferric reducing antioxidant power (FRAP) assays (Table 6). Netzel et al. (2007) investigated the ABTS radical cation scavenging activity of

Australian native fruits and found that Kakadu plum and Burdekin plum exhibited the highest activity, followed by Cedar bay cherry, muntries, mountain pepper berry and Illawarra plum. An earlier study from the same group investigated the DPPH and FRAP antioxidant capacities in native fruits (Netzel et al., 2006). The results exhibited a similar trend to the aforementioned data set, with the Burdekin plum possessing the highest DPPH and FRAP values, followed by muntries. Interestingly, Konczak et al. (2010a) found that Kakadu plum exhibited a substantially higher ferric reducing activity than those previously reported for Burdekin plum and muntries. It was also revealed that the berry of mountain pepper exhibited the highest peroxyl scavenging activity, followed by quandong, Kakadu plum and lemon aspen. Australian limes, Davidson’s plum and riberry, in contrast, exhibited lower peroxyl scavenging activities.

The antioxidant capacity of Australian native herbs and spices was not as well studied as the native fruits (Table 7). Native herbs anise myrtle, mountain pepper leaf and lemon myrtle possessed comparable, if not higher, FRAP and ORAC values than those reported for native fruits Illawarra plum, muntries and native currant (Tan et al., 2011b). Out of all native fruits, herbs and spices, Kakadu plum exhibited the highest ferric reducing capacity with a FRAP value of 4538.4 μmol Fe2+/g, DW. Konczak et al. (2010b) investigated the ferric reducing power and peroxyl scavenging activities of anise myrtle, bush tomato, lemon myrtle and mountain pepper leaf. Among the herbs and spices, anise myrtle exhibited the highest FRAP (2158 μmol Fe2+/g, DW) followed by mountain pepper leaf (1315 μmol Fe2+/g, DW), while the latter exhibited a higher ORAC value (3504 μmol TE/g, DW) than anise myrtle (2446 μmol TE/g, DW). When these results are examined together, it is clear that the ferric reducing capacity of the native plants does not correspond to the peroxyl scavenging capacity. Indeed, correlation analysis showed that the level of phenolic compounds in native herbs was correlated with the peroxyl scavenging activity but not with the ferric reducing effect (Sakulnarmrat and Konczak, 2012). This suggests that different phenolic compounds may exert their antioxidant effects in different ways. Some compounds may be excellent free radical scavengers while poor iron reducers, while others may have the opposite effects.

Although reagent based assays are simple to implement and can quickly provide information on the antioxidant capacity of the plant extracts, these assays do not

42 necessarily reflect their physiological effect in the body. To provide better insight in this respect, some studies have investigated the cellular antioxidant capacity of Australian native fruits (Tan et al., 2011b) and herbs (Sakulnarmrat et al., 2013) by using HepG2 tissue cell line cultures. Using such methods, it was revealed that the Kakadu plum exhibited the highest cellular antioxidant capacity, followed by the Illawarra plum, native currant and muntries. As opposed to results of chemical based assays, the Illawarra plum exhibited much higher antioxidant activity than that of the native currant when assessed using the tissue culture assays.

Furthermore, results of tissue culture assays showed that native herbs had higher cellular antioxidant activities than native fruits. Of the native herbs studied, Tasmannia pepper leaf showed the highest activity, followed by anise myrtle and lemon myrtle (Sakulnarmrat et al., 2013). These findings generally correlated with those obtained from the reagent based ORAC assay with the activities showing the following order: Tasmannia pepper leaf > anise myrtle > lemon myrtle. However, the same trend was not observed for the FRAP results obtained by previous researchers (Konczak et al., 2010b).

Table 6. Antioxidant capacity in crude methanolic extracts of commercially edible Australian native fruits (fresh weight basis)

Common name Antioxidant Capacity1

(Scientific name) TPC ABTS DPPH FRAP ORAC References µmol µmol µmol µmol µmol GE/g TE/g TE/g Fe2+/g TE/g

Australian desert 10.8 34.8 44.9 (Konczak et al., 2010a) lime (Citrus glauca) Brush cherry 12.6 27 (Netzel et al., 2007) (Syzygium australe) Burdekin plum 100.5 192 27.1 283.4 (Netzel et al., 2006, Netzel et (Pleiogynium al., 2007) timorense) Cedar bay cherry 65.0 129.5 9.6 233.3 (Netzel et al., 2006, Netzel et (Eugenia al., 2007) reinwardtiana) Davidson’s plum 15.9- 36.5 3.2 49.3- 83.1 (Konczak et al., 2010a, (Davidsonia pruriens) 16.8 53.9 Netzel et al., 2007, Netzel et al., 2006) Finger Lime2 6.8 (g) 12.6 (g) 45.9 (g) (Konczak et al., 2010a) (Citrus australasica) 9.2 (p) 23.3 (p) 65.1 (p) 8.7 (r) 13.8 (r) (Netzel et al., 2007) 10.9 (y) 16.2(y) Illawarra plum 68.2 122.8 8.9 214.8 (Netzel et al., 2006, Netzel et (Podocarpus elatus) (864.2) (1111) al., 2007, Tan et al., 2011b) Kakadu plum 74.7- 204.8 691 315 (Konczak et al., 2010a, (Terminalia 160 (4538) (1817) Netzel et al., 2007, Tan et al., ferdinandiana) 2011b) Lemon aspen 10.8 14.0 132 (Konczak et al., 2010a) (Acronychia acidula) Molucca raspberry 21.9 45.1 5.3 66.6 (Netzel et al., 2006, Netzel et (Rubus moluccanus) al., 2007) Mountain pepper 82.5 123.2 11.8 186.7 779.5 (Konczak et al., 2010b, berry Netzel et al., 2007, Netzel et (Tasmannia lanceolata) al., 2006) Muntries 67.1 123.8 15.4 267.6 (Netzel et al., 2006, Netzel et (Kunzea pomifera) (520) (867) al., 2007, Tan et al., 2011b) Quandong 50.4 123 501 (Konczak et al., 2010a) (Santalum acuminatum) Riberry 7.5 28.1 33.2 49.9 (Konczak et al., 2010a, (Syzygium luehmannii) Netzel et al., 2007) Native currant (1256) (1402) (Tan et al., 2011b) (Acrotriche depressa) 1GE: gallic acid equivalent; TE, trolox equivalent; Fe2+: iron (II) equivalent; TPC: total phenolic content; FRAP: ferric reducing antioxidant power; ORAC: oxygen radicals absorbance capacity (hydrophilic compounds). Data expressed in dry weight basis are in brackets. 2Finger limes: green (g), pink (p), red (r) and yellow (y).

Table 7. Antioxidant capacity in crude methanolic extracts of commercially edible Australian native herbs, spices and seeds (dry weight basis)

Common name Antioxidant Capacity

(Scientific name) TPC ABTS DPPH FRAP ORAC Reference mg µmol µmol µmol µmol GE/g TE/g TE/g Fe2+/g TE/g

Anise myrtle 55.9 2158 2446 (Konczak et al., (Syzgium anisatum) 2010b) Bush tomato 12.4 206 913 (Konczak et al., (Solanum centrale) 2010b) Lemon myrtle 31.4 1225 1890 (Konczak et al., (Backhousia citriodora) 2010b) Mountain pepper 102.1 1315 3504 (Konczak et al., (Tasmannia lanceolata) 2010b) Wattle seed 0.8 17.8 53.4 (Konczak et al., (Acacia spp.) 2010b) 1GE: gallic acid equivalent; TE, trolox equivalent; Fe2+: iron (II) equivalent; TPC: total phenolic content; FRAP: ferric reducing antioxidant power; ORAC: oxygen radicals absorbance capacity (hydrophilic compounds).

2.3.2. Antimicrobial properties of Australian native plants

The use of food preservatives, chemicals with antimicrobial and antioxidant properties, is an integral part of the food industry’s overall strategy to combat pathogenic and spoilage microorganisms in food to ensure the products are safe and have an acceptable shelf-life. While chemicals such as nitrite/nitrate, sulphur compounds and organic acids are the most commonly used preservatives in food, the general public are demanding natural alternatives that are perceived to be less harmful to health and the environment (Gutierrez et al., 2008). Hence, natural antimicrobial substances derived from plants have attracted increasing interest from both the industry and research community. A large number of studies have explored the antimicrobial properties of plant essential oils and non-volatile extracts, and many of their components, e.g., terpenoids and flavonoids, were found to have significant antimicrobial activities (Dorman and Deans, 2000).

The mode of action of antibacterial activity of essential oils has been reviewed by Calo et al. (2015). Several mechanisms have been suggested, but only a few have been examined in detail. The structural and functional damage of microbial cell membrane 45 induced by the hydroxyl groups of phenolic constituents was suggested as one of the most dominant mechanisms. Essential oils act by penetrating the microbial cell wall, dissolving into the lipid layer of the membrane structure, and binding to the hydrophobic region. This in turn affects the permeability of the cell, resulting in the release of intracellular materials, such as ribose and sodium glutamate, which are essential for the survival of microorganisms (Toldrá, 2009). Essential oils can also affect membrane functions by interfering with electron transport, oxygen and nutrient uptake, synthesis of proteins and nucleic acids, and enzyme activity (Bajpai et al., 2007). Another proposed mechanism can be explained by alkyl substitution into the phenol nucleus or the aromatic ring of the phytochemical, resulting in the formation of phenoxyl radicals (Dorman and Deans, 2000). These radicals interact with isomeric alkyl substituents that are present on bacterial surfaces thereby inhibiting their growth.

The mechanisms of antimicrobial activity of flavonoids have also been reviewed and which are more complex than those of essential oils (Cushnie and Lamb, 2005). The B ring of flavonoids has been suggested to inhibit nucleic acid synthesis by intercalation or hydrogen bonding with the stacking of nucleic acid bases (Mori et al., 1987). The inhibition of cytoplasmic membrane function was another mechanism proposed, and the cause of which may be explained by two theories. The flavonoid may rupture the lipid bilayer and penetrate it, thereby disrupting the barrier function; or alternatively, it could cause membrane fusion, inducing leakage of intramembranous materials and aggregation (Ikigai et al., 1993, Burt, 2004).

The antimicrobial activity of native Australian plant essential oils and extracts against food-related microorganisms have been studied by some researchers, with agar disk diffusion (measuring the size of inhibition zone by the sample on the target organism) and minimum inhibitory concentration (MIC) being the most commonly used methods (Burt, 2004). As mentioned previously, Fulton (2000) investigated the essential oils of mint bush, cut leaf and Christmas bush, in which the former two mints had some degree of inhibitory activity against S. aureus and B. subtilis, with inhibition zones between 3 and 4 mm. Methanolic extracts of cut leaf showed larger inhibition zones (8 mm), though interestingly no inhibition was observed for water and hexane extracts (Zhao and Agboola, 2007). In this particular study, 18 native bushfoods were screened against a range of pathogenic and spoilage bacteria and spoilage yeasts. Some degree of

46 antimicrobial activity was observed in most Australian native plants, and methanol extracts were found to show much higher activities than water and hexane extracts. Overall, Tasmannia pepper leaf and berry exhibited the strongest activities, showing large zones of inhibition against pathogenic and spoilage bacteria. Tasmannia pepper leaf was the strongest against B. subtilis (16 mm), Listeria monocytogenes (15.8 mm), Vibrio cholera (14.7 mm), S. aureus (14.5 mm), and had an exceptionally high activity against Schizosaccharomyces octosporus (34.7 mm). Other native bushfoods with high antimicrobial activity were aniseed myrtle, quandong, eucalyptus oil and wild limes. However, eucalyptus oil inhibited the broadest spectrum of food spoilage bacteria (Zhao and Agboola, 2007). Weerakkody et al. (2010) studied the antimicrobial activities of lemon ironbark and Tasmannia pepper leaf in comparison to common herbs such as rosemary and oregano. Their study demonstrated that both native herbs have high antimicrobial activities, with lemon ironbark and Tasmannia pepper leaf being the most effective against L. monocytogenes (25.6 and 14.4 mm, respectively), followed by S. aureus (21.8 and 13.2 mm, respectively). Cock (2008) also reported that methanolic extracts of lemon myrtle had a moderate degree of activities against four food-related bacteria Aeromonas hydrophilia (8.3 mm), B. bacillus (7.6 mm), B. subtilis (7.6 mm) and P. fluorescens (7.6 mm). These findings were comparable with the results for A. hydrophilia (7.5 mm) and B. bacillus (7.7 mm) reported by Zhao and Agboola (2007), although the latter study found a lower inhibition (4.0 mm) against B. subtilis.

Dupont et al. (2006) also examined the inhibitory activities of five Australian native herb extracts obtained by three different solvents (water, ethanol and hexane). It was found that most native herbs could inhibit S. aureus at low concentrations, with anise myrtle showing the strongest effect (7.8 µg/mL) followed by strawberry gum (15.6 µg/mL) and cut leaf (15.6 µg/mL). Lemon iron bark also showed a strong inhibitory effect against P. aeruginosa and Sal. Enteritidis, exhibiting a minimum inhibitory concentration of 31.3 µg/mL for both organisms. Winnett et al. (2014) specifically studied the MIC of Tasmannia species (berry, leaf and peppercorn), which showed a broad spectrum of antimicrobial activity and low MICs against several food-related bacteria, such as S. aureus (77 µg/mL) and Bacillus cereus (93 µg/mL), suggesting its potential as a therapeutic agent. However, no activity against fungi (Candida albicans and Saccharomyces cerevisiae) was observed for Tasmannia berry and leaf, though

47 some inhibition was observed in peppercorn, a finding contradicting that of Zhao and Agboola (2007).

2.3.3. Inhibitory properties of Australian native plants on digestive enzymes

The traditional indigenous dietary pattern and active lifestyle played an important role in the health and well-being of the indigenous people (Bannerman, 2006). Prior to the introduction of westernised food, the carbohydrate in indigenous diet was mainly obtained from bush plants. Hence, their diet contained low carbohydrate, over half of which was derived from fruit and honey in the form of sugars (Brand-Miller and Holt, 1998). Indigenous dietary intake of carbohydrate is also low in glycaemic index, which means it is slowly digested, absorbed and metabolised. This results in a slower increase in blood glucose and hence insulin levels, thereby reducing the risk of diabetes and related diseases including coronary heart disease and obesity (Brand-Miller and Holt, 1998).

Metabolic syndromes are a group of increasingly prevalent chronic conditions in wealthy societies that can lead to diabetes, obesity and hypertension (Dandona et al., 2005). These conditions are related to an accumulation or over activity of digestive enzymes such as α-glucosidase, α-amylase and pancreatic lipase, which induce the breakdown of dietary starch and fats, respectively, into smaller components readily absorbed by the body. Normally, the breakdown of these dietary compounds is necessary to facilitate their absorption in the body. However, with the oversupply of carbohydrates and fat in the diet, excessive break down of starch and fats by digestive enzymes can induce nutrient imbalance in the body (e.g., the imbalance of glucose), glucose intolerance, dyslipidemia and obesity (Sakulnarmrat and Konczak, 2012). On the other hand, enzymes such as angiotensin-I converting enzyme (ACE) play a key role in the regulation of blood pressure. Over-expression of ACE could lead to high levels of angiotensin II with resultant hypertension.

Recent studies revealed that the consumption of native plants could contribute to lowering the risks of metabolic syndromes by inhibiting key enzymes involved (i.e. α- glucosidase, pancreatic lipase and ACE) (Sakulnarmrat and Konczak, 2012). These enzyme inhibitory activities have been suggested to be attributed to phenolic 48 compounds present in the plant food, which can interact directly with digestive enzymes. Furthermore, dietary polyphenols can also target the receptors that are involved in insulin signal transduction, hence playing a role in this aspect of the human physiology (Dandona et al., 2005). As illustrated in Figure 7, polyphenols can influence a number of different pathways that are related to the management of blood glucose in type 2 diabetes (Bahadoran et al., 2013). However, studies about the effects of Australian native plants on metabolic syndromes are limited and require further exploration.

Plant polyphenols

Inhibit α- ↓ gluconeogenesis ↑ insulin dependent Protect pancreatic β glucosidase & α- & glucose output of glucose uptake via cells against amylase the liver glucose transporter oxidative damage

(GLUT4) and apoptosis Inhibit intestinal ↑ glycogenesis & Na+ dependent glycogen content of Activate signalling Alleviate imposed glucose transporters the liver pathways pressure on β cell (SGLT1 & SGLT2) ↑ glycolysis & Regulate production glucose oxidation & secretion of insulin

↓digestion & Regulate Improve glucose Improve β-cell intestinal absorption carbohydrate uptake in muscle function and insulin of dietary metabolism cell and adipocytes action carbohydrate

Improve glucose homeostasis and insulin resistance

Figure 7. Influence of polyphenols on the management of blood glucose in type 2 diabetes (Bahadoran et al., 2013).

2.3.3.1. α-Glucosidase and α-amylase inhibitory effects of Australian native plants

α-Glucosidase, an exo-glucosidase found mainly in the brush border of the small intestine, is a key digestive enzyme that acts to break down starch in the diet into glucose for absorption and utilisation by the body. The enzyme is involved in cleaving disaccharides (maltose) and polysaccharides (starch, glycogen) molecules at the 1,4- alpha linkage yielding the monosaccharide glucose (Simpson et al., 2012). Glucose can be rapidly absorbed into the body and the accumulation of glucose in the blood stream after a starch-rich meal can lead to hyperglycemia (Kwon et al., 2006). Research has shown that repeated hyperglycemia over time can lead to insulin resistance, a key sign of type 2 diabetes (Asgar, 2013, Li et al., 2009). α-Amylase is another key enzyme involved in the digestion of dietary starch, releasing oligosaccharides that can be further digested by α-glucosidase to glucose (Gropper and Smith, 2012). Therefore, the inhibition of α-glucosidase and α-amylase is an important measure in the overall strategy for prevention and treatment of type 2 diabetes. Inhibitors of α-glucosidase and α-amylase, e.g., acarbose, are standard anti-diabetic drugs used in the treatment of type 2 diabetes. Studies have reported that plant-based phenolic compounds have inhibitory effect on these digestive enzymes (Tadera et al., 2006).

Inhibitory activities of culinary herbs and fruits against α-glucosidase and α-amylase have been reported, with the activities often associated with their phenolic content (Wongsa et al., 2012, Cazzola et al., 2011, El-Beshbishy and Bahashwan, 2012, McDougall et al., 2005, Apostolidis et al., 2006, Kwon et al., 2006). Total phenolic content was found to have significant correlation with both α-glucosidase and α- amylase inhibitory activities of culinary herbs and fruits (Wongsa et al., 2012). Different phenolic compounds were found to have a different effect on the inhibitory activities for these enzymes. Specifically, caffeic acid showed a stronger correlation with α-amylase inhibitory activity than p-coumaric acid, while both phenolic acids displayed good correlation with α-glucosidase inhibitory activities. Furthermore, quercetin was found to be a more potent α-glucosidase inhibitor (~80% inhibition) than the antidiabetic drug acarbose (40%) (Li et al., 2009). While myricetin (53.9%), quercetin (52.1%) and kaempherol (50.7%) exhibited inhibitory activities comparable to α-amylase (52.1%). Kwon et al. (2006) also investigated a range of phenolic compounds against α-glucosidase. It was found that with pH adjustment (pH 6.5-7.5), 50 caffeic acid, rosmarinic acid and catechin exhibited high responses (>80% inhibition). Lower inhibition was observed when pH was not adjusted (pH 3.5-4.5). However, some phenolic acids, such as hydroxybenzoic acid, protocatechuic acid and vanillic acid, respond more effectively under an acidic environment. This indicates that pH plays an important role in the enzyme inhibitory actions of certain phenolic compounds.

Some studies have investigated the inhibitory activity of Australian native fruits and herbs on carbohydrate-digesting enzymes. Sakulnarmrat and Konczak (2012) evaluated the α-glucosidase inhibitory effect of several polyphenolic-rich extracts of native herbs.

They reported that anise myrtle (IC50 = 0.30 mg/mL), lemon myrtle (IC50 = 0.13 mg/mL) and Tasmannia pepper leaf (IC50 = 0.83 mg/mL) exhibited a stronger effect than the reference material bay leaf (IC50 = 3.21 mg/mL), with lemon myrtle being the most potent inhibitor. Moreover, Sakulnarmrat et al. (2014) examined the polyphenolic- rich extracts of Davidson’s plum and quandong with regard to their response against α- glucosidase. Davidson’s plum and quandong effectively suppressed the activities of α- glucosidase, displaying an IC50 of 0.13 mg/mL and 0.39 mg/mL, respectively. Overall, however, research in this area is limited to only a few plants, and needs to be expanded to more native fruits, herbs and spices.

2.3.3.2. Lipase inhibitory effect of Australian native plants

Another enzyme that can induce metabolic syndrome is pancreatic lipase, which acts by hydrolysing dietary triglyceride fats and converting them into monoglycerides and fatty acids (Birari and Bhutani, 2007). Breakdown of triglycerides into their smaller components is necessary for them to be taken up into the body. The prolonged excessive absorption and storage of fats in adipocytes, as a result of high dietary fat content, can lead to chronic weight gain and obesity. Polyphenolic compounds extracted from plants, such as culinary herbs, fruits and vegetables, have been shown to exhibit pancreatic lipase inhibitory effects (Bustanji et al., 2010, Ikarashi et al., 2011). For example, phenolic acids (e.g., ellagic, ferulic and gallic acids) and flavonoids (e.g., luteolin, genistein, quercetin and kaempherol) have been reported to be inhibitors of pancreatic lipase (Sergent et al., 2012). Phenolic compounds in peanut shell extract, which contains coumarin derivatives and flavonoid glycosides, and mango leaf, which 51 contains phenolic acids, phenolic esters, flavan-3-ols and mangiferin, were also identified as strong inhibitors of pancreatic lipase (Birari and Bhutani, 2007). Pancreatic lipase inhibitory activities of Australian native plants have also been investigated by Sakulnarmrat and Konczak (2012) and Sakulnarmrat et al. (2014). Their studies revealed that native herbs are much more effective as lipase inhibitors than bay leaf

(IC50 = 6.3 mg/mL) while some were even more potent than blueberries (IC50 = 1.02 mg/mL). Tasmannia pepper leaf and quandong showed the highest response (IC50 = 0.6 mg/mL), followed by anise myrtle (IC50 = 1.6 mg/mL), Davidson’s plum (IC50 = 1.7 mg/mL) and lemon myrtle (IC50 = 2.5 mg/mL). In brief, studies on the lipase inhibitory effects of native Australian food plants have only been conducted on a small number of species with limited scope and depth in the subject, but most importantly, the role of phenolic compounds in lipase inhibition have not been investigated in most cases.

2.3.4. Angiotensin I-converting enzyme inhibitory effect of Australian native plants

The renin-angiotensin system plays a crucial role in blood pressure regulation (Kwon et al., 2006). The enzyme renin converts angiotensinogen into angiotensin I by cleaving the peptide bond between the leucine and valine residues on angiotensinogen, which is turned into angiotensin II by angiotensin I-converting enzyme (ACE) (Figure 8). Over- expression of ACE in the body can cause hypertension, which is a major risk factor for a number of debilitating and life-threatening diseases, including cardiovascular disease (Boschin et al., 2014). Plant phenolics, particularly flavonoids and proanthocyanidins, have been shown to exhibit properties that could inhibit ACE (Lacaille-Dubois et al., 2001). Kwon et al. (2006) compared the ACE inhibition effect of several Lamiaceae herbs and their phenolic constituents. Inhibition was observed only for some herbs: rosemary (90%), lemon balm (>80%) and oregano (~40%). No inhibition of this enzyme was found in sage and chocolate mint. Furthermore, the authors also examined the 18 phenolic compounds that are commonly found in these herbs. The majority of them did not show any inhibition and those that did exhibit inhibitory effects were not very potent in the activity: resveratrol (<25%), hydroxybenzoic acid (<20%) and coumaric acid (<3%). Even rosmarinic acid, which is one of the most common and abundant phenolic compounds across the mint family, did not show any inhibition.

Lacaille-Dubois et al. (2001) also reported no significant ACE inhibition by phenolic acids, though inhibition was observed from flavonoids: isoorientin (48%), isovitexin

(46%), procyanidin C1 (45%), (-)-epicatechin (34%) and isoquercitrin (32%). This indicates that not all phenolic compounds in herbs exhibit ACE inhibitory activities.

The ACE-inhibitory activities of Australian native bushfoods have been explored by some researchers (Sakulnarmrat and Konczak, 2012, Ee et al., 2012, Sakulnarmrat et al., 2014). The polyphenolic-rich extracts of the native fruit Davidson’s plum showed great potential to inhibit ACE with a 91.3% inhibitory effect (Sakulnarmrat et al., 2014). Moderate inhibition was observed from Tasmannia pepper leaf (29.6%), anise myrtle (25.9%), quandong (22.2%) and lemon myrtle (13.0%), most of which gave higher inhibition than bay leaf (13.9%) (Sakulnarmrat and Konczak, 2012, Sakulnarmrat et al., 2014). Ee et al. (2012) also studied the ACE inhibitory activity, but with the aqueous extract of wattle seeds, which gave an inhibition of 18.9%.

Angiotensin I- converting Renin enzyme (ACE) Angiotensinogen → Angiotensin I → Angiotensin II

↓

Increase blood pressure

↓

Excessive increase leads to hypertension

Figure 8. Renin-angiotensin system – the role of renin and ACE in the increase of blood pressure and potentially lead to hypertension.

2.3.5. In vitro studies on anti-inflammatory properties and related cancer chemopreventive properties of Australian native plants

Inflammatory response can be induced by oxidative stress, hence antioxidants, which can neutralise the reactive species, may play a significant role in reducing inflammation (Reuter et al., 2010). Phenolic components of herbs and spices have been found to be 53 partly responsible for their antioxidant activity and also attribute to mitigating the inflammatory responses produced by reactive species (Peter, 2012). However, inflammatory processes can also be caused by enzyme-related pathways, involving pro- inflammatory enzymes such as hyaluronidase, cyclooxygenase-1 (COX-1), cyclooxygenase-2 (COX-2) and inducible nitric oxide synthases (iNOS), and also their products, i.e., prostaglandin E2 (PGE2) and nitric oxide (NO) (Soberón et al., 2010, Tan et al., 2011d, Guo et al., 2014). Hence, the number of studies about inhibitory effect of plant extracts towards these inflammation-related enzymes and their products have gradually increased.

Hyaluronidase acts by hydrolysing hyaluronic acid, which causes degranulation of mast cells, inducing the release of the chemical mediators histamine and leukotrienes (Ippoushi et al., 2000). Hyaluronidase inhibitors can prevent this by suppressing the enzyme and thus prevent the degranulation process. In addition to allergic reactions, this enzyme is also involved in cancer metastasis (Moon et al., 2009). Thus, measurement of hyaluronidase inhibitory effect is a simple in vitro approach to determine the potential anti-inflammatory activity of phytochemicals and it has been used in a number of studies (Ito et al., 1998, Ippoushi et al., 2000, Moon et al., 2009, Soberón et al., 2010). Several studies have investigated the effect of certain flavonoids against hyaluronidase. Soberón et al. (2010) showed that quercetin glycosides exhibited a relatively high hyaluronidase inhibitory effect compared with several other flavonoids, including isoquercitrin, hyperoside and rutin. Ippoushi et al. (2000) evaluated 46 vegetables and herbs for their inhibitory effects against hyaluronidase. Their study found that herbs from the Lamiaceae family exhibited higher inhibitory effects than other vegetable and herb families. It was revealed that lemon balm possessed the highest effect and the authors identified rosmarinic acid as the main hyaluronidase inhibitor in lemon balm. Although no research has studied the inhibitory effects of Australian native plants on hyaluronidase, inhibition of other inflammatory-related enzymes and their products have been examined.

The inhibition of COX-1, COX-2 and iNOS expression and their products, PGE2 and NO, is not only involved in anti-inflammatory activities but also proposed to have the potential for chemoprevention (Tan et al., 2011a). Several studies on Australian native fruits, herbs and spices have been directed towards modulating inflammation

54 mechanism and evaluating their anti-proliferative and proapoptotic properties (Tan et al., 2011d, Tan et al., 2011a, Guo et al., 2014, Sakulnarmrat et al., 2015, Vuong et al., 2014). Among the native fruits, Kakadu plum, Illawarra plum, muntries and native currant have been explored by Tan et al. (2011a). Their findings demonstrated that

Kakadu plum was the most effective in inhibiting COX-2, iNOS, PGE2 and NO. Tan et al. (2011d) further investigated the anti-inflammatory effects as well as proapoptotic activities of purified polyphenolic fractions of Kakadu plum and Illawarra plum. The same group also studied the anti-proliferative activities of these fruits in human cell lines (Tan et al., 2011c, Tan et al., 2011b). Their findings demonstrated Kakadu plum’s potential to act on multiple pathways to prevent the occurrence of inflammation and related adverse health effects. Moreover, polyphenolic-rich native herb extracts of anise myrtle and lemon myrtle were shown to be effective inhibitors of COX-2 and iNOS, while Tasmannia pepper leaf had pronounced effect on COX-1. Anise myrtle, lemon myrtle and Tasmannia pepper leaf were also shown to reduce the proliferation of cancer cells (Sakulnarmrat et al., 2013).

2.4. Australian native mints from the Lamiaceae family selected for study in this thesis 2.4.1. Australian native mints from the Mentha genus

The Mentha genus (Lamiaceae family) comprises an estimated of 25-30 species distributed across Europe, Asia, Africa, North America and Australia (Lawrence, 2007, Dorman et al., 2003), with at least seven species endemic to Australia, all of which contain strong aromatic foliage that is uniquely mint. Due to their flavour, aroma and biological functions, the essential oils and extracts from Mentha are frequently used as flavourings in food, cosmetic, pharmaceutical, dental and insecticidal products and also for aromatherapy purposes (McKay and Blumberg, 2006). Three Mentha species, Mentha australis, Mentha diemenica and Mentha satureioides, were recorded for traditional medicinal uses by indigenous people in Australia (Table 8).

Table 8. Description of three Australian native Mentha species used by indigenous people for medicinal purposes.

Species Distribution1 Usage References

River/Native mint Headaches, coughs and (Lassak and McCarthy, (M. australis R. Br.) colds. 2001, Williams, 2010)

Slender mint Fever, diaphoretic, stomach (Lassak and McCarthy, (M. diemenica disorder, emetics, diuretics, 2001, Williams, 2010) Spreng.) menstrual disorder, respiratory disorder,

abortifacient.

Native pennyroyal Tonics, vitamin (Lassak and McCarthy, (M. satureioides R. deficiencies, blood 2001, Williams, 2010) Br.) impurifiers, analgesics, earache, coughs and colds,

stomach disorders, emetics, menstrual disorder. 1Distribution maps are obtained from The Atlas of Living Australia (www.ala.org.au) (ALA, 2015).

Currently, only Mentha australis R. Br., also known as river or native mint (Figure 9), is available commercially and is consumed as herb or used as garnish. This mint is widespread throughout Australia and is usually found by rivers and creeks. Decoction or inhalation of crushed leaves of M. australis was traditionally used to treat colds and coughs, while inhaling the crushed mint and chewing the stems relieved headaches. It was also used as an abortifacient by the indigenous people (Lassak and McCarthy, 2001, Williams, 2010). However, there has yet been any scientific investigation into the phenolic composition of Australian native Mentha species and information on their bioactive properties is also rather limited.

Plantae (Kingdom)

Lamiales (Order)

Lamiaceae (Family)

Mentha Prostanthera (Genus) (Genus)

M. australis P. rotundifolia (Species) (Species)

Figure 9. Classification of Australian native mints Mentha australis R. Br. and Prostanthera rotundifolia R. Br.

2.4.2. Phytochemicals identified from the Mentha genus and their potential biological activities

The main phytochemicals investigated in the common Mentha species are essential oil components (terpenoids, terpenes) and phenolic compounds (McKay and Blumberg, 2006, Arumugam et al., 2010). The chemical structures of common essential oils and phenolic compounds in mints are illustrated in Figure 10. Plant essential oils from mints, in particular, have been extensively studied as these volatile components contribute to various flavours and aromas (Hussain et al., 2010, Sulieman et al., 2011, McKay and Blumberg, 2006). The main components identified in essential oils of 57 spearmint, peppermint and water mint are carvone, menthol and cineole, respectively (Table 9). However, research has also been directed towards the non-volatile polar components, particularly phenolic acids and flavonoids, which are well-known bioactive compounds with numerous reported health beneficial functions (Mimica- Dukic and Bozin, 2008). In non-volatile mint extracts, phenolic acids (e.g., caffeic acid and rosmarinic acid) and flavonoids (e.g., luteolin, luteolin-O-glucoside, apigenin, isorhoifolin, thymonin, eriocitrin and naringenin-7-O-glucoside) were identified in most Mentha species (Dorman et al., 2003, McKay and Blumberg, 2006, Kosar et al., 2004, Guedon and Pasquier, 1994). The levels of rosmarinic acid have been found to be high in a majority of the Mentha species (Dorman et al., 2003, Wang et al., 2004).

The composition of other components such as pigments (chlorophyll and carotenoids) and ascorbic acid has also been investigated. Capecka et al. (2005) examined ascorbic acid and carotenoid composition of three species from the Lamiaceae family (peppermint, lemon balm and oregano). Their study showed that after drying, the ascorbic acid level reduced markedly for all species with more than 90% reduction in peppermint and lemon balm. In the same study, chlorophyll was found to decline by around 45-50% in the three herbs. Moderate levels of total chlorophylls and low levels of carotenoid were also found in spearmint (Arumugam et al., 2010).

The health benefits of essential oils and solvent extracts of common Mentha are well studied. The antioxidant activities of common mint essential oils and non-volatile extracts have been well documented (Dorman et al., 2003, Gulluce et al., 2007, Scherer et al., 2013, Uribe et al., 2015). Mint essential oils have also been shown to exhibit antibacterial and cytotoxic properties (Sharafi et al., 2010). Furthermore, there is evidence that the non-volatile constituents of Mentha spp. exhibit antimicrobial (Gulluce et al., 2007), α-glucosidase inhibitory (Wongsa et al., 2012), pancreatic lipase inhibitory (Marrelli et al., 2014), antiallergic (Inoue et al., 2002) and anti-inflammatory effects (Shen et al., 2011, McKay and Blumberg, 2006).

Compared with the relatively extensive studies on common mints, research on the phytochemical composition and biological activities of native Australian mints from the Mentha genus is rather limited. Qualitative screening of phytochemicals in aqueous and methanolic extracts of M. australis has been conducted (Wright et al., 2015). The methanolic extracts were found to be qualitatively rich in phenolics (water soluble and 58 insoluble), saponins and flavonoids, while moderately rich in cardiac glycosides, triterpenes, alkaloids and tannins. However, the identification and quantification of major phytochemicals in any of the Mentha Australian species is yet to be conducted.

Zhao and Agboola (2007) reported that M. australis exerted inhibition against pathogenic bacteria (Bacillus cereus, Clostridium perfringens, Staphylococcus aureus, Vibrio cholera, Acinetobacter baumannii), and spoilage yeast (Schizosaccharomyces octosporus). In the same study, M. australis was found to have a relatively high content of total phenolics (98.1 mg gallic acid equivalent (GE)/L), moderate β-carotene bleaching inhibition and DPPH radical scavenging activities compared to other native bushfoods. However, this study did not attempt to identify phenolic compounds or explore other biological properties of the herb.

Limonene Menthone Menthol (S)-carvone (R)-carvone

Menthyl acetate 1,8-cineole (-)-globulol Prostantherol Menthofuran

Caffeic acid Rosmarinic acid Luteolin Apigenin

Hesperidin Narirutin Cynaroside

Isorhoifolin Eriocitrin

Figure 10. Chemical structures of common essential oil constituents and phenolic compounds identified in the Mentha and Prostanthera species.

Table 9. Major phenolic compounds and essential oil constituents identified in common Mentha species

Mentha species Major compounds References

Spearmint Rosmarinic acid, luteolin, (Dorman et al., 2003, Arumugam et (M. spicata L.) cynaroside, apigenin, isorhoifolin. al., 2008, Zheng and Wang, 2001, Zheng et al., 2006) Carvone (50-67%), limonene (5- (Hussain et al., 2010, Soković et 14%), cineole (4-9%). al., 2009, Mkaddem et al., 2009, Scherer et al., 2013)

Peppermint Eriocitrin, narirutin, hesperidin, (Inoue et al., 2002, Zheng and (M. piperita L.) luteolin-7-O-rutinoside, isorhoifolin, Wang, 2001, Guedon and Pasquier, diosmin, rosmarinic acid, 5,7- 1994) dihydroxycromone-7-O-rutinoside.

Menthol (33-37%), menthyl acetate (Soković et al., 2009, Yang et al., (7-17%), menthone (13-21%), 2010) limonene (6.9%), menthofuran (6- 7%), cineole (6%).

Water mint Rosmarinic acid, cynaroside and (Dorman et al., 2003, Zheng and (M. aquatica L.) eriocitrin. Wang, 2001)

Cineole (27%), menthofuran (23%), (Morteza-Semnani et al., 2006) β-caryophyllene (13%), limonene (5%), germacrene D (5%), β-pinene (4%).

2.4.3. Australian native mints from the Prostanthera genus

Unlike the Mentha genus, Prostanthera is more exclusively Australian. This genus, which also belongs to the Lamiaceae family, is collectively known as mint bushes and consists of at least 100 species which are all endemic to Australia (Palá-Paúl et al., 2006). Their leaves are distilled for essential oil and dried as spice for flavouring. Traditionally, the indigenous people extracted oils from crushed leaves of Prostanthera to make ointments for medicinal and aromatherapeutic purposes (Collins et al., 2014). Ointments made from Prostanthera rotundifolia were applied to sores, skin diseases, aches and pains (Lassak and McCarthy, 2001), while those from Prostanthera striatiflora were used for respiratory infections, skin sores and malaise (Barr, 1993) (Table 10). However, majority of the Prostanthera mint bushes were ornamental (Lassak and McCarthy, 2001), including Prostanthera lasianthos (Victorian Christmas bush) and Prostanthera melissifolia (Balm mint bush), while Prostanthera incisa, in contrast, was used to provide flavour. Only P. rotundifolia (mint bush) and P. incisa (cut leaf) are currently available commercially as edible herbs.

Table 10. Distribution and usages of three Australian native Prostanthera species for medicinal and culinary purposes in indigenous community

Species Distribution1 Usage References

Mint bush Ointment to treat sores, (Lassak and McCarthy, (P. rotundifolia R. Br.) skin diseases, aches and 2001) pains

Striped Mintbush Respiratory infections, (Barr, 1993) (P. striatiflora F. Muell.) skin sores and malaise

Cut leaf Herbs used for flavour (Dupont et al., 2006, (P. incisa R. Br.) Agboola and Radovanovic-Tesic, 2002)

1Distribution maps are obtained from The Atlas of Living Australia (www.ala.org.au) (ALA, 2015). 62

2.4.4. Phytochemicals identified from the Prostanthera genus and their biological activities

The identification of essential oil components has been reported for several species of the Prostanthera genus, and most of the studies also investigated the antimicrobial and antioxidant properties of the oils (Table 9). The predominant component of the essential oils of P. rotundifolia, P. incisa and P. lasianthos is cineole, constituting >40% of the steam-distilled product (Fulton, 2000, Dupont et al., 2006, Palá-Paúl et al., 2006). Dellar et al. (1994) revealed three sesquiterpenes using NMR techniques as the major components in the oil of the mint bush family (Dellar et al., 1994, Collins et al., 2014). To date, there have been very few studies which reported the identification of the phenolic compounds in solvent extracts of Prostanthera as opposed to the essential oils. One study on Prostanthera melissifolia F. Muell identified apigenin, ursolic acid, martynoside, isomartynoside, verbascoside, isoverbascoside, betonyoside F and isobetonyoside F in the herb (Kisiel and Piozzi, 1999). Fulton (2000) studied the chemical profile of mint bush (P. rotundifolia), Christmas bush (P. lasianthos) and cut leaf (P. incisa) by thin layer chromatography (TLC), which indicated the presence of flavonoid and phenol carboxylic acid compounds. The results from TLC also showed the presence of chlorogenic acid, which was detected as weak bands. Methanolic extractions of mint bush (Prostanthera rotundifolia) revealed high levels of phenolic compounds particularly flavonoids, but identification of individual compounds was not conducted (Wright et al., 2015, Maen and Cock, 2015).

Some researchers have investigated the antimicrobial activities of essential oils and phenolic extracts of P. incisa (Zhao and Agboola, 2007). In this study, some degree of antimicrobial activity was observed from the phenolic extracts and essential oils against the bacteria Aeromonas hydrophila, Vibrio cholera, Acinetobacter baumannii and Bacillus subtilis, and also the yeasts Pichia membranifaciens and Schizosaccharomyces octosporus. The essential oils of P. incisa were also found to exert significant inhibition against S. aureus (Dupont et al., 2006). Fulton (2000) investigated the antimicrobial properties of the essential oils of P. rotundifolia, P. incisa and P. lasianthos over a five -month storage period. The activity was only observed from the former two mints against S. aureus and B. subtilis after five months. No activity was observed against Escherichia coli, Pseudomonas aeruginosa, Erwinia carotovora subs P. carotovora

63 from these herbs. In addition to antimicrobial properties, the antioxidant capacity has also been investigated for the Prostanthera species. Zhao and Agboola (2007) determined the antioxidant activity by using β-carotene bleaching and di(phenyl)-(2,4,6- trinitrophenyl)iminoazanium (DPPH) free radical scavenging methods. P. incisa was reported to exhibit a moderate β-carotene bleaching inhibition (67%), while the DPPH radical scavenging activity was low (12.4%).

Table 11. Phenolic compounds and essential oil constituents identified in Prostanthera species and their biological functions

Biological Plant species Major compounds References functions Mintbush Antimicrobial Cineole (38.2%), p-cymene (Fulton, 2000, Dellar et al., (P. rotundifolia) (3.4%), prostantherol 1994)

Christmas bush Cineole (64%), α-pinene (6- (Palá-Paúl et al., 2006, (P. lasianthos) 8%), β-pinene (9.2), B- Fulton, 2000) elemene (8.2%), p-cymene (6.5%)

Cut leaf Antioxidant, Cineole (58.6%), globulol (Dupont et al., 2006, Zhao (P. incisa) Antimicrobial (15.9%), terpinyl acetate and Agboola, 2007, Fulton, (5.4%) p-cymene (3.9%) 2000)

Australian desert Antimicrobial Prostantherol (75.5%), β- (Collins et al., 2014) plant gurjunene (8.7%) (P. centralis) Purple mint bush Cineole, p-cymene, ether cis (Southwell and Tucker, (P. ovalifolia) dihyroagarofuran 1996)

Balm mint bush Antimicrobial Prostantherol; apigenin, (Dellar et al., 1994, Kisiel (P. melissifolia) ursolic acid, martynoside, and Piozzi, 1999) isomartynoside, verbascoside, isoverbascoside, betonyoside F and isobetonyoside F

Torrington mint- Limonene (26.9%), α-pinene (Russell et al., 2001) bush (10.7%), verbenone (8.4%), α- P. staurophylla phellandren-8-ol (5.7%) 64

2.5. Methods for identification of polyphenolic compounds in plants 2.5.1. Sample preparation for chromatographic techniques

Studies on plant-derived food, including edible Australian native plants, have shown that many of them are rich in phenolic compounds that have significant potential health- enhancing properties. These bioactive compounds are most commonly extracted by solvents, such as methanol, ethanol, acetone, hexane, water or their combination. The choice of solvent is important as it impacts the yield and type of phenolic compounds extracted (Dai and Mumper, 2010, Robards, 2003). For instance, aqueous acetone extracts higher molecular weight flavonols efficiently, while aqueous methanol is more effective for lower molecular weight polyphenols (Dai and Mumper, 2010). In most native plant studies, phenolic compounds are obtained by solvent extraction usually with aqueous methanol. Zhao and Agboola (2007) have compared the extraction efficiency of hexane, methanol and water and the methanolic extracts were found to have the most phenolic compounds. In the extraction of anthocyanin-rich phenolic compounds, acidified organic solvents are commonly used. The acidified solvent aids in breaking down the cell wall of the plant material as well as dissolving the anthocyanins (Dai and Mumper, 2010). However, though with the use of acid phenolic compounds can easily be hydrolysed in the process. In the studies of anthocyanins in Australian native plants extraction was mostly carried out by 80% aqueous methanol in 0.1-1% hydrochloric acid (Konczak et al., 2010a, Konczak et al., 2010b, Netzel et al., 2006). Other solvents have also been used to extract phenolic compounds. For wattle seed flour, phenolic extraction was performed by 70% aqueous acetone (Ee et al., 2012). In most of the research on fruits, vegetables and grains, 80% methanol is the most commonly used solvent for the extraction of phenolic compounds (Capecka et al., 2005, Velioglu et al., 1998, Wojdyło et al., 2007).

With the complexity of components in plant matrices, crude extracts usually contain large amounts of non-phenolic substances such as carbohydrates, proteins, pigments and lipids. Hence, purification and/or concentration of the phenolic compounds is frequently necessary prior to analytical identification (Dai and Mumper, 2010). These steps will improve the sensitivity of techniques such as High Performance Liquid Chromatography (HPLC), which will also remove the non-phenolic impurities that could interfere during analysis (Pietrzyk and Chu, 1977). Purification is often conducted 65 by liquid-liquid partitioning or solid-phase extraction (SPE) using C18 cartridges or XAD Amberlite® resins (Pyrzynska and Biesaga, 2009, Dai and Mumper, 2010). Phenolic compounds are separated based on their differing polarities, often by sorbents such as silica-based material (e.g., C18 resin) and Amberlite® polymeric adsorbents (e.g., XAD-7 and XAD-16). This technique is rapid, sensitive and also economical and has been widely used in the isolation and separation of phenolic compounds. The Amberlite® polymeric resins have been found to be successful in the purification of polyphenolic compounds (Aehle et al., 2004), especially anthocyanins (Hurst, 2008). Two of these resins, XAD-7 (Tan et al., 2011b, Tan et al., 2011c, Tan et al., 2011a) and XAD-16 (Sakulnarmrat and Konczak, 2012, Guo et al., 2014, Sakulnarmrat et al., 2013) have been successfully used in the purification of phenolic fractions from the extracts of native Australian fruits, herbs and spices.

2.5.2. Chromatographic techniques for qualitative and quantitative analysis of essential oils and phenolic compounds

2.5.2.1. Gas chromatography coupled with mass spectrometry in native plant studies

The extraction process of essential oils in native fruits, herbs and spices were often carried out by steam distillation (Brophy et al., 1995, Brophy et al., 2004, Palá-Paúl et al., 2006, Southwell and Brophy, 1992), however organic solvents (dichloromethane) have also been used on the peel of native citrus fruits (Delort and Jaquier, 2009). Identification of plant essential oil constituents is usually conducted by Gas Chromatography-Mass Spectrometry (GC-MS), which is a proven technique that has been widely used due to its high accuracy, sensitivity and robustness compared to many other separation and identification techniques (Halket et al., 2005). This technique provides excellent resolution, which facilitates the separation of extremely complex mixtures of compounds (Skoog et al., 2007). In addition, GC-MS is advantageous over other analytical techniques due to its large, well-established mass spectra library (e.g., NIST and Wiley databases) for the identification of volatile compounds (Halket et al., 2005, Rai, 2016).

The most common ionisation technique used on GC-MS is electron impact (EI) (Proestos et al., 2006, Fiamegos et al., 2004, Zhang and Zuo, 2004), in which the sample molecular ion collides with fast electrons producing ions. The mass spectra obtained can be matched from the existing records in the library. EI is advantageous as it is highly standardised and reproducible across GC-MS systems (Rai, 2016). However, in some cases EI can produce unstable fragmentation as a result of excess energy being imported to the sample analyte (Hansen et al., 2011). This can produce complex molecular fragmentations, making it difficult for library search, or inducing complete breakdown of the fragment ions.

Although GC-MS is primarily used to identify volatile compounds, it can also be employed to analyse non-volatiles, while a chemical derivatisation procedure is often required (Proestos et al., 2006). Chemical derivatisation is a process which converts analytes in solution into molecules that can easily be volatilised and with improved thermal stability (Fiamegos et al., 2004). The most common base catalyst used in plant phenolic studies are pyridine, N,O-bis-(trimethylsilyl)-trifluoroacetamide (BSTFA) and N-(tert-butyldimethylsilyl)-N-methyltrifluoroacetamide (TBDMS) (de Rijke et al., 2006). This particular method has been used to derivatise phenolic compounds in a number of food samples including cranberry juice (Zhang and Zuo, 2004, Zuo et al., 2002), various liquid plant extracts (Proestos et al., 2006, Proestos and Komaitis, 2013) and culinary herbs (Fiamegos et al., 2004). In these studies, phenolic compounds are commonly derivatised by silylation reactions which replace the hydroxyl groups of the phenolic compounds with trimethylsilyl (TMS) (Halket et al., 2005). Derivatisation coupled with GC-MS has been suggested as a very good alternative to liquid chromatography techniques for identification of phenolic acids (Proestos and Komaitis, 2013).

Although silylated derivatisation coupled with GC-MS has been successful in identifying phenolic acids, its performance in regards to the identification of flavonoids is poor in general (Proestos et al., 2006). In the study of cranberry juice using this method, Zuo et al. (2002) could only identify phenolic acids initially, although their group did identify some flavonoids (catechin, quercetin and myricetin) in a following study (Zhang and Zuo, 2004). It has been suggested that GC-MS may not be suitable for the analysis of flavonoids due to the limited volatility of flavonoid glycosides

(Cuyckens and Claeys, 2004). Furthermore, TMS spectra are not available for a large number of phenolic compounds; hence there is a need for more research towards identifying silyl derivatives by this method.

2.5.2.2. Reverse Phase-High Performance Liquid Chromatography

Reverse phase-High Performance Liquid Chromatography (HPLC) coupled with detection by ultraviolet-visible absorption or, more commonly nowadays, Diode Array Detection (DAD) (also referred as Photodiode Array Detection (PDA)) has become one of the most useful and preferred analytical techniques in the separation and determination of plant polyphenolics (Ignat et al., 2011). It is often implemented for the initial analysis and identification of active components in plant matrices including fruits (Schieber et al., 2001), herbs (Guedon and Pasquier, 1994, Wang et al., 2004) and vegetables (Bergman et al., 2001, Nuutila et al., 2002). HPLC-DAD is especially useful in this regard as the DAD provides an absorption spectrum of a resolved peak, which can be diagnostic of the compound. In chromatographic techniques, the feature of the spectrum (i.e., the absorption bands) does not represent the whole molecule, rather only the chromophore is displayed on the chromatogram (Pietrzyk and Chu, 1977). The chromophore may be associated with certain functional groups within a molecule such as double bonds in carbonyl groups and aromatic rings (Field et al., 2012). Each group of phenolic compounds exhibits different and, sometimes, unique characteristics of absorption spectra based on their diverse functional groups. For instance, the detection of aromatic rings or conjugated structures of phenolic compounds is apparent as they exhibit strong absorbance in the ultraviolet region (Merken and Beecher, 2000). In the detection of flavonoids, two characteristic absorption bands arise from the B-ring (band

I) with absorption maxima (λmax) between 300-550 nm and from the A-ring with λmax between 240-285 nm (Merken and Beecher, 2000). Generally, hydroxybenzoic acids and flavanols have λmax at 280 nm, hydroxycinnamic acid and stilbenes are detected at 320 nm, flavonols at 370 nm, and anthocyanins at 520 nm (Konczak et al., 2010a). The spectral characteristics of common phenolic compounds are illustrated in Table 12.

However, HPLC-DAD as a technique of identification alone is frequently not sufficient when there are a large number of compounds that have similar chemical structures or 68 when isomers co-exist, both of which are frequently encountered in the study of plant matrices. Structurally similar compounds or isomers may produce similar or identical ultraviolet absorption spectra, which could be difficult to distinguish (Cappiello et al., 2008). In such circumstances, co-chromatography with a reference standard is often used, and necessary, to determine whether the target analyte co-elute with the reference standard forming an overlapping peak. However, the matrix effect of plant extracts may influence the elution of the compounds, introducing compounds which makes it difficult to identify (Ignat et al., 2011).

Another limitation of this technique is that the detection of chromophores by UV absorption only allows the deduction of the structural fragments in the molecule, but not the molecule as a whole (Field et al., 2012). This will only provide information on possible structural elements of a compound. The technique is particularly useful when known standards are available for comparison, but its value diminishes if reference standards are not accessible. The latter situation often occurs when identification of compounds is performed with a rarely studied matrix of which little or no prior information about its phenolic composition can be used as a reference point. For most plant samples, it may be necessary for the crude extracts to be purified in order to produce a simpler chromatogram for identification as purification eliminates unwanted compounds. Notwithstanding these improvements, the identification of complex unknown compounds in new or less well studied plant matrices, such as native Australian plants, usually requires more sophisticated techniques such as mass spectrometry or Nuclear Magnetic Resonance (NMR) Spectroscopy to provide direct and more detailed information on the chemical structure of the analyte.

Table 12. Characteristic ultraviolet spectral properties of common phenolic compounds

Compound Structure Spectrum λmax (nm) Hydroxycinnamic acid

Caffeic acid 242, 322

p-coumaric acid 310

Chlorogenic acid 246, 325

Flavanones

Naringenin 288

Hesperetin 285

Hesperidin 283

Flavones

Apigenin 26, 335

Flavonols

Quercetin 254, 350

Kaempherol 262, 342

Rutin 262, 342

2.5.2.3. Liquid chromatographic techniques coupled with mass spectrometry

Chromatographic techniques alone cannot provide detailed structural information for the identification of closely related phenolic compounds. HPLC is a useful approach for preliminary analysis of known compounds with available reference standards. However, as mentioned previously, plant matrices can be complex and frequently contain compounds that are difficult to distinguish by retention times and UV spectra. There is a need for detection by mass spectrometry, which enables the detection and identification of molecules based on their masses (Halket et al., 2005). This technique can separate two compounds with the same retention time by mass, which is unachievable by HPLC- DAD. Over the years, many different LC-MS variants have been developed for various analytical purposes. Of these, Liquid Chromatography-High Resolution Mass Spectrometry (LC-HRMS) and Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS) are often utilised for the analysis of polyphenols from different biological matrices to elucidate the chemical structure of unknown compounds (Justesen, 2000, Ignat et al., 2011). Furthermore, an advantage of LC-MS is that this instrument does not require the non-volatile compounds to undergo a derivatisation step, which is necessary for GC-MS (Halket et al., 2005).

In mass spectrometric analysis of phenolic compounds, as for many other types of analytes, it is important to consider the method of ionisation of the compounds, and whether detection is the best when the compounds are in the cation or anion form. As previously mentioned, EI is the main ionisation source employed on GC-MS (Proestos et al., 2006, Fiamegos et al., 2004). With the LC-MS method, however, the precursor ions of phenolic compounds are commonly generated by ionisation, namely electrospray ionisation (ESI) and atmospheric pressure chemical ionisation (APCI). Recently, an additional heater component has been added to the traditional electrospray. This Heated Electrospray Ionisation (HESI) has been found to be more efficient, which promotes solvent evaporation at a higher flow rate (Naushad and Khan, 2014). ESI or HESI can potentially ionise compounds with high polarity or high molecular mass, while APCI can detect compounds with moderate polarity and molecular mass (Halket et al., 2005). These molecules can be ionised into anions or cations in ESI or APCI. In the study of plant matrices, either positive or negative ion mode of ESI has been implemented in the analysis of phenolic compounds (Corradini et al., 2011), though 71

APCI has become more popular due to its better responses for some compounds in this mode.

The mass of a molecular ion and its fragmentation is useful in the identification of unknown or novel compounds. In LC-HRMS, the mass of an ion or its fragments can be determined to an accuracy of approximately 2 ppm of a mass unit (Krauss et al., 2010, Wu et al., 2012). Based on exact mass and isotopic patterns, HRMS can provide a range of molecular formulae which narrows the search for new compounds (Kaufmann et al., 2010). In particular, molecules with the same nominal mass yet different exact masses may also be distinguished with very high precision (Krauss et al., 2010). However, a limitation of LC-HRMS is that it cannot differentiate between compounds with the same molecular formula such as stereoisomers (Wu et al., 2012). Moreover, the detection limit is not as low as that of LC-MS/MS.

LC-MS systems often consist of one mass analyser such as an ion trap, quadrupole or time-of-flight (Halket et al., 2005). In the case of LC-MS/MS, a combination of multiple mass analysers is employed. In the LC-MS/MS triple quadrupole system, there are three compartments – two mass analysers (Q1 and Q3) and a collision cell that contains an inert gas (e.g. argon and nitrogen) (Pitt, 2009). The role of the first mass analyser (Q1) is to select and fragment the precursor ions measuring the mass-to-charge (m/z) of a compound (McMaster, 2005). This is followed by an unscanned Q2 quadruple which causes fragmentation of ions by inducing collisions with an inert gas. The fragments are then analysed by a second mass analyser (Q3), which further select the product ions that have been initially generated from the fragmentation of precursor ions in the first mass analyser (Pitt, 2009). LC-MS/MS is highly sensitive and selective due to the selected reaction monitoring (SRM) or multiple reaction monitoring (MRM) mode (Wu et al., 2012). The sensitivity of LC-MS/MS is largely due to the drastic reduction of noise, by 100-fold or more, while the specificity is enhanced by the scanning of both precursor and product ions (Polettini, 2006, Wu et al., 2012). However, as mentioned previously, the accuracy of mass obtained by LC-MS/MS is much lower than LC-HRMS.

The MRM mode of LC-MS/MS can differentiate two compounds of same molecular weight by identifying the different product ions generated from fragmentation (Wu et al., 2012). However, even though isomeric compounds display different fragmentation 72 patterns, reference standards may still be required for direct comparison in order to accurately differentiate isomeric compounds from each other. Nevertheless, isomers can be relatively easily distinguished without reference standards by Nuclear Magnetic Resonance (NMR) Spectroscopy (Cuyckens and Claeys, 2004), which specifically examines the interactions of proton (1H) and carbon-13 (13C) nuclei within a chemical structure as well as their arrangement in space.

The structure of plant phenolic compounds, i.e. phenolic acids and flavonoids, may range from simple molecules to complex high molecular weight polymer. When identifying these compounds in plants, tandem mass spectrometry has been found useful in providing structural information, identifying unknowns, detecting structurally related analytes that produce a common fragment ion and detecting target analyte with higher sensitivity (Pitt, 2009). Due to these benefits, tandem mass spectrometry has been utilised in plant phytochemistry in the identification of proteins (Peng et al., 2003, Andon et al., 2002) and phenolic compounds (Robards, 2003, Wu and Prior, 2005). LC- MS/MS on a Quantum triple stage quadrupole (TSQ) coupled with ESI ionisation mode was the main instrument used in identifying polyphenolic compounds in Australian native fruits, herbs and spices (Konczak et al., 2010a, Netzel et al., 2006, Tan et al., 2011b).

2.5.3. Nuclear Magnetic Resonance Spectroscopy in structural elucidation of phenolic compounds

Nuclear Magnetic Resonance (NMR) Spectroscopy is a powerful technique for elucidating the chemical structure of a compound (de Rijke et al., 2006). This method is often employed for structural confirmation when the reference standard is unavailable and a sufficient amount of pure analyte can be collected (Cuyckens and Claeys, 2004). Samples can be prepared by fraction collection using preparative HPLC, which isolates the compound in relatively large quantities. NMR does not only differentiate isomers and substitution patterns of molecular compounds, but it also provides structural information that cannot be acquired by GC-MS, LC-HRMS and LC-MS/MS (de Rijke et al., 2006). However, this instrument is expensive, has low sensitivity, and requires a significantly larger amount of near pure sample than mass spectrometry methods (Ignat et al., 2011).

In principle, the nuclei in each molecule or atom have a charge with a spin-like behaviour, or a half-integer spin, which generates a magnetic field. The NMR instrument measures the resonance frequency (chemical shift) of each nucleus (usually 1H or 13C nuclei) and their intensity of absorption (Field et al., 2012). On the NMR spectrum, the chemical shift of each signal, its coupling constant and the relative intensities are parameters used to determine the structure of a compound as each compound exhibits characteristic spectrum that is useful for identification (Nollet, 2004). In the identification of phenolic compounds, the general chemical shifts on the NMR spectra are ~7 ppm (aromatic rings), 3-5 ppm (sugar groups for flavonoid glycosides) and ~1 ppm (methyl groups).

Other than 1H and 13C NMR methods, two-dimensional NMR techniques have been used to determine structural properties of complex structures in various plant extracts containing phenolic compounds (Ignat et al., 2011). The main ones that have been used in food analysis include Heteronuclear Single Quantum Coherence (HSQC), Heteronuclear Multiple-Bond Correlation (HMBC) and Correlation Spectroscopy (COSY) (Dimitrios, 2006). NMR techniques have been found to be useful in identifying and elucidating the complex chemical structures of compounds as it provides unequivocal structural identification. However, milligram quantities of near pure compound are required. Although LC-MS can separate a complex mixture, only 74 tentative identification can occur. As a result, NMR and LC-MS are used as complementary or hyphenated techniques (Dai et al., 2009).

In recent years, NMR analysis on food and plant material has become increasingly popular. NMR techniques have been applied to identify polyphenolic compounds in a number of plant matrices, including herbs (Gulluce et al., 2015, Park, 2011), fruits (Shoji et al., 2003), vegetables (Bergman et al., 2001) and plant-based oils (Adhvaryu et al., 2000, Clark et al., 1997, Guillén and Ruiz, 2001). To date, no studies have been reported to use NMR techniques for identification of phenolic compounds in Australian native plants. As discussed in preceding sections, this method, especially used in conjunction with other instrumental techniques, can be a powerful tool in the identification of phytochemical components in plant matrices. Furthermore, NMR techniques provide unequivocal structures, which are beneficial for the identification of new plants, and thus should be applied to the study of phenolic compounds in Australian native plants.

2.6. Conclusion

Australian native edible plants (i.e., fruits, herbs, spices and seeds) are important sources of bioactive phytochemicals that have attracted increasing research attention for their health beneficial biological functions. In these studies, polyphenolic-rich extracts from native bushfoods were revealed to exhibit a number of health-promoting effects including antioxidation activity, inhibitory effects against digestive enzymes that are related to chronic diseases (diabetes and obesity), and also anti-inflammatory effects. However, research in this area is still in a relatively early stage. There are still a large number of native edible plants yet to be studied. Among them, Australian native mints from the Mentha and Prostanthera genus were two of the least studied herbs. Mints from the Lamiaceae family have been reported to contain high levels of phenolic constituents with a wide range of health beneficial properties, indicating that Australian native mint species are worthy subjects of systemic scientific investigation. Currently, very little research has been conducted to identify and quantified the phenolic compounds in any of the Australian native mint species. Furthermore, research conducted on the health related biological functions of these mints are also very limited 75 in scope. This PhD project is thus conceived to partly bridge this important gap of scientific knowledge.

Chapter 3 Materials and Methods

The general experimental approach taken to achieve the project aim is outlined in Figure 11.

3.1 Plant materials

Fresh plants (leaves and stems in bunches) of river mint (Mentha australis R. Br., Lamiaceae) and bush mint (Prostanthera rotundifolia R. Br., Lamiaceae) were purchased from Outback Pride, Adelaide, South Australia; while spearmint (Mentha spicata L., Lamiaceae), which was used as reference sample, was purchased from Sydney nursery in February and July, 2014. The mints (2 kg) from Outback Pride were harvested at commercial maturity and shipped by air freight the following day to our laboratory where they were frozen immediately and freeze-dried (-109°C, 0.015 kPa; Labogene ScanVac Coolsafe 110-4 Pro Freeze Dryer, Lynge, Denmark). The lyophilised plant materials were vacuum-packed (Vacumatic Pty. Ltd. Melbourne, Australia) and stored at -80°C until use.

3.2 Chemicals and reagents 3.2.1 Chemicals used for sample preparation

The major chemicals used for sample preparation, analytical grade methanol and Amberlite® resin (XAD-7) were from Sigma Aldrich (Sydney, Australia). Milli-Q water (Millipore, Sydney, Australia) was used in all experiments, unless otherwise stated.

3.2.2 Chemicals and reagents used for antioxidant capacity assays

Gallic acid, trolox, Folin-Ciocalteu phenol reagent, ABTS (2, 2’-azino-bis(3- ethylbenzothiazoline-6-sulphonic acid) diammonium salt), DPPH (di(phenyl)-(2,4,6- trinitrophenyl)-iminoazanium), potassium persulfate, sodium acetate trihydrate, TPTZ (2,4,6-tripyridin-2-yl1,3,5-triazine), iron(III) chloride hexahydrate, fluorescein, and AAPH (2,2’-azobis(2-amidino-propane)) were purchased from Sigma Aldrich, Sydney, Australia.

Phenolic compounds Ground freeze-dried Proximate analysis Mint composition (moisture, lipid, ash, (total phenolic & total flavonoid content) material protein, carbohyrate) Antioxidant capacity (ABTS, DPPH, FRAP, ORAC) In vitro activities Inhibition of enzymes Methanolic (α-amylase, α-glucosidase, pancreatic lipase, hyaluronidase) extraction Purification (purified extract) (crude extract) High Performance Liquid Chromatography- Photodiode Array detector (HPLC-PDA)

Gas Chromatography-Mass Spectrometry (GC-MS)

Identification and Liquid Chromatography High Resolution-Mass quantification Spectrometry (LCHR-MS)

Liquid Chromatography-Tandem Mass Spectrometry (LC-MSMS)

Nuclear Magenetic Resonance (1H, 13C NMR, HSQC, HMBC, COSY)

Figure 11. Outline of the experimental approach for the thesis. 78

3.2.3 Chemicals and reagents used for enzyme-based assays

All chemicals and reagents used were purchased from Sigma Aldrich (Sydney, Australia) and included: acarbose, p-nitrophenyl-β-ᴅ-glucopyranoside (P-NPG), dinitrosalicylic acid, sodium potassium tartrate tetrahydrate, dimethyl sulfoxide (DMSO), 4-methylumbelliferyl oleate, orlistat, hyaluronic acid, 4-dimethylaminobenzaldehyde (DMAB), α-glucosidase, α-amylase, pancreatic lipase and hyaluronidase.

3.2.4 Chromatographic and Nuclear Magnetic Resonance analyses

Methanol (HPLC grade) was purchased from Burdick & Jackson (Muskegon, MI, USA); glacial acetic acid from Ajax Finechem Pty. Ltd. (Sydney, Australia); acetonitrile (HPLC grade), ammonium formate, TMCS (trimethylchlorosilane), BSTFA (N,O-bis(trimethylsilyl)trifluoroacetamide) and pyridine from Sigma-Aldrich

Corporation (Sydney, Australia); and deuterated methanol CD3OD was purchased from Cambridge Isotope Laboratories (Andover, MA, USA).

Reference standards of phenolic compounds: rosmarinic acid, apigenin, naringenin and hesperetin were purchased from Aladdin Chemistry Co. Ltd (Shanghai, China), hesperidin from Zhengzhou Lion Biological Technology Co. Ltd (Zhengzhou, China) and p-coumaric acid, caffeic acid and chlorogenic acid from Sigma Aldrich (Sydney, Australia).

3.3 Preparation of phenolic extracts 3.3.1 Preparation of crude extracts

The steps used in the preparation of the crude mint extract are presented in Figure 12. Each freeze-dried plant sample was ground to pass 0.5 mm mesh using a rotor mill (Pulverisette 14, Fritsch GmbH, Idar-Oberstein, Germany) and extracted with aqueous methanol (80% v/v) at a ratio of 1:20 g/mL for 10 min with sonication and centrifuged (10 min, 3000 g). The pellet was re-extracted two more times and the supernatants were pooled. The crude methanolic solutions were filtered through 0.45 μm polytetrafluoroethylene (PTFE) membrane and stored at -20°C. Crude extracts were used for antioxidant analysis within three days. Whereas for enzyme inhibition 79 experiments, the crude extracts were further concentrated by vacuum rotary evaporation (37°C; Orme Scientific Ltd., Manchester, UK), freeze-dried and stored at -20°C prior to analysis. The crude yield was calculated as follows:

Weight of crude extract (g) Crude yield (%) = x 100 Weight of sample (g)

3.3.2 Preparation of purified extracts

Five grams (dry weight) of freeze-dried mint samples were extracted three times, each with 100 mL 80% (v/v) aqueous methanol. In the first extraction, the plant/solvent mixture was sonicated for 10 min and then stirred with a magnetic stirrer for 2 h (4°C), followed by centrifugation at 10,000 g for 15 min (Thermo Scientific Fiberlite, Sydney, Australia). The second and third extractions were carried out at 4°C for 2 h and overnight, respectively, without sonication. The supernatants were combined and the methanolic solvent was removed via vacuum rotary evaporation (40°C; Orme Scientific Ltd., Manchester, UK). The remaining aqueous extract was centrifuged to remove the residual.

The aqueous extract was purified by non-ionic aliphatic acrylic polymer adsorbent (Amberlite® XAD-7 resin) following the procedure of Sakulnarmrat and Konczak (2012). Amberlite® resin has the ability to adsorb non polar compounds from aqueous solutions and polar compounds from non-polar solvents, was packed into an open glass column (300 x 25 mm i.d.; Figure 12). Briefly, the aqueous extract (50 mL) was loaded onto the column and washed with Milli-Q water by gravity until the impurities had passed through (i.e. when only colourless water flows through the column). Phenolic compounds were eluted with 80% (v/v) methanol. Methanol was removed by vacuum rotary evaporation at 40°C and the remaining aqueous extract was lyophilised (-109°C, 0.015 kPa; Labogene ScanVac Coolsafe 110-4 Pro Freeze Dryer, Lynge, Denmark). This lyophilised powder represented the purified extract and was used for identification of phenolic compounds and characterisation of potential bioactive properties. The purified extract was stored at -20°C prior to analysis and homogenised by mixing prior to chemical analysis and bioassays. The purification yield was calculated as follows:

Weight of purified extract (g) Purification yield (%) = x 100 Weight of sample (g) 80

Ground freeze-dried material

Methanolic extraction (80% v/v)

Sonication & centrifugation (stirred 2 x 2 h & overnight at 4°C)

Crude extract

Rotary evaporation 37°C

Purification by reverse chromatography (Amberlite® XAD-7 resin)

Rotary evaporation 40°C

Freeze-dry

Purified extract

Figure 12. The preparation steps of the crude and purified extracts of native mint samples.

3.4 Proximate analysis of Australian native mints 3.4.1 Protein content

The protein content was determined based on the amount of nitrogen in the sample using a LECO TruSpec CN Analyser (LECO Corporation, St Joseph, MI, USA) according to the approved American Association Cereal Chemists (AACC 46-30, 2000) standard procedures (LECO, 2006). The instrument was calibrated prior to analysis. Briefly, the lyophilised plant material (0.2 g) was placed in the loading head and any atmospheric gases that have entered during sample loading were purged. The sample was placed into the hot furnace (950oC) and flushed with oxygen for very rapid and complete combustion, followed by a secondary furnace (850oC) for further oxidation and particulate removal. A thermal conductivity cell was used to determine the nitrogen content (N%). Protein in the mints was calculated from nitrogen content, where a conversion factor of 5 was used to convert nitrogen to protein content (Yeoh and Wee, 1994). The analysis was carried out in duplicates.

3.4.2 Lipid analysis

The crude lipid content was determined using a Soxhlet system according to the approved AACC (30-25, 2000) standard procedures. Briefly, freeze-dried mint samples (2 g, duplicates) were extracted with diethyl ether for 8 h at room temperature. The solvent was filtered and the mint samples were then dried in oven at 105°C until constant weight was reached. The crude lipid content was calculated based on percentage residue per gram of sample weight.

Weight of extract Crude lipid (%) = x 100 Weight of sample

3.4.3 Moisture determination

The moisture content was analysed according to the approved AACC method (44-01, 2000). Prior to moisture analysis, aluminium dishes were dried and stored in a desiccator. The moisture content of the fresh mints was analysed within three days upon arrival. Fresh mint samples (5 g, duplicates) were placed in the pre-weighed dish. The samples were placed in a preheated vacuum drying oven (Townson & Mercer Ltd, Croydon, England) for 24 h at 70°C (i.e. until constant weight). The moisture content on a wet basis (MCwb) was calculated as follows:

Wi-Wf MCwb (%) = x 100 Wi

where, Wi = weight of initial sample; Wf = weight of final sample.

3.4.4 Ash determination

The ash content was determined according to the approved AACC method (08-01, 2000). Prior to the analysis, aluminium dishes were dried and stored in a desiccator. Dried mint samples (2 g) were placed in the pre-weighed dish. Duplicates were placed in a preheated furnace for 5 h at 520°C. The ash content was calculated as follows:

Wi-Wf Ash (%) = x 100 Wi

where, Wi = weight of initial sample; Wf = weight of final sample.

3.4.5 Carbohydrate content

The carbohydrate content in the mint samples was determined by difference and calculated as follows:

Carbohydrate (%) = 100 – [Protein (%) – Lipid (%) – Ash (%)]

3.5 Assays of antioxidant capacity of Australian native mints 3.5.1 Total phenolic content (Folin-Ciocalteu) assay

Total phenolic content in mint extracts was determined using the Folin-Ciocalteu assay according to the procedure described by Dorman et al. (2003) with minor modifications. The methanolic crude mint extract was diluted (1:25) with phosphate buffer (75 mM, pH 7.2), while the lyophilised purified extract was first dissolved in phosphate buffer (1 mg/mL) and further diluted (1:25) in the same buffer. The diluted extracts and standards (30 µL) were mixed with Folin-Ciocalteu phenol reagent (140 µL) and the mixture was allowed to react for 5 min at room temperature in a 96-well microplate (Sarstedt Australia, Technology Park, SA, Australia). The reaction was terminated by the addition of 20% sodium carbonate (30 µL) and the mixture was shaken for 5 s. The mixture was left at room temperature for 2 h and the absorbance was measured at 765 nm on a SpectraMax M2 microplate reader (Molecular Device, Sunnyvale, CA, USA) against a phosphate buffer blank. A reagent control was analysed the same way except that the sample extract was replaced by phosphate buffer. Gallic acid solutions in the concentration range of 1-100 mg/mL in phosphate buffer were analysed the same way to produce a standard curve which was used to determine the equivalent amount of gallic acid that was present in the plant extracts (Appendix 1). Total phenolic content in the crude and purified extracts were expressed as milligram gallic acid equivalent (GE) per gram dry weight plant material (mg GE/g, DW). The Folin-Ciocalteu assay was independently repeated at least three times (n = 3).

3.5.2 Determination of total flavonoids

Total flavonoid content was determined using the aluminium chloride colorimetric method modified from Padmini et al. (2008). The methanolic crude mint extracts were diluted (1:4) with 80% (v/v) methanol solution, while the lyophilised purified extracts were dissolved (3 mg/mL) using the same methanol solution. Subsequently, the solutions of crude and purified extracts (20 μL) were mixed with 10% (w/v) aluminium chloride (20 μL), 1 M potassium acetate (20 μL) and Milli-Q water (180 μL) in a 96- well microplate (Sarstedt Australia, Technology Park, SA, Australia). The mixtures

84 were allowed to react for 30 min at room temperature. Absorbance was measured at 415 nm on a SpectraMax M2 microplate reader (Molecular Device, Sunnyvale, CA, USA) against a blank of 80% methanol solution. The reagent control was analysed the same way except that the sample extract was replaced by 80% (v/v) methanol solution. Rutin hydrate in the concentration range between 25-400 μg/mL prepared in 80% (v/v) methanol was used to produce a standard curve. Total flavonoids were expressed as milligram rutin hydrate equivalent (RE) per gram dry weight plant material (mg RE/g, DW) for the crude and purified extract. This assay was independently repeated at least three times (n = 3).

3.5.3 ABTS radical scavenging capacity assay

The ABTS (2,2'-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) radical cation scavenging capacity assay was based on Danh et al. (2012) with modifications. ABTS stock solution was prepared as follows: ABTS powder (19.2 mg) was dissolved in Milli-Q water (3.5 mL), followed by the addition of 8.17 mM potassium persulfate powder. The mixture produced a dark green colour and was left for 12-16 h at room temperature in darkness. The ABTS stock solution (1 mL) was diluted with Milli-Q water (39 mL) prior to analysis. The methanolic crude mint extract was diluted (1:15) with phosphate buffer (75 mM, pH 7.2), while the lyophilised purified extract was first dissolved (1 mg/mL) in phosphate buffer and further diluted with the buffer (1:10). Standard solutions of trolox in phosphate buffer solution with concentrations between 100-1000 µM were also prepared. For each sample extract and standard, 10 μL was added onto a 96-well microplate. ABTS solution (190 μL) was added to each well and gently shaken for 5 s. The mixture was allowed to react in darkness for 5 min and the absorbance at 734 nm was measured on a SpectraMax M2 microplate reader (Molecular Device, Sunnyvale, CA, USA) against a phosphate buffer blank. The reagent control was analysed the same way except that the sample extract was replaced by phosphate buffer. ABTS radical scavenging activity of crude and purified extracts was expressed as micromole trolox equivalent (TE) per gram dry weight plant material (μmol TE/g, DW). This experiment was independently repeated at least three times (n = 3).

3.5.4 DPPH radical scavenging capacity assay

The DPPH (di(phenyl)-(2,4,6-trinitrophenyl)iminoazanium) radical scavenging assay was conducted according to the procedure described by Danh et al. (2012) with some modifications. The activity was determined based on the ability of the antioxidants present in the plant extracts to act as radical scavengers towards the stable free radical, DPPH. The methanolic crude mint extract was diluted (1:25) with phosphate buffer (75 mM, pH 7.2), while the lyophilised purified extract (1 mg/mL) was dissolved in phosphate buffer and further diluted with the buffer (1:25). The extract or standard (40 μL) were mixed with methanolic solution of DPPH (160 μL, 0.2 mM) and the mixture was kept at ambient temperature for 30 min in darkness. Absorbance was then measured at 517 nm wavelength on the SpectraMax M2 microplate reader (Molecular Device, Sunnyvale, CA, USA) against a phosphate buffer blank. The reagent control was analysed the same way except that the sample extract was replaced by phosphate buffer. Gallic acid in phosphate buffer was used to produce a standard curve (10-100 µM). DPPH radical scavenging activity of crude and purified extracts was expressed as micromole gallic acid equivalent (GE) per gram dry weight plant material (μmol GE/g, DW). The DPPH radical scavenging assay was independently repeated at least three times (n = 3).

3.5.5 Ferric Reducing Antioxidant Power (FRAP) assay

The FRAP assay was conducted based on the method reported by Konczak et al. (2010b) with modifications. The FRAP reagent (10:1:1) was prepared by mixing acetate buffer (20 mL, 300 mM, pH 3.6) with TPTZ (2 mL, 10 mM) in hydrochloric acid and iron(III) chloride hexahydrate (2 mL, 20 mM). The methanolic crude mint extract (1:25) was diluted with Milli-Q water. The lyophilised purified extract (1 mg/mL) was dissolved in Milli-Q water and further diluted with water (1:25). Crude and purified extracts and standards (20 μL) were mixed with FRAP reagent (150 μL). The mixture was left in darkness for 8 min and absorbance was measured at 600 nm on a SpectraMax M2 microplate reader (Molecular Device, Sunnyvale, CA, USA) against Milli-Q water blank. The reagent control was analysed the same way where the sample extract was replaced with Milli-Q water. The FRAP assay was independently repeated 86 at least three times (n = 3). The reducing capacity of crude and purified extracts was expressed as micromole trolox equivalent per gram dry weight plant material (µmol TE/g, DW). A series of trolox standard solutions in the range of 25-1000 M were used to produce a standard curve.

3.5.6 Oxygen Radicals Absorbance Capacity (ORAC) assay

The ORAC assay was conducted based on Konczak et al. (2010b) with modifications. Crude extracts were dissolved (1:550) in phosphate buffer solution (75 mM, pH 7.0). Stock solutions of lyophilised purified extracts (1 mg/mL) were further diluted (1:350) with phosphate buffer solution (75 mM, pH 7.0). Standard solution of trolox (100 M) was prepared in the same phosphate buffer solution. The extracts and standards (25 μL) were mixed with fluorescein (150 μL, 70 nM) and AAPH (2, 2’-azobis(2- amidinopropane)dihydrochloride) (25 μL, 140 mM). The kinetics of the reaction was monitored on a FLUOstar OPTIMA plate reader (BMG Labtech, Durham, NC) (excitation wavelength, 485 nm; emission wavelength, 520 nm) over 120 min at 37°C. Antioxidant capacity of the crude was expressed as micromole trolox equivalent per gram dry weight plant material (µmol TE/g, DW). A calibration curve was constructed by plotting the calculated differences of area under the fluorescein decay curve between the blank and the sample for a series of trolox standard solutions in the range of 10-100 M. The control consisted of fluorescein only. The ORAC assay was independently repeated three times (n = 3).

3.6 Assays of enzyme inhibition activities by extracts of Australian native mints 3.6.1 α-Glucosidase inhibitory activity assay

The α-glucosidase inhibition assay was conducted according to Zhang et al. (2013) with modifications. All solutions were prepared freshly and set at room temperature before use. Samples of lyophilised crude (5 mg/mL) and purified (3 mg/mL) extracts were dissolved in sodium phosphate buffer (100 mM, pH 6.9) and diluted into a series of concentrations to determine the dose dependent response. Acarbose was used as a positive control and standard with a concentration range of 100-1000 µM. Sample

87 extracts and standard solutions (30 µL) were mixed with α-glucosidase (30 µL, 0.1 U/mL) in a 96-well microplate. The mixture was warmed to 37°C and kept for 10 min, followed by the addition of p-nitrophenyl-β-D-glucopyranoside (P-NPG) (60 µL, 0.375 mM) as the substrate and incubation of the mixture for 30 min at 37°C. The reaction was terminated with the addition of Na2CO3 (120 µL, 0.2 M) and the absorbance was measured at 405 nm on the SpectraMax M2 microplate reader (Molecular Device, Sunnyvale, CA, USA) against a blank which contained no enzyme. The control reagent was prepared the same way where the sample extract was replaced with sodium phosphate buffer. This assay was independently repeated three times (n = 3). The α- glucosidase inhibitory activity was expressed as per cent inhibition and calculated using the following equation:

(A -A )-(A -A ) Inhibition (%) = C CB S SB X 100 AC-ACB

where, AC = absorbance of control, ACB = absorbance of control blank, AS = absorbance

of sample, ASB = absorbance of sample blank.

3.6.2 α-Amylase inhibitory activity assay

The α-amylase inhibition assay was conducted based on Nampoothiri et al. (2011), with modifications. Dinitrosalicylic acid colour reagent was prepared by dissolving 472.6 mg dinitrosalicylic acid in Milli-Q water (20 mL), to which sodium potassium tartrate tetrahydrate (12 g), sodium hydroxide (8 mL, 2 M), and Milli-Q water (12 mL) were added. The mixture was gently heated and stirred constantly to dissolve all constituents and stored overnight in darkness. All other solutions were prepared fresh and set at room temperature before use. Samples of lyophilised crude (5 mg/mL) and purified (10 mg/mL) extracts were dissolved in sodium phosphate buffer (0.02 M, pH 6.9 with 0.006 M sodium chloride) and diluted into a series of concentrations to determine the dose dependent response. Acarbose solutions in the concentration range 25-400 µg/mL prepared in sodium phosphate buffer were used to produce a standard curve. Sample extracts and standards (100 µL) were mixed with starch solution (100 µL, 1% v/v in the same buffer) and warmed to 25°C for 10 min. This was followed by the addition of α- amylase (100 µL, 0.5 mg/mL, 200 U) and allowed to react for 10 min at 25°C. The

88 reaction was terminated with dinitrosalicylic acid colour reagent (200 µL) and incubated at 100°C for 5 min. The mixture was cooled to room temperature and 50 µL was transferred to microplate and diluted with Milli-Q water (200 µL). The absorbance was measured at 540 nm on the SpectraMax M2 microplate reader (Molecular Device, Sunnyvale, CA, USA) against a blank which contained no enzyme. The control was prepared the same way except that the sample extracts were replaced with phosphate buffer. This assay was independently repeated three times (n = 3). The inhibition of α- amylase was expressed as percent inhibition and calculated using the following equation:

(A -A )-(A -A ) Inhibition (%) = C CB S SB X 100 AC-ACB

where, AC = absorbance of control, ACB = absorbance of control blank, AS = absorbance

of sample, ASB = absorbance of sample blank.

3.6.3 Pancreatic lipase inhibitory assay

The pancreatic lipase inhibitory assay was conducted based on the method reported by Sakulnarmrat and Konczak (2012) with modifications. All solutions was prepared fresh and set at room temperature before use. Lipase enzyme was prepared by dissolving lipase enzyme (0.085 g) in McIlvaine’s buffer (pH 7.4). The mixture was centrifuged for 10 min at 10,000 g and the supernatant was collected for the enzymatic reaction. Samples of lyophilised crude (5 mg/mL) and purified (5 mg/mL) extracts were dissolved in McIlvaine’s buffer, which contained 0.2 M disodium phosphate and was adjusted to pH 7.4 with 0.1 M citric acid. The extracts were diluted into a series of concentration to determine the dose dependent response. Orlistat solutions in the concentration range 0.2-1.0 mg/mL were prepared in dimethyl sulfoxide (DMSO). Sample extracts and standard (50 µL) were mixed with 4-Methylumbelliferyl oleate (100 µL, 0.1 mM, in DMSO). This was followed by the addition of pancreatic lipase enzyme (50 µL, 0.085 g/mL) and allowed to react for 20 min at 37 °C. The reaction was terminated with hydrochloric acid (1 mL, 0.1 N) and sodium citrate (2 mL, 0.1 M). Fluorescence was measured at the emission wavelength of 460 nm and excitation wavelength of 320 nm on a Varian Cary Eclipse fluorescence reader (Agilent 89

Technologies Inc., Sydney, Australia) against a buffer blank containing no substrate. The control was prepared the same way except that the extract was replaced with buffer and DMSO solution. This assay was independently repeated three times (n=3). The inhibition of lipase was expressed as % inhibition and is calculated using the following equation:

(F -F ) - (F -F ) Inhibition (%) = 1 - C CB S SB X 100 (FC - FCB)

where, FC = Fluorescence of control, FCB = Fluorescence of control blank, FS =

Fluorescence of sample, FSB = Fluorescence of sample blank.

3.6.4 Hyaluronidase inhibitory activity assay

The hyaluronidase inhibitory assay was carried out by following the method reported by Muckenschnabel et al. (1998) with some modifications. Samples of lyophilised crude (5 mg/mL) and purified (5 mg/mL) extracts were dissolved in DMSO. The extracts were diluted into a series of concentration to determine the dose dependent response. Aliquots (13.5 µL) of samples were mixed with citrate-phosphate buffer (150 µL, 100 mM, pH 5). Bovine serum albumin solution (37.5 µL, 0.2 g/L in Milli-Q water), hyaluronic acid substrate (37.5 µL) and Milli-Q water (61.5 µL) were added to the extract. To start the reaction, hyaluronidase enzyme (37.5 µL, 100 U/mL) was added and the reaction mixture was incubated for 60 min at 37°C. The reaction was stopped with an alkaline solution (82.5 µL) prepared from borate solution (10 mL, containing

17.3 g H3BO3 and 7.8 g KOH in 100 mL Milli-Q water) and potassium carbonate solution (1 mL, 8 g K2CO3 in 10 mL Milli-Q water). The reaction mixture was immediately heated for 5 min, then cooled and transferred to microplate (90 µL). The colour reagent 4-dimethylaminobenzaldehyde (DMAB) solution (110 µL) was then added and incubated for 20 min at 37°C. The DMAB solution was prepared with DMAB (20 g) in concentrated HCl (25 mL) and glacial acetic acid (75 mL). The solution was diluted with four volumes of glacial acetic acid prior to use. The absorbance was measured at 590 nm on a SpectraMax M2 microplate reader (Molecular Device, Sunnyvale, CA, USA) against a blank, which contained no enzyme. The control was prepared the same way except that the sample extracts were replaced with DMSO. 90

This assay was independently repeated three times (n = 3). The inhibition of hyaluronidase was calculated as follows:

As - ASB Inhibition (%) = x 100 AC - ACB

where, AC = absorbance of control, ACB = absorbance of control blank, AS = absorbance

of sample, ASB = absorbance of sample blank

3.7 Identification and quantification of phenolic compounds in native mints

3.7.1 High Performance Liquid Chromatography-Photodiode Array Detector analysis

HPLC analysis was performed using a Shimadzu Prominence Ultra-Fast Liquid Chromatography (UFLC) system (Shimadzu, Japan), equipped with a SIL-20A HT autosampler, a DGU-20A5 degasser, two LC-20AD pump, a CTO-20A column oven, a SPD-M20A photodiode array detector, a CBM-20A bus module and a FC-10A fraction collector. The column was a Phenomenex (Sydney, Australia) Luna 5 µm C18(2) column (250 mm x 4 mm i.d.) maintained at 30°C. The HPLC elusion procedure followed that of Wang et al. (2004) with modifications. The gradient elution system consisted of 2.5% v/v aqueous acetic acid (solvent A) and 100% acetonitrile (solvent B). The solvents were filtered through a 0.45 µm PTFE membrane (Grace Davidson Discovery Sciences, Melbourne, Australia) before use. For analysis, the lyophilised purified extracts of the mint samples (1 mg), obtained as described in Section 3.3.2, were dissolved in methanol (1 mL) and 20 µL of which was injected into the HPLC system. The samples were monitored at 280 (hydroxybenzoic acids and flavanols), 320 (hydroxycinnamic acids) and 370 nm (flavonols). This experiment was independently repeated at least five times (n = 5).

For the analysis of M. australis, the flow rate was maintained at 1 mL min-1 for 34 min and the following gradient was applied: 23% B (0 min), 23% B, (6 min), 32.5% B (11.5 min), 34% B (15.5 min), 34% B (16.5 min), 36% B (17.5 min), 90% B (18.5 min), 90% B (21.5 min), 23% B (27 min), 23% (34 min).

For P. rotundifolia, the flow rate was maintained at 0.8 mL min-1 and a 34 min gradient was applied: 15% B (0 min), 15% B, (6 min), 29% B (8.5 min), 31.5% B (17.5 min), 90% B (19 min), 90% B (21.5 min), 15% B (27 min), 15% B (34 min).

3.7.2 Isolation of phenolic compounds by semi-preparative HPLC

The HPLC system used for fraction collection was the same as that for analytical HPLC described in Section 3.7.1. Fraction collection of the major peaks was performed using a Phenomenex semi-preparative column (Luna 10 µm C18(2) 100 Å, 250 x 15 mm) with the same gradient elution procedure as for the analytical HPLC, but the injection volume was increased to 100 µL and the flow rate to 4 mL min-1. The collected peaks were purged with nitrogen to remove solvents, lyophilised and stored at -20°C prior to NMR analysis.

3.7.3 Liquid Chromatography-High Resolution Mass Spectrometry analysis

Identification of phenolic components was initially carried out using high resolution mass spectrometry for accurate mass determinations, followed by tandem mass spectrometry. Chromatographic separation for LC-HRMS was performed on an Orbitrap LTQ XL system (Thermo Fisher Scientific, San Jose, CA) equipped with an Accela 900 UPLC system. A Phenomenex Synergi column (250 mm x 1 mm i.d., 4 µm Hydro-RP) was used and maintained at 45°C. The mobile phase consisted of 5 mM ammonium formate in water, pH 7.4, adjusted with ammonium hydroxide (solvent A) and 5 mM ammonium formate in 90% methanol, pH 7.4 (solvent B). The method was based upon Shen et al. (2011) with modifications. A flow rate of 150 µL min-1 was used with a gradient elution program: 20% B (0 min), 30% B, (2 min), 30% B (4.5 min), 45% B (5.5 min), 45% B (7 min), 75% B (12 min), 90% B (14 min), 20% B (14.1 min), 20% B (19 min). Furthermore, to obtain information on the molecular masses that corresponded with the peaks eluting from the HPLC-PDA analysis described in Section 3.8.1, LC-HRMS was also conducted using the same HPLC method with the same gradient elution conditions and column as those of the HPLC-PDA analysis. The injection volume was 20 µL for all analyses and each analysis was repeated at least

92 twice. Atmospheric pressure chemical ionisation (APCI) and heated electrospray ionisation (HESI) sources were used in both positive and negative modes to generate ions from the analytes in solution, recorded over the mass range of m/z 120-1000 (full scan mode) at a resolution of 60,000. The limit of detection (LOD) was set at the signal- to-noise ratio of ≥20. LC-HRMS was calibrated at the start of each run using positive and negative calibration solutions (as per instrument manual) on each day of analysis using direct infusion into the HESI source. The LOD for LC-HRMS was 0.625 ng rosmarinic acid. The experiment was independently repeated at least five times (n = 5).

3.7.4 Liquid Chromatography-Tandem Mass Spectrometry analysis

LC-MS/MS analysis was carried out on a Quantum TSQ Access triple quadrupole instrument (Thermo Fisher System, San Jose, CA), equipped with an Accela pump and in the multiple reaction monitoring (MRM) mode. The chromatographic and column conditions were identical to those described in the HPLC analysis section. APCI and HESI sources were used in both positive and negative modes to generate ions from the analytes in solution. Data was acquired in Q3 full scan mode and the product scans were performed on analyte ions of purified extracts and reference standards. Product ions of reference standards were directly infused using a syringe pump at 10 L min-1, and subjected to negative and positive HESI. The MRM conditions were optimised using the tuning software on Xcalibur 2.2. The injection volume for the extract samples was 20 µL and each analysis was repeated at least three times. Ions were measured over a mass range of m/z 120-700 over 34 min (full scan mode).

The instruments (LC-HRMS and LC-MS/MS) were optimised for sensitivity on both solvent and compound using LC tuning software. The Qual Browser feature on Xcalibur 2.2 software (Thermo Fisher Scientific, San Jose, CA) was used for analysis of mass spectrometry data. The LOD for LC-MS/MS was 0.25 ng of apigenin. This experiment was independently repeated at least five times (n = 5).

3.7.5 Identification of phenolic compounds by Gas Chromatography-Mass Spectrometry 3.7.5.1 Derivatisation of phenolic compounds

Due to the complexity in both composition and chemical structure, flavonoid glycosides in plants are difficult to identify. To aid the identification of flavonoid glycosides, they were first hydrolysed to cleave the glycosidic bonds between the aglycones and the sugar components, followed by identification of aglycones and sugars (if necessary). For the GC-MS procedure used in the present study, hydrolysis of flavonoid glycosides was performed on the crude extracts of native plant samples using 1% (v/v) hydrochloric acid (6 M) and was mixed carefully for 2 min (Proestos et al., 2006). Prior to GC-MS analysis, all sample solutions were derivatised in order to obtain the alkyl silylated derivatives. Derivatisation was carried out by treating the hydrolysed crude methanolic extracts of mints (200 µL) with BSTFA (N,O- bis(trimethylsilyl)trifluoroacetamide, 200 µL) containing 1% TMCS (trimethylchlorosilane anhydrous) and 800 µL of pyridine. The mixture was heated at 60°C for 1 h in the dark and the derivatised samples was filtered through 0.45 µm PTFE membrane filter then subjected to GC-MS analysis. Pyridine was evaporated prior to GC analysis. Standard solutions of phenolic compounds, namely p-hydroxybenzoic, chlorogenic, caffeic, p-coumaric, ferulic and trans-cinnamic acids, were derivatised in exactly the same way.

3.7.6 Compound identification by GC-MS

The GC-MS system used was a Thermo Electron Corporation (Waltham, MA) Trace DSQ mass spectrometer, equipped with a TR-5MS SQC capillary non polar column (15 m x 0.25 mm, i.d.; 2.5 μm film thickness). The GC-MS procedure followed that of Proestos and Komaitis (2013) with some modifications. The inlet temperature was set at 280°C while the oven temperature was initially at 70°C and then increased to 135°C at 2°C/min (held for 10 min) followed by an increase to 220°C at 4°C/min (held for 10 min) and finished at 270°C at 3.5°C/min (held for 20 min). To equilibrate for the next injection, a post run of 70°C for 10 min was carried out. The mass spectrometer was operated in the positive ionisation mode with ionisation energy of 70 eV. The mass

94 range was set at m/z 25-700 and ion source temperature at 280°C. Detection was performed in selective ion monitoring (SIM) and total ion current (TIC) modes. The peaks were identified using target ions mass spectra and retention times. Standard solution (1 μL) and sample extracts (1 μL) were injected into the GC column. Identification of phenolic compounds in the extracts was achieved by comparing the retention times and molecular and fragment masses with those of the standards and the mass spectra NIST 2.0 library (The National Institute of Standards and Technology, Gaithersburg, MD, USA) on the Xcalibur 2.2 software (Thermo Fisher Scientific, San Jose, CA). This experiment was independently repeated twice.

3.7.7 Structural assignment of phenolic compounds by Nuclear Magnetic Resonance Spectroscopy

The purified extracts and collected fractions of major peaks from the preparative HPLC were subjected to 1H NMR (600 MHz) and 13C NMR (150 MHz) analysis in deuterated methanolic (CD3OD) solutions. A 600 MHz Bruker NMR (Rheinstetten, Germany) spectrometer with a cryoprobe at room temperature was used to acquire the data. The 2- D NMR data sets Heteronuclear Single Quantum Coherence (HSQC), Heteronuclear Multiple-Bond Correlation (HMBC), and Correlation Spectroscopy (COSY) (Berger and Braun, 2004) were used for analysis of unknown components. Topspin 3.2 (Bruker, Rheinstetten, Germany) and Mnova 10.0 NMR (Mestrenova, Santiago de Compostela, Spain) software were used for data analysis. This experiment was independently repeated twice.

3.7.8 Quantification of phenolic compounds by HPLC-PDA

The quantification of phenolic components in the mint samples was carried out in the HPLC-PDA system with standard curves established using reference standards where they were available. For qualification of components where standard compounds were not available, gallic acid was used and the amount was expressed as milligram gallic acid equivalent per gram purified extract (dry weight; mg GE g-1). The calibration curves had good linearity (r2 > 0.99) over the measured range for all the standards. Quantification of each phenolic compound was independently repeated at least five 95 times (n = 5). The results of the main compounds are expressed in percentage and are calculated as follows:

Concentration of a compound % concentration = Total concentration of all major peaks

3.8 Statistical Analysis

Each experiment was independently repeated at least twice and each assay was conducted at least in duplicates. The mean and standard deviation of results were calculated based on at least three independent measurements (n = 3) for all analyses, except for proximate analyses where two independent measurements (n = 2) were used. One-way ANOVA, Tukey’s multiple comparison tests and correlation analyses were performed using Graphpad Prism 6 (Graphpad Software, CA, USA) software to assess the differences between the mean values of each sample. Correlation analyses were performed using the Pearson’s correlation coefficient (r). Results were considered significantly different when p < 0.05. A second sample of Australian native mints from an alternate season was tested and similar trend of results were obtained.

Chapter 4 Phenolic composition and in vitro biological activities of Mentha australis and Prostanthera rotundifolia

4.1. Introduction

Mints from the Mentha genus of the Lamiaceae family comprise a variety of aromatic plants, many of which are used as culinary or medicinal herbs. Mints are widely distributed in most parts of the world with an estimate of more than 30 species (Lawrence, 2007, Dorman et al., 2003). In comparison, the native Australian species are not well documented. Furthermore, the native Australian mint family is not limited to the Mentha genus alone. The Prostanthera genus, also a member of the Lamiaceae family, is also considered as a mint. This genus is uniquely Australian as all of the Prostanthera species are natively grown in Australia (Palá-Paúl et al., 2006). Similar to the Mentha genus, Prostanthera leaves were distilled for their essential oil and used as spice for flavouring. Presently, only Mentha australis, Prostanthera rotundifolia and Prostanthera incisa are commercially grown as edible herbs (Ahmed and Johnson, 2000, Zhao and Agboola, 2007).

Other than anecdotal evidence passed down by generations of Australian indigenous people, only a handful of scientific studies have investigated the health beneficial properties of the commercially grown native mint species. These mints are rich in essential oils with significant antimicrobial activities as demonstrated in the studies on M. australis (Zhao and Agboola, 2007), P. rotundifolia (Fulton, 2000, Dellar et al., 1994), P. incisa (Dupont et al., 2006, Zhao and Agboola, 2007, Fulton, 2000), P. lasianthos (Palá-Paúl et al., 2006, Fulton, 2000), P. centralis (Collins et al., 2014) and P. melissifolia (Dellar et al., 1994, Kisiel and Piozzi, 1999). Zhao and Agboola (2007) have also conducted antioxidant testing on two native mints – river mint (M. australis) and cut leaf (P. incisa). The study showed that crude methanolic extracts of M. australis exhibited higher DPPH radical scavenging activity than P. incisa – 66.8% compared to 12.4%, whilst the β-carotene bleaching activity of P. incisa was more superior than M. australis – 67% compared to 41.8%. However, biological properties of the non-volatile components of these mints have not yet been studied.

The non-volatile components of herbs usually comprise carotenoids and phenolic compounds, with the latter being the focus of this study. In addition to antioxidant activities, some plant polyphenolic compounds have also been shown to inhibit digestive enzymes such as amylase and lipase, thus potentially playing a role in mitigating the risk of type 2 diabetes and obesity by reducing starch and fat absorptions in the body. Moreover, mints have a history of being used for its anti-inflammatory effects and have been shown to reduce histamine and inhibit metabolites that induce inflammation (Park, 2011, Shen et al., 2011, Arumugam et al., 2008). It is well recognised that hyaluronidase plays a key role in inflammatory reactions and a major approach to reduce its effect is to block the action of this enzyme, which increases during chronic inflammation (Soberón et al., 2010). Hence, to provide an insight into the potential antidiabetic, anti-obesity and anti-inflammatory properties of the phenolic compounds present in Australian native mints, in vitro enzyme inhibitory assays were conducted. Although these in vitro assays do not provide direct evidence regarding their physiological effects, the information obtained from these tests could be used as preliminary proof of their potential health properties.

The objectives of this chapter were to determine the general phenolic composition of Australian river mint and mint bush and to screen them for potential health-benefiting in vitro effects. Antioxidant capacities of the mints were examined using four different assays and the inhibition of key digestive enzymes (α-amylase, α-glucosidase and pancreatic lipase) as well as the inflammation-related hyaluronidase was also investigated.

4.2. Results and discussion 4.1.1 Proximate composition of Australian native mints

Results of proximate analysis for protein, moisture, crude lipid, ash and carbohydrate of river mint and mint bush are given in Table 13. Values for spearmint and peppermint were obtained from the literature for the purpose of comparison. The moisture content of fresh mint reported at harvest is 75-85% (Uribe et al., 2015), and this was confirmed with the Australian mints. The proximate composition of river mint and mint bush showed similar characteristics to those of the common spearmint. River mint showed lower moisture content (75.8%) compared to mint bush (85.8%) and spearmint (86.0%) and a slightly higher protein, ash and carbohydrate content. The level of crude lipid appeared to be higher in both the Australian mints – river mint (1.7%) and mint bush (2.1%) – than that of spearmint (0.4%). Low crude lipid level of 0.24% was also reported for peppermint (Mentha piperita) (Uribe et al., 2015).

Table 13. Proximate analysis of native Australian mints (w/w, fresh weight)1

Plant species Protein Moisture Crude Lipid Ash Carbohydrate2

River mint 3.9% 75.8% 1.7% 3.0% 15.6% (Mentha australis) Mint bush 2.4% 85.8% 2.1% 0.7% 9.0% (Prostanthera rotundifolia) Spearmint3 2.3% 86.0% 0.4% 1.7% 9.6% (Mentha spicata) Peppermint4 3.4% 72.2% 0.2% 3.3% - (Mentha piperita) 1Protein converted by a factor of 5. Moisture (n = 3); all other data (n = 2). 2Carbohydrate calculated by difference: % carbohydrate = [100 – (protein – lipid – ash)]. 3Spearmint was used as a reference sample. Data obtained from Scherer et al. (2013). 4Peppermint results were obtained from Uribe et al. (2015).

4.1.2 Purification yield of Australian native mints

The type of solvent used in the extraction process of plant phenolic compounds is important. The extraction of phenolic compounds in leaves using various solvents has been investigated by a number of studies and 80% (v/v) aqueous methanol demonstrated a higher extraction yield of total phenolics and total flavonoids than other solvents (i.e. 80% ethanol, 80% acetone and distilled water) (Butsat and Siriamornpun, 2016, Sultana et al., 2009). Hence, 80% (v/v) aqueous methanol was selected for the extraction of the mint samples. The mint extracts were lyophilised and weighed to calculate the extraction yields (Table 14). The crude mint extracts contained components other than phenolic compounds, such as lipids, sugars and pigments. To facilitate analysis of their biological activities and identification of compounds present in them, crude mint extracts were purified by Amberlite® resin (XAD-7) to enrich the phenolic components. The purified extracts were also lyophilised and weighed to calculate the yield (Table 14). Overall, P. rotundifolia (mint bush) produced a higher yield (12.2%) than that of M. australis (10.6%) and M. spicata (10.1%), which gave similar crude yields. The mint bush also gave the highest yield of purified extract (6.7%). However, the river mint had a lower yield (4.3%) of purified extract than spearmint (5.8%), indicating that more plant materials were removed from this mint sample during purification.

Table 14. Extraction yield of crude and purified native Australian mints (%, DW)1

Plant species Crude extract Purified extract

River mint 10.6% 4.3% (Mentha australis) Mint bush 12.2% 6.7% (Prostanthera rotundifolia) Spearmint 10.1% 5.8% (Mentha spicata) 1Yield was expressed as per cent of the original dried plant sample (%, DW).

100

4.1.2.1 Recovery of phenolic components during purification

As phenolic compounds have different affinities to the XAD-7 column, it was expected that the recovery of a particular compound would vary to another after the purification process. To test this hypothesis, the recoveries of five pure phenolic standards, p- coumaric acid, syringic acid, caffeic acid, gallic acid and quercetin, were determined. It was found that p-coumaric acid, syringic acid and caffeic acid had a recovery of >70%. Gallic acid and quercetin had a much lower recovery at 46% and 45%, respectively. These results are generally comparable to those reported by Pyrzynska and Biesaga (2009), who investigated the recoveries of several phenolic compounds with an Amberlite® XAD-2 column. Relatively high levels of recovery (>80%) were obtained for kaempherol, p-coumaric acid, syringic acid, caffeic acid, and vanillic acid, while the recovery for quercetin was considerably lower at 54%. Although the use of XAD resins is simple, low cost, efficient and can purify a large volume of sample at one time (Soto et al., 2011), its performance may not be as effective for certain phenolic compounds (e.g. gallic acid) as shown in this study. Hence, the use of an alternative method e.g. C- 18 solid phase extraction may be better prior to phenolic quantification.

4.1.2.2 General discussion on the extraction and purification of phenolic compounds from the mint samples

As phenolic compounds are prone to degradation due to high temperature and oxidation, their extraction method and storage conditions must be carefully chosen to prevent the loss of analytes. The first cause of analyte loss is often degradation during storage of plant and extract samples, which can be largely reduced by storing the samples appropriately. In the present study, all samples containing phenolic compounds were stored at -20°C and kept in darkness to avoid oxidation; purified extracts were also purged with nitrogen gas prior to storage. For long term storage (>14 days), samples were vacuum-packed and stored at -80°C.

The choice of solvent can significantly affect the phenolic yield and composition. In plant studies, a number of solvents, such as methanol, ethanol, acetone, hexane or a combination of those, have been used for the extraction of phenolic compounds (Dai and Mumper, 2010). Aqueous solutions of methanol (e.g., 80% v/v) have been widely 101 used in the extraction of phenolic acids and flavonoids from fruits, cereals and vegetables including those in the Mentha and Thymus genus (Pereira and Cardoso, 2013), and hence this solvent was selected for this study. A sonication step was included, which facilitated the extraction process by producing mechanic shearing through sound waves and cavitation (Proestos and Komaitis, 2006), which helped rupture plant tissue cell wall and improve solvent penetration, thereby releasing the contents within. To minimise oxidation, the entire extraction process was carried out at 4°C.

Plant materials are largely composed of carbohydrates and, to a lesser extent, proteins, lipids and pigments (Dai and Mumper, 2010). The use of 80% (v/v) methanol enabled the extraction of these compounds as well as secondary metabolites such as phenolic compounds in the crude extracts. Due to the presence of these impurities, crude extracts usually are not suitable for direct analysis by LC-MS as they may contaminate the column or damage the instrument. Furthermore, as phenolic compounds are usually present at low concentrations in the crude extracts in comparison to macro components, such as carbohydrates and lipids, they are often difficult to be identified with GC-MS as the high levels of sugar and lipid peaks frequently dominate the mass spectra. Hence, for chemical identification using GC-MS, it is best to purify extracts using Amberlite® XAD resins or SPE to remove the non-phenolic compounds using water. Purifying the crude extracts resulted in a concentration of the polyphenolic-rich fractions, which would facilitate the identification of phenolic compounds, as well as characterisation of their biological activities.

4.1.3 Phenolic composition of Mentha australis and Prostanthera rotundifolia

Total phenolic content of crude and purified extracts of river mint and mint bush was determined using the Folin-Ciocalteu procedure (Table 15). When the phenolic contents of the three different herbs were compared, spearmint was found to always have the highest total phenolic content in both the crude and the purified extracts. This was followed by river mint, while mint bush always contained the least amount of phenolic components in both the crude and purified extracts. Furthermore, the differences in phenolic content between the two Mentha species were relatively small. In contrast, the 102 phenolic contents in the mint bush (Prostanthera rotundifolia) were considerably lower than both of the two Mentha species. When the phenolic content in the crude and purified extracts were compared, it was found that the crude extract always had higher phenolic content than the purified one, for all three plants. This was not surprising given that the crude extracts were expected to contain impurities such as sugars, proteins, lipids and pigments, some of which could also react with the Folin-Ciocalteu reagent. This is because the Folin-Ciocalteu reagent is not specific for phenolic substances; rather, almost any reducing substances can react with the reagent and shift the absorbance upwards (Popova et al., 2004). Thus, the phenolic content in the crude extracts, as measured by this assay, was in fact a combined result of both the phenolic content and the impurities that have reducing capacities. When the crude extracts were purified, most of the impurities were removed, which would result in a reduction of the “measured” phenolic content in the purified extract. Furthermore, as shown in Section 4.1.2.1, the recovery of phenolic components during the purification process was not 100%, and for some compounds such as gallic acid and quercetin, the recovery was less than 50%. The partial loss of phenolic compounds during purification would also lead to a decrease in the total phenolic content in the purified extracts.

The crude extract of mint bush contained a higher level of total flavonoids (24.8 ± 1.2 mg RE/g, DW) than those of the Mentha species (Table 15). This is particularly true for the river mint, which had a total flavonoid content of 14.7 ± 0.6 mg RE/g for its crude extract. The trend remained the same with the purified extracts where the mint bush had the highest total flavonoid content, while the values for the two Mentha species were statistically the same. Surprisingly, the purified extracts showed a higher flavonoid content than the crude for all three herbs. Considering the fact some phenolic components were partially lost during purification, this result was probably due to interference of impurities in the crude extracts with the flavonoid assays. Such interference has been reported previously where it was shown that interference with the aluminium chloride reagent caused a significant bathochromic shift (Joubert et al.,

2008).

Shahrbandy and Hosseinzadeh (2007) reported a total phenolic content of 84 mg GE/g (DW) for the crude extract of spearmint, which is consistent with the result of the present study for this herb. Total phenolic contents reported for other mint species are

103 generally between 34.5 and 58 mg GE/g (DW) (Gulluce et al., 2007, Shan et al., 2005, Kırca and Arslan, 2008). These values are broadly comparable with that of the mint bush but lower than that of Australian river mint and spearmint With regards to total flavonoids, the two native Australian mints, especially the mint bush, had a higher content than another common herb, rosemary (14.7-17.5 mg RE/g, DW) as reported by Yosr et al. (2013).

104

Table 15. Phenolic content in crude and purified extracts obtained from native Australian mints and reference sample of spearmint1

Total phenolics Total flavonoids

Plant species (mg GE/g, DW) (mg RE/g, DW) Crude Purified Crude Purified

River mint 76.3 ± 3.8a 19.7 ± 0.2a 14.7 ± 0.6a 25.6 ± 1.2a (M. australis) Mint bush 42.1 ± 4.3b 16.4 ± 0.5b 24.8 ± 1.2b 29.2 ± 1.5b (P. rotundifolia) Spearmint 83.1 ± 5.8a 22.8 ± 0.3c 20.5 ± 1.0c 24.6 ± 1.5a (M. spicata) 1Results are expressed as milligram gallic acid equivalent (GE) and rutin hydrate equivalent (RE) per gram dry weight of plant material. The data represent the mean ± standard deviation of at least three independent experiments. Means with different letters in the same column were significantly different with p < 0.05 (n = 3).

Different levels of total phenolic and flavonoid contents reported in various studies could be attributed to a number of factors, including differences in genotype within species, environmental factors (geographical and climatic conditions), different parts of the plant tested, sample preparation, collection time, drying conditions and the extraction method used (Shan et al., 2005, Mimica-Dukic and Bozin, 2008). These factors can result in significant differences in phenolic and flavonoid contents even in the same plant species. For example, in a study where analysis was conducted on 23 accessions of basil, it was found that the total phenolic content ranged from 22.9 to 65.5 mg GE/g (DW) (Javanmardi et al., 2003). Furthermore, methods used in measuring the total phenolic and flavonoid contents can also significantly affect the results. Since extracts of biological matrices, such as plant materials, almost always contain impurities, their influences on the assays need to be quantified, but rarely considered in most studies. This remains a deficiency in the quantification of phenolic compounds by colorimetric methods, which should be addressed in future studies.

105

4.1.4 Antioxidant capacity of Mentha australis and Prostanthera rotundifolia 4.1.4.1 ABTS free radical scavenging assay

ABTS radical cation scavenging is a commonly used method for measuring antioxidant activities of phytochemicals. The assay is based upon the capacity of phytochemicals in the sample to transfer either electrons or hydrogen atoms to the ABTS radical cation with the resultant formation of a stable compound (MacDonald‐Wicks et al., 2006). Overall, the Mentha species showed higher ABTS radical cation scavenging activities than mint bush, as shown in Table 16. The crude extract of river mint (398.5 ± 19.3 µmol trolox equivalent (TE)/g, DW) exhibited a comparable antioxidant activity to that of spearmint (403.5 ± 14.8 µmol TE/g, DW), with no significant difference (p > 0.05) in the measured values between the two mint samples. In comparison, the crude extract of mint bush demonstrated a significantly lower ABTS radical scavenging activity of 262.4 ± 2.2 µmol TE/g (DW). For the purified extracts, spearmint exhibited the highest activity (134.8 ± 2.1 µmol TE/g, DW), followed by river mint and mint bush (129.9 ± 6.4 and 75.5 ± 9.4 µmol TE/g, DW, respectively). The values of river mint and spearmint were not significantly different (p > 0.05). As mentioned previously, the crude extracts contained sugars, lipids and pigment, as well as phenolic compounds, which would contribute to the antioxidant activity (Van den Ende and Peshev, 2013, Konczak, 2009). As a result, it is expected that the purified extracts would show a lower antioxidant activity than that of the crude extracts because most of the impurities were removed by purification and recoveries are not 100%.

Previous studies have reported a range of trolox equivalent antioxidant capacity (TEAC) values for phenolic extracts of Mentha species. For example, crude phenolic extracts of M. piperita exhibited a TEAC value of 461 µmol TE/g (DW) (Kırca and Arslan, 2008), which was higher than those in this study while lower TEAC values were reported for M. canadensis (338.3 µmol TE/g, DW) (Shan et al., 2005) and M. longifolia (265.6 µmol TE/g, DW) (Kırca and Arslan, 2008). Shan et al. (2005) also reported that ABTS radical scavenging activities of a number of common herbs were lower than those obtained for the mint samples in the present study, such as rosemary (378.0 µmol TE/g, DW), basil (295.9 µmol TE/g, DW), coriander (70.2 µmol TE/g, DW), dill (63.6 µmol TE/g, DW) and parsley (63.1 µmol TE/g, DW).

106

Table 16. Antioxidant capacity of crude and purified samples (per gram dry plant material) obtained from native Australian mints and reference sample of spearmint1

ABTS DPPH FRAP ORAC Plant species (µmol TE/g, DW) (µmol GE/g, DW) (µmol TE/g, DW) (µmol TE/g, DW)

Crude Purified Crude Purified Crude Purified Crude Purified

River mint 398.5 ± 19.3a 129.9 ± 6.4 a 178.9 ± 8.4 a 107.5 ± 1.0a 1933 ± 63a 690.3 ± 17.1a 1727 ± 184a,c 1011 ± 56a (M. australis)

Mint bush 262.4 ± 2.2b 75.5 ± 9.4 b 110.2 ± 9.0 b 62.2 ± 4.6b 606 ± 28b 337.4 ± 5.7b 1289 ± 82b 584 ± 65b (P. rotundifolia)

Spearmint 403.5 ±14.8a 134.8 ± 2.1 a 168.2 ± 6.9 a 145.3 ± 4.1c 1859 ± 9c 537.5 ± 16.3c 1551 ± 137b,c 918.1 ± 46c (M. spicata)

1Crude and purified extracts are measured as micromole trolox equivalent (TE) or gallic acid equivalent (GE) per gram dry weight plant material. The data represent the mean ± standard deviation of at least three independent experiments. Means with different letters in the same column were significantly different with p < 0.05 (n = 3).

106

107

4.1.4.2 DPPH free radical scavenging assay

In this in vitro assay, antioxidant activity was determined based on the ability of the antioxidants to scavenge the stable DPPH free radicals and the results are presented in Table 16. Similar to the results of ABTS radical scavenging assay, the crude extracts of river mint exhibited the highest DPPH radical scavenging activity (178.9 ± 8.4 µmol GE/g, DW), which was comparable to that of spearmint (168.2 ± 6.9 µmol GE/g, DW). However, the purified extract of spearmint exhibited a higher activity than that of river mint. As expected, due to the removal of sugars, lipids and pigments, the antioxidant activity of each plant extract decreased after purification.

Park (2011) evaluated the DPPH radical scavenging activities of 15 phenolic compounds that are commonly present in herbs. The results showed that rosmarinic acid, caffeic acid and chlorogenic acid exhibited high DPPH radical scavenging activities. As will be shown later in Chapter 5 and 6, these three phenolic acids were the main phenolic compounds identified in the Mentha species. This suggests that these compounds might be part of the reason why the crude and purified extracts of the two Mentha species possessed a significantly higher DPPH radical scavenging activity than that of mint bush.

4.1.4.3 Ferric Reducing Antioxidant Power (FRAP) assay

Unlike the free radical scavenging assays, the FRAP assay measures the capacity of antioxidants to reduce ferric ion (Fe3+) to ferrous ion (Fe2+). The crude extracts of river mint showed the highest ferric reducing activity (1933 ± 63 µmol TE/g, DW), followed by spearmint (1859 ± 9 µmol TE/g, DW), with the mint bush having the lowest activity (Table 16). The same trend as in the case of ABTS and DPPH assay results was observed where the purified extracts gave lower FRAP values than the crude extracts. Overall, mint bush exhibited a lower FRAP value than the Mentha species in both crude and purified extracts. Interestingly, compared to the results obtained by ABTS and DPPH assays, TEAC values produced by the FRAP assay were considerably higher. This may be due to phenolic compounds being more efficient in reducing ferric iron

108 than scavenging free radicals as a result of steric hindrance as suggested by Wong et al. (2006).

Wong et al. (2006) investigated FRAP values of 21 aqueous plant extracts, and mint (M. arvensis) was found to have higher FRAP value (<200 µmol TE/g, DW) than other herbs such as Thai basil (<150 µmol TE/g, DW) and coriander (100 µmol TE/g, DW). In the present study, the mints were found to have a significantly higher FRAP result compared to that reported by Wong et al. (2006). This is likely due to the different extraction methods used in the extraction of phenolic compounds. In the present study, extraction was carried out by 80% (v/v) methanol, whereas Wong et al. (2006) used water as the extraction medium. Previous studies have generally found that antioxidant activities were higher in alcoholic extracts than in aqueous extracts of plants. For example, Lu et al. (2011) evaluated ethanolic extracts of 19 commonly consumed spices in China, and obtained higher FRAP values than those of the aqueous extracts reported by Wang et al. (2004).

4.1.4.4 Oxygen Radicals Absorbance Capacity (ORAC) assay

The ORAC assay has been regarded as one of the in vitro antioxidant assays that give the closest approximation to physiological conditions in the body (MacDonald‐Wicks et al., 2006). Although this assay has a number of drawbacks, which has been described by Schaich et al. (2015), the ORAC assay still offers some advantages over DPPH and ABTS in the determination of antioxidant capacity of phytochemicals. As shown in Table 16, crude extracts of river mint exhibited the highest peroxyl radical inhibiting activity (1727 µmol TE/g, DW), followed by spearmint (1551 µmol TE/g, DW) and mint bush (1289 µmol TE/g, DW). The purified extract also followed the same trend with river mint exhibiting the highest ORAC activity. As with the results obtained by the other antioxidant assays, the purified extracts showed lower ORAC values compared with their crude counterparts.

Herbs have been reported to possess high ORAC values. Wu et al. (2004) conducted a comprehensive study on the ORAC properties on a variety of food including fruits, vegetables, nuts, spices and grain-based food, and the ORAC values for herbs were 109 found to be much higher than the other foods. The highest ORAC value among vegetables was dried beans (145 µmol TE/g, DW), and for fruits it was cranberries (92.5 µmol TE/g, DW). Among herbs, the ORAC values of oregano and parsley were 1831.4 and 740.9 µmol TE/g (DW), respectively (Wu et al., 2004). Kratchanova et al. (2010) studied the antioxidant capacity of 25 medicinal plants including a number of herbs. The acetone-extracted peppermint exhibited the highest ORAC value (2917 ± 52 µmol TE/g, DW). Other acetone extracts of herbs, such as thyme (1637 ± 59 µmol TE/g, DW) and wild basil (1437 ± 52 µmol TE/g, DW), showed comparable ORAC values to those of mint samples found in this study, while spearmint exhibited a lower capacity at 748 ± 57 µmol TE/g (DW).

4.1.4.5 General discussion on the antioxidant capacity of Australian native mints

Extensive studies have been carried out on the antioxidant capacity of plant species belonging to the Lamiaceae family (Zheng and Wang, 2001, Shan et al., 2005, Wojdyło et al., 2007). These studies have demonstrated that herbs from this family possess good antioxidant capacity. Of particular interest is spearmint, which was reported to exhibit excellent antioxidant capacity, comparable to that of the synthetic antioxidant – butylated hydroxytoluene (Kanatt et al., 2007) – and hence was used as a reference material in this study. Results obtained in the present study revealed that the antioxidant capacity of crude river mint extracts was either comparable or superior to that of spearmint. Overall, the antioxidant capacities of crude and purified extracts of mint bush were not as high as those of the Mentha species. Nevertheless, mint bush demonstrated high antioxidant capacity when compared to many common herbs as mentioned previously.

To date, only Zhao and Agboola (2007) have conducted antioxidant testing on M. australis and P. incisa. In their study, 17 Australian native bushfoods were investigated for their phenolic content, antioxidant capacity as well as antimicrobial effect. It was found that M. australis had the highest total phenolic content (98.1 mg gallic acid (GE)/L extract) among the bushfoods. For the Prostanthera species P. incisa, total phenolic content was significantly lower (18.4 mg GE/L) with a DPPH radical

110 scavenging activity of 12.4%. The general trend was in agreement with the current study where P. rotundifolia exhibited lower antioxidant capacity than M. australis.

There are a variety of phytochemicals, including lipids, pigments and reducing sugars, which may possess some antioxidant activities. These non-phenolic phytochemicals were expected to be present in crude extracts but were mostly removed during the purification process. Consequently, crude extracts were found to have a higher total phenolic content and antioxidant capacities than their purified counterparts. Furthermore, because not all phenolic compounds were recovered during purification, the partial loss of phenolic compounds would contribute to the lower antioxidant activities of the purified extracts. However, the purified extracted represented a phenolic-rich fraction of the extracts, and their relatively high antioxidant capacities, albeit lower than their crude counterparts, indicated that most of the antioxidant effects of the crudes came from the phenolic components of the herbs.

4.1.4.6 Correlation analysis

Correlation analysis was carried out between the total phenolic and total flavonoid contents of the mint herbs and their antioxidant capacities obtained by the four different assays (Table 17). Significant, positive correlations were observed between total phenolic content and antioxidant capacities obtained by all four antioxidant assays. However, total flavonoid content only showed significant, positive correlations with ABTS (r = 0.925, p < 0.01) and DPPH (r = 0.898, p < 0.05) values, but not with FRAP (r = 0.806) and ORAC (r = 0.787) results. This appears to indicate that phenolic compounds in the mint samples had direct relationship with the antioxidant mechanisms of free radical scavenging, ferric chelation and peroxyl inhibition, while flavonoids were more closely associated with the free radical scavenging, but less so with ferric chelation and peroxyl inhibition. This could be related to the fact that the antioxidant capacity of flavonoids usually depends on the position of hydroxyl group on the B-ring and whether it has the ability to donate hydrogen or electron to a free radical (Miliauskas et al., 2004). Moreover, correlation between antioxidant activity and phenolic compounds can be affected by synergistic/antagonistic effects of different

111 compounds present in the reaction mixture, mechanism of antioxidant reactions and the experimental conditions (Mustafa et al., 2010).

Total flavonoid content of the mint extracts also showed a significant correlation with the total amount of phenolics (r = 0.863, p < 0.05), which was also found in other studies (Miliauskas et al., 2004). Significant, positive correlation between phenolic content and antioxidant activity has been reported in a number of studies on plant food (Shan et al., 2005).

Table 17. Pearson correlation coefficient (r) and level (p) for relationships between total phenolic content, total flavonoid content and antioxidant assays of native Australian mints1.

TPC TF ABTS DPPH FRAP ORAC

TPC 1 0.863* 0.933** 0.946** 0.991*** 0.925**

ns TF 1 0.925** 0.898* 0.806ns 0.787

ABTS 1 0.997*** 0.918** 0.957**

DPPH 1 0.941** 0.977***

FRAP 1 0.947**

ORAC 1

1Significant correlation at p < 0.05, 0.01 and 0.001 are indicated by *,** and ***, respectively; ns = not significant. TPC: total phenolic content, TF: total flavonoid, ABTS: 2,2’-azino-bis(3- ethylbenzothiazoline-6-sulphonic acid), DPPH: di(phenyl)-(2,4,6-trinitrophenyl)iminoazanium, FRAP: ferric reducing capacity power, ORAC: oxygen radicals absorbance capacity.

112

4.1.5 Inhibitory effects of Mentha australis and Prostanthera rotundifolia against health related enzymes

4.1.5.1 Inhibition against α-glucosidase

α-Glucosidase is a key enzyme involved in the digestion of starch. It breaks down dietary starch into glucose molecules, which can be readily absorbed by the body. However, rapid and excessive absorption of glucose can lead to hyperglycemia, so the inhibition of this enzyme is often used for the management of type 2 diabetes. The α- glucosidase inhibitory activity of the crude and purified extracts of the three herbs at three different concentrations (1.25-5.00 mg/mL) was investigated and the results are presented in Figure 13. Among the crude extracts, river mint exhibited the strongest α- glucosidase inhibitory effect (IC50 = 5.6 ± 0.5 mg/mL), while the activities of the mint bush and spearmint were significantly lower. This was especially true at high (5.00 mg extract/mL) and low concentrations (1.25 mg/mL). For example, at the concentration of 5.00 mg/mL, 45% of the α-glucosidase activity was inhibited by the crude extracts of river mint, while the activities for the other two mint samples were less than 20%. However, at 2.50 mg/mL, the crude extracts of river mint and mint bush showed comparable inhibitory activities against α-glucosidase (24-25%), while spearmint showed a lower activity of 16% inhibition.

With regards to the purified extracts, inhibition was only observed at lower concentrations of 1.25 and 2.50 mg/mL. No inhibition was observed at 5.00 mg/mL for all mints, which could be due to interference by plant materials in the concentrated samples affecting the spectrophotometer reading. At the concentration of 2.5 mg/mL, purified extracts of river mint and spearmint exhibited similar α-glucosidase inhibitory activities of 76.7% and 83.6%, respectively, while the inhibition of purified mint bush extract was at a lower inhibition of 51.2%.

Studies have shown that phenolic compounds exhibited α-glucosidase inhibitory effects (Kwon et al., 2006, Lee et al., 2014). Kwon et al. (2006) evaluated the α-glucosidase inhibitory activity of pure phenolic compounds that are commonly found in herbs. Their study showed that the inhibitory effect of caffeic and rosmarinic acids was especially high, with 91.3% and 85.0% inhibition of this enzyme at the concentration of 1 mg/mL, respectively. As will be described in detail in Chapters 5 and 6, these two phenolic acids

113 were also present in the mint herbs examined in the current study, suggesting that they could have a significant contribution to their α-glucosidase inhibitory effects. Several other phenolic compounds present in the mint herbs (e.g., 4-methoxycinnamic acid, see details in Chapters 5 and 6), have also been reported to exhibit significant inhibition against α-glucosidase (Adisakwattana et al., 2004), which would also be expected to contribute to the inhibtitory activity of the mints against the enzyme.

1 0 0 1 .2 5 m g /m L 2 .5 m g /m L 8 0

n 5 m g /m L

i t

i 6 0

i h

n 4 0

% 2 0

0 ) ) ) ) ) ) C C C (P P (P ( ( ( ( s a a is a a li li t l li t a a a a o c r o c tr if i t if i s d p s d p u n s u n s . . a u a u . t . t M o M o M r M r . . P P

Figure 13. α-Glucosidase inhibitory activity of crude (C) and purified (P) extracts obtained from native Australian mints (Mentha australis and Prostanthera rotundifolia) and reference sample of spearmint (Mentha spicata). The data represent the mean ± standard deviation of at least three independent experiments.

4.1.5.2 Inhibition against α-amylase

Similar to α-glucosidase, α-amylase also plays a crucial role in the digestion of dietary starch, and the inhibition of this enzyme is an important part of the overall strategy in managing type 2 diabetes. The capacity of purified extracts of the mints to inhibit α- amylase activity was evaluated at a range of concentration from 5 to 20 mg/mL of purified extracts (Figure 14). Analysis of crude extracts for their inhibitory effects 114 against α-amylase was not successful due to strong interference by impurities in the samples. Results obtained for the purified extracts revealed that the α-amylase- inhibitory activity of the mint samples was concentration dependent, but the effect of the extract concentration was not consistent among the different samples. For both of the Mentha species, the initial increase in extract concentration from 5 to 10 mg/mL, resulted in a sharp increase in the activity. However, a further increase in concentration to 20 mg/mL caused marked decline in the activity. For Prostanthera rotundifolia, the activities were similar at the concentrations of 5 and 10 mg/mL. When the concentration was increased to 20 mg/mL, however, a significant increase in the activity was observed. At the lowest extract concentration of 5 mg/mL, Prostanthera rotundifolia gave the highest activity, followed by the spearmint and river mint. At 10 mg/mL, the spearmint exhibited the highest activity, while the two native mints showed similar activities. At the high extract concentration of 20 mg/mL, the Prostanthera rotundifolia gave the highest activity, while spearmint and river mint exhibited the lowest activity.

Low α-amylase inhibitory activity was also reported for other Mentha species (Marrelli et al., 2014, Figueroa-Pérez et al., 2014). Aqueous crude extract of peppermint at the concentration of 10 mg/mL was found to exhibit low α-amylase inhibitory activity (0.5% inhibition) (Figueroa-Pérez et al., 2014). The maximum α-amylase inhibitory capacity of the peppermint extract was achieved at 45% with concentrations between 60-80 mg/mL. Furthermore, Figueroa-Pérez et al. (2014) found a weak correlation between rosmarinic acid in mints and α-amylase inhibitory activity.

115

5 0 5 m g /m L

n 4 0 1 0 m g /m L o

i 2 0 m g /m L t

i 3 0

i h

n 2 0

% 1 0

is ia ta l l a a o c tr if i s d p u n s a u . . t o M M r . P

Figure 14. α-Amylase inhibitory activity of purified extracts obtained from native Australian mints (Mentha australis and Prostanthera rotundifolia) and reference sample of spearmint (Mentha spicata). The data represent the mean ± standard deviation of at least three independent experiments.

4.1.5.3 Inhibition against pancreatic lipase

Pancreatic lipase hydrolyses dietary fats into smaller components such as fatty acids and monoglycerides, which can then be absorbed by the body. Accumulated absorption of fats can cause excess fat deposition which leads to weight gain and obesity (Garza et al., 2011). Inhibition of pancreatic lipase can prevent this occurrence, thus reducing the risk of overweight and obesity (Birari and Bhutani, 2007). Figure 15 shows the inhibition of pancreatic lipase by the crude and purified extracts of the three mint samples at various concentrations (1.25-5.00 mg/mL). It is clear that the inhibition of pancreatic lipase was dependent on the concentration of the mint extracts. Results revealed comparable pancreatic lipase inhibitory activities for all mint samples. However, for a better comparison of the relative inhibitory activity of the mint samples, the data was extrapolated to determine the IC50 values – concentration of the extract with 50% inhibition of the lipase activity (Table 18).

116

For crude extracts, mint bush was the strongest inhibitor of pancreatic lipase with an

IC50 value of 0.39 ± 0.01 mg/mL, which was followed by river mint (0.61 ± 0.01 mg/mL). The crude extract of the spearmint had the lowest inhibitory activity against the enzyme with an IC50 value of 1.10 ± 0.01 mg/mL. However, after purification, the spearmint extract exhibited a stronger activity (IC50 = 1.90 ± 0.01 mg/mL) than the purified extracts of the two Australian mints (>3.97 ± 0.01 mg/mL).

Furthermore, it appeared that the purified extracts exhibited a lower inhibitory activity compared with their crude counterparts. A possible explanation of this phenomenon is that the non-phenolic constituents such as lipids, reducing sugars and pigments, which were mostly removed during purification, also contributed to lipase inhibition. For example, lipids were likely present in the crude extracts as they are partially soluble in methanol which was the main solvent of the extraction medium (80% v/v aqueous methanol). The endogenous lipids would be expected to react with some of the lipase in the reaction system, making it unavailable for reaction with the substrate in the assay. This would be equivalent to a decrease in the lipase concentration, or a reduction in its activity. Moreover, compounds such as saponins and terpenes extracted from plant sources have been reported as pancreatic lipase inhibitors (Birari and Bhutani, 2007, Garza et al., 2011) and their presence in the crude extracts, if occurred, would also contribute to the greater lipase inhibitory activities observed in the crude extracts.

Additionally, phenolic compounds were not equally effective as lipase inhibitors. This was demonstrated by Bustanji et al. (2010), in which pure rosmarinic acid was shown to have a higher IC50 (125.2 µg/mL) than that of other phenolic compounds: chlorogenic acid (96.5 µg/mL), caffeic acid (32.6 µg/mL) and gallic acid (10.1 µg/mL). Interestingly, the same study found that crude rosemary extract exhibited a higher inhibition of lipase with an IC50 of 13.8 µg/mL than most of the individual phenolic compounds. This suggests that non-phenolic compounds in the rosemary extract might contribute to the lipase inhibitory activity, or some forms of synergism existed among the different phenolic components. Their results agreed with this study as the crude extracts of all mint samples produced higher lipase inhibition than the purified extracts.

117

1 0 0 1 .2 5 m g /m L

8 0 2 .5 m g /m L

n 5 m g /m L

i t

i 6 0

h n

I 4 0

% 2 0

0 ) ) ) ) ) ) C C C (P P (P ( ( ( ( s a a is a a li li t l li t a a a a o c r o c tr if i t if i s d p s d p u s u n s n . . a u a u . t . t M o M o M r M r . . P P

Figure 15. Pancreatic lipase inhibitory activity of crude (C) and purified (P) extracts obtained from native Australian mints (Mentha australis and Prostanthera rotundifolia) and reference sample of spearmint (Mentha spicata). The data represent the mean ± standard deviation of at least three independent experiments.

Table 18. Lipase inhibitory activity of crude and purified extracts obtained from native mints with spearmint as reference sample

1 Pancreatic lipase inhibition [IC50 (mg/mL)] Plant species Crude extract Purified extract

River mint 0.61 ± 0.01a 3.97 ± 0.01a (Mentha australis)

Mint bush 0.39 ± 0.01b 4.70 ± 0.03b (Prostanthera rotundifolia)

Spearmint 1.10 ± 0.01c 1.90 ± 0.01c (Mentha spicata)

1 IC50: half maximum inhibitory concentration. The data represent the mean ± standard deviation of at least three independent experiments. Means with different letters in the same column were significantly different with p < 0.05 (n = 3).

118

4.1.5.4 Inhibition against hyaluronidase

Hyaluronidase is an enzyme known to be involved in allergic responses and inflammation. This enzyme acts by hydrolysing the polysaccharide hyaluronic acid, in which excessive amounts of the latter can lead to chronic inflammation (Soberón et al., 2010). The hyaluronidase inhibitory assay is an indirect way to assess anti-inflammation activity. In this study, crude extracts of river mint showed the highest inhibitory effect with an IC50 of 1.83 ± 0.04 mg/mL, while spearmint and mint bush had a much lower effect of 25.9% and 5.3%, respectively (Figure 16). With regards to the purified extracts, river mint showed a dose response relationship with an IC50 of 6.38 ± 0.04 mg/mL. However, spearmint exhibited a low hyaluronidase inhibitory effect with maximum inhibition (6.8%) at the concentration of 2.50 mg/mL and no inhibition was observed in mint bush. As explained in Section 4.1.3, the high inhibition of hyaluronidase by crude river mint extracts may be due to compounds that were not retained after purification, suggesting that these compounds were not phenolic acids or flavonoids.

Phenolic compounds, such as flavonoids and phenolic acids, have been reported to act as inhibitors of hyaluronidase (Samee et al., 2009, Ippoushi et al., 2000, Soberón et al., 2010, Moon et al., 2009). Rosmarinic acid, in particular, has been reported to have inhibitory activity against hyaluronidase (Nunes et al., 2015, Ito et al., 1998). Ippoushi et al. (2000) studied the hyaluronidase inhibitory effects of vegetables and herbs. Their study showed that rosmarinic acid extracted from lemon balm possessing the strongest inhibitory effect. Extracts of other herbs from the Lamiaceae family also exhibited high hyaluronidase inhibitory effects. Peppermint was reported to suppress the activity of this enzyme by 26.5% (Ippoushi et al., 2000), rosemary by 35.6%, thyme by 35.5% and sweet basil by 23.5%. In the current study, rosmarinic acid was found to be present in high levels in both Mentha species (described in Chapters 5 and 6 in detail), which could contribute significantly towards their inhibitory activity against hyaluronidase.

119

8 0 1 .2 5 m g /m L 2 .5 m g /m L

n 6 0

o 5 m g /m L

i b

i 4 0

% 2 0

) ) ) ) ) ) P C C (C (P P ( ( ( ( s a a is a ta li li t l li a a a a o c r o c tr if i t if i s d p s d p u s u n s n . . a u a u . t M . t M o o M r M r . . P P

Figure 16. Hyaluronidase inhibitory activity of crude (C) and purified (P) extracts obtained from native Australian mints (Mentha australis and Prostanthera rotundifolia) and reference sample of spearmint (Mentha spicata). The data represent the mean ± standard deviation of at least three independent experiments.

120

4.3. Conclusion

This chapter demonstrated that the two native Australian mints, river mint and mint bush, possessed high antioxidant capacities when compared with common herbs, as assessed by four different assays. In most of the assays, crude and purified extracts of river mint exhibited higher antioxidant capacity than the spearmint extracts, while mint bush extracts had lower capacity than both of the Mentha species. Overall, the crude extracts exhibited higher total phenolic content and antioxidant capacities than the purified extracts. This may be due to the presence of a variety of non-phenolic phytochemicals in the crude extracts, such as reducing sugars, lipids and pigments, which could also contribute to the measured phenolic content and antioxidant capacities. Antioxidant activity was positively correlated with total phenolic content for results obtained by all antioxidant assays, whereas total flavonoid content was significantly correlated only with the results of ABTS and DPPH assays.

The native Australian mint extracts were also found to exhibit significant inhibitory activities against α-glucosidase, pancreatic lipase and hyaluronidase enzymes. The Australian mints exhibited stronger inhibition against pancreatic lipase than the other enzymes. The crude extracts of river mint exhibited a higher or comparable inhibition of α-glucosidase, pancreatic lipase and hyaluronidase enzymes to that of spearmint. The crude extracts of mint bush, on the other hand, produced a lower inhibition of α- glucosidase and hyaluronidase compared with the two Mentha species. All three herbs exhibited relatively low α-amylase inhibitory activities.

The purified, polyphenolic-rich extracts of the mint herbs exhibited greater α- glucosidase inhibitory activities than their crude counterparts. However, purified extracts were weaker inhibitors of pancreatic lipase and hyaluronidase than the crude extracts, and only some levels α-amylase inhibition were observed when the concentration of the purified extracts was increased to 20 mg/mL. This suggests that inhibition of α-amylase by crude extracts was mainly caused by strong interference by impurities or non-phenolic compounds, which were removed by the purification process.

121

Chapter 5 Identification and quantification of phenolic compounds in Mentha australis R. Br.

5.1. Introduction

The Mentha genus (Lamiaceae) is widely distributed worldwide (Lawrence, 2007, Dorman et al., 2003). At least seven Mentha species are endemic to Australia, though only three (Mentha australis, Mentha diemenica and Mentha satureioides) have been reported to have medicinal value (Lassak and McCarthy, 2001, Williams, 2010). To date, no scientific studies have investigated the phenolic composition of these Australian native species. In Chapter 4, it was reported that mint extracts from the Mentha genus exhibit several potentially health-promoting activities. As mentioned in previous chapters, the chemical constituents and health benefits of essential oils and extracts of the Mentha genus, particularly spearmint and peppermint, have been well studied (Soković et al., 2009, Dorman et al., 2003). The phenolic compounds that are most common among the Mentha species include rosmarinic acid, caffeic acid, chlorogenic acid, luteolin, cynaroside, luteolin-O-rutinoside and eriocitrin (Dorman et al., 2003, Inoue et al., 2002, Guedon and Pasquier, 1994). However, little is known about the phenolic compounds present in the Australian mints in the Mentha genus.

In most of the studies on Mentha species, High Performance Liquid Chromatography- Photodiode Array Detection (HPLC-PDA) was employed in the analysis of phenolic compounds. This method provides quantification and tentative identification of phenolic compounds, especially when reference standards are available for comparison. Other analytical techniques were usually needed to confirm the identities of the compounds and to elucidate their structures, which exceeds the abilities of HPLC-PDA analysis. Mass spectrometric techniques have proved particularly useful and more powerful in identifying phenolic compounds in plants (Justesen, 2000).

Gas Chromatography-Mass Spectrometry (GC-MS) techniques are well developed for identifying volatile components, and those that can be volatilised through chemical derivatisation. The large numbers of mass spectra that are available in various GC-MS database libraries are especially helpful for the identification of unknown components as they allow mass spectra of the sample to be matched with those in the database. On 122 the other hand, Liquid Chromatography-High Resolution Mass Spectrometry (LC- HRMS) and Tandem Mass Spectrometry (LC-MS/MS) are more direct techniques for analysing non-volatile plant components as they do not require sample derivatisation. In cases where reference standards were unavailable, Nuclear Magnetic Resonance (NMR) Spectroscopy was often used to provide full structural characterisation of an organic compound in solution by analysing the resonance frequency of each nucleus (1H or 13C nuclei) of the molecule and their intensity of absorption.

The objectives of this chapter were to identify and quantify the major phenolic components in the Australian native mint species M. australis using HPLC-PDA as well as other state-of-the-art instrumental techniques, including LC-HRMS and GC-MS. LC- HRMS was conducted using a high resolution mass detector for accuracy and LC- MS/MS for specificity and sensitivity. Furthermore, the structural features of the major peaks collected from HPLC were elucidated by NMR.

123

5.2. Results and discussion 5.2.1. Determination of phenolic compounds in Mentha australis by High Performance Liquid Chromatography-Photodiode Array Detector 5.2.1.1. Optimisation of HPLC-PDA method

Good resolution of phenolic compounds by HPLC is essential for their identification and quantification as it not only allows proper comparison of retention times between sample peaks and standards, but also ensures that the compounds can be sufficiently separated for peak collection and mass spectrometry analysis. To this end, several gradient elution programs were trialled to optimise the procedure. The initial gradient program started with 3% of solvent B (acetonitrile, 100%) and 97% solvent A (2.5% acetic acid in water, v/v), followed by 9% B (0-5 min), 16% B (5-15 min), 50% B (15- 45 min) and 90% B (45-51 min). This, however, was found to give poor resolutions of the phenolic compounds in the purified extract (Figure 17). The mobile phase gradient was adjusted to improve the resolution by minimising the time where no peaks eluted. The target analytes were found to elute from the column between 16-17 min at a gradient of 16% acetonitrile and 84% acetic acid in water, hence the starting elution gradient was modified to 23% acetonitrile and 77% acetic acid. It was also found that more peaks were displayed with greater heights at the wavelength of 320 nm than at 280 nm and 370 nm and this wavelength was used in subsequent displays. The modified HPLC elution system produced much improved resolutions of compounds in the purified extract of M. australis (Figure 18), where several major and minor peaks of phenolic compounds were well resolved.

124

Absorbance (mAU) Absorbance

Figure 17. HPLC chromatogram at 320 nm of purified M. australis extract with a run

time of 60 min.

Absorbance (mAU) Absorbance

5.2.1.2. Preliminary identification of the phenolic compounds with HPLC-PDA

The extract of the river mint, like extracts of most plant materials, was expected to contain a diverse range of phenolic compounds and the identification of these compounds can be a difficult and complex process. In this study, a five-step identification and confirmation process, as illustrated in Figure 19, was used to ascertain the identity of the major phenolic compounds in the extract. First, the retention times of the peaks in the sample were compared with those of phenolic acid (caffeic, chlorogenic, rosmarinic, p-coumaric, p-hydroxybenzoic, syringic, cinnamic) and flavonoid (apigenin, naringenin, hesperetin, rutin hydrate, kaempherol, luteolin, naringin, quercetin) standards to achieve preliminary identification. Confirmation of the identity of the compounds was further sought by comparing the UV spectra of the suspect peaks with those of the standards, and by comparing the MS spectra of the suspected compounds with those of the standards using GC-MS and LC-MS. Finally, for certain unknown peaks, fractions of the peak were collected and subjected to NMR to facilitate their identification.

2. Co- 4. Confirmation 1. Comparison of 3. UV absorption chromatography by GC-MS, LC- 5. Confirmation retention time by spectra with standard HRMS & LC- by NMR standards comparison compounds MSMS

HPLC-PDA analysis

Figure 19. Processes used to identify phenolic compounds in the purified extract of M. australis.

126

Following this general process, six compounds, namely chlorogenic acid, caffeic acid, rutin hydrate, rosmarinic acid, ferulic acid and naringin, were tentatively identified based on their retention time, ultraviolet spectra (Appendix 2) and co-chromatography with standards.

Based on this analysis, two of the peaks were identified as chlorogenic acid (peak 1, 3.9 min, λmax = 324 nm) and rosmarinic acid (peak 4, 10.8 min, λmax = 328 nm; Figure 18), and their chemical identities were confirmed by LC-HRMS (Section 5.2.3.2). The identity of rosmarinic acid was further confirmed by 1H NMR, 13C NMR analysis (Section 5.2.4.1) and 2-D NMR analyses (Section 5.2.4.2). Peak 7 was found to contain two compounds, naringenin and apigenin, which were confirmed with LC-MS/MS. The remaining peaks could not be identified by HPLC-PDA, hence further analysis using mass spectrometry was performed to determine their chemical identities.

5.2.2. Determination of phenolic compounds by Gas Chromatography-Mass Spectrometry

Initial mass spectrometry analysis of the mint extracts was conducted on Gas Chromatography-Mass Spectrometry (GC-MS). Crude extracts of the mint samples were analysed using the NIST 2.0 software which contained a library of reference mass spectra for compound identification. The phenolic compounds in the crude methanolic extracts were converted to volatiles by derivatisation with N,O- bis(trimethylsilyl)trifluoroacetamide (BSTFA) reagent. During this process, a trimethylsilyl group was attached to the functional group of analytes by replacing the active hydrogen (Orata, 2012). For example, after treatment with BSTFA the active hydrogens on the hydroxyl groups of naringenin were replaced with silyl groups, converting the molecule to silylated naringenin (Figure 20). Pyridine was added as a catalyst which facilitated the reaction of the sterically hindered groups (Orata, 2012). Mass spectra of known standards were compared with the sample. Matching compounds were also analysed with the NIST library software, which provided a reverse match factor. The reverse match factor is a numerical value, scaling from 0-999, which takes both the m/z value and the intensities of all peaks of the mass spectra of the unknown compound into consideration in the calculations. Generally, a reverse factor of 800 and above is regarded as a good match and this value was used in the present study. 127

An example of such comparison was shown in Figure 21, and the list of phenolic compounds identified by GC-MS is given in Table 19.

A B

Figure 20. Naringenin (A) and the silylated naringenin treated with BSTFA reagent (B).

However, using the methods described above, only the phenolic acids, caffeic acid and rosmarinic acid were identified and confirmed. Other suspected phenolic compounds identified previously by HPLC-PDA analysis were not found by GC-MS. The main reason for this was that the GC-MS spectra were dominated by peaks of sugar compounds, which might exist either in free form or as sugar moieties bound with phenolic compounds such as in flavonoid glycosides. Interestingly, the GC-MS results showed a relatively high intensity of the caffeic acid peak, whereas HPLC-PDA analysis showed that the peak intensity of caffeic acid was very low, especially when compared to that of rosmarinic acid. The large intensity of caffeic acid on the GC-MS spectrum may result from the hydrolysis of chlorogenic acid or rosmarinic acid, as both contain at least one caffeic acid derivative as their constituent.

To facilitate the identification of flavonoid glycosides by GC-MS, the crude mint extract was hydrolysed by 1% HCl to break the phenolic compounds into their constituting aglycone and sugar components to obtain information on the aglycones. Hydrolysis also resulted in the breakdown of some of the free sugars, thus improving the mass spectra in general. Figure 22 displays the GC-MS chromatograph before and after hydrolysis. However, the phenolic compounds were still suppressed by the sugar peaks and were difficult to identify. Nevertheless, using this method, the fragmented

128 components obtained from hydrolysis provided information regarding the aglycones, from which predictions could be made on the possible phenolic acids and flavonoid glycosides. An example of this was mentioned previously where high levels of caffeic acid could be a result of the hydrolysis of chlorogenic acid or rosmarinic acid. Although a small amount of silylated naringenin was detected on the GC-MS chromatogram, a high level of the compound was also identified occurring at a different retention time. It was suspected that naringenin may be an aglycone fragment which was hydrolysed from a glycoside compound. Further analysis is necessary to support this finding.

Overall, the use of GC-MS enabled the identification of six phenolic compounds, caffeic acid, rosmarinic acid, naringenin, t-cinnamic acid, syringic acid and p-coumaric acid. The dominance of sugar peaks prevented the identification of other phenolic components in the mint samples by this technique. However, information obtained by GC-MS provided additional evidence for the identification of some phenolic compounds in the mint samples.

129

73 100 219

396 O O

Si Si O O Si 50

191 381

249 307 147 293 115 175 233 59 89 203 267 0 30 60 90 120 150 180 210 240 270 300 330 360 390 (w9n11) CAFFEIC ACID-TMS-ETHER

219 100

73 50 396

191 249 381 28 45 59 88 102 115 133 147 163 175 203 233 267 280 293 307 0 39 51 66 81 93 123 259 321 337 59 89 163 175 203 267 280 115 133 147 233 293 249 307 45 50 191 381

100 396 73 219 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400 410 caffeic#10442-10454 RT: 63.05-63.12 AV: 13 Head to Tail MF=881 RMF=934 CAFFEIC ACID-TMS-ETHER Figure 21. GC Mass spectrum confirmation of trimethylsilyl-derivatised caffeic acid using the NIST library (above) and the head to tail

128 matching (below).

130

Table 19. Molecular ions present in the mass spectra of silylated phenolic compounds1 in the methanolic crude extract of Mentha australis identified by Gas Chromatography-Mass spectrometry

Retention Time Molecular ion Reverse Peak2 Silylated Compound MW3 Fragmented ions (%) (min) [M]+ match factor

2 Caffeic acid 63.08 396 396 935 191(19), 219(100), 396(89)

4 Rosmarinic acid 97.42 720 720 833 179(15), 219(50), 324(30), 396 (100)

7 Naringenin 86.91 488 488 - 28(39), 32(17), 45(14), 473(12), 296(11),

44(10), 147(9)

- t-cinnamic acid 35.09 220 220 884 103(83), 131(100), 161(69), 205(98), 220(27)

- Syringic acid 56.41 342 342 832 253(69), 297(66), 312(76), 327(100), 342(65)

- p-coumaric acid 57.59 308 308 935 219(89), 249(84), 293(100), 308(65)

1Identified as trimethylsilyl derivatives. 2Peak obtained from HPLC-PDA analysis. 3MW: molecular weight of silylated compound.

129

131

C:\Users\...\2013-10-24_KittyTang\M 24/10/2013 4:47:31 PM

RT: 9.63 - 99.91 61.31 NL: 100 1.15E9 Base Peak 90 F: + c Full 63.14 81.86 A ms 80 82.39 [25.00- 700.00] 70 MS M 60

50 84.12 56.53 40 80.80 64.96

Relative Relative Abundance 97.79 30 96.22 34.82 79.25 20 55.22 65.68 69.44 10 54.88 72.31 89.34 18.09 19.89 33.55 41.34 94.10 0 10 20 30 40 50 60 70 80 90 C:\Users\...\2013-10-24_KittyTang\Ma 24/10/2013 6:41:31 PM Time (min) RT: 9.88 - 99.56 M #1 RT: 5.00 AV: 1 NL: 4.23E7 61.33 NL: T: +100 c Full ms [25.00-700.00] 1.18E9 58.05 100 Base Peak 90 B m/z= 90 25.00-700.00 F: 80 + c Full ms 80 [25.00-700.00] 70 63.11 MS Ma 70 15.65 60 67.95 60 60.83 50 56.52 50 81.82 40 54.88

40 Relative Relative Abundance 30 Relative Relative Abundance 78.30 30 88.68 94.75 20 75.04 20 19.88 52.29 69.45 34.80 47.91 10 146.12 72.30 10 77.04 25.69 35.92 0 170.04 244.06 281.06 396.47 441.32 515.80 555.00 630.55 657.42 0 10 20 30 40 50 60 70 80 90 100 200 300 400 500 600 700 Time (min) m/z Ma #1 RT: 5.00 AV: 1 NL: 1.33E7 FigureT: + c Full 22 ms. [25.00-700.00]Total ion current (TIC) GC-MS chromatogram of non-hydrolysed (A) and 58.07 hydrolysed100 crude extract of M. australis.

90 75.04 80

40 Relative Relative Abundance 30 77.04

20 147.08 179.08 10 117.05 267.14 396.19 309.01 470.09 496.40 596.23 645.86 679.76 0 100 200 300 400 500 600 700 m/z 132

5.2.3. Identification of phenolic compounds in M. australis by Liquid Chromatography-High Resolution Mass Spectrometry

5.2.3.1. Optimisation of LC-HRMS method

As the GC-MS library database did not provide sufficient information, LC-HRMS was implemented. LC-HRMS provides high resolution mass spectra with an accuracy of less than 2 ppm. The choice of ionisation method is important and depends on the nature of the analyte molecules. To optimise ionisation results for all the unknown components in the analytes, cations and anions of the target analytes were examined using both atmospheric pressure chemical ionisation (APCI) and Heated Electrospray Ionisation (HESI) sources. This was due to the fact that some analytes can be detected in their anion form while not in their cation form, or ionised with the APCI source but not the HESI. For example, the HESI method was found to be the best ionisation technique for rosmarinic acid, especially at low concentrations. As this study was aimed at investigating phenolic compounds, a mass range was set at the mass-to-charge ratio (m/z) of 120-700.

Initially, the same gradient and mobile phase used in the HPLC-PDA analysis was also employed in the LC-HRMS APCI and HESI methods. Although this method worked well with major compounds, the minor compounds that corresponded with the HPLC chromatogram were found difficult to distinguish or did not appear on the mass spectra. The use of 2.5% acetic acid in water and acetonitrile (100%) as mobile phases produced weak or no mass spectra and a lot of noise and the analytes did not ionise well with high acetic acid concentration. Only the peaks at retention times 13.1 and 17.2 min could be identified, which had the molecular ions [M-H]- of 359 and 593, respectively (Figure 23). Furthermore, the peaks at 13.1 and 17.2 min corresponded with peak 4 and peak 6, respectively, on the HPLC-PDA chromatogram as shown in Figure 18.

An alternative elution system using 5 mM ammonium formate (pH 7.4) and 5 mM ammonium formate in 90% methanol (pH 7.4) as the mobile phases enhanced ionisation of the analytes leading to more distinct spectra. This is evident in Figure 24 where the total ion current chromatogram produced by ammonium formate-methanol elution also presented significantly less noise and better baseline than the acetonitrile-acetic acid elution. The peak at 2.0 min showed the molecular ions [M-H]- of 359 and [M+H]+ of

133

361 to be most abundant, while peak at 13.4 min had the molecular ions [M-H]- of 593 and [M+H]+ of 595, which had the highest intensities. The HESI technique also produced similar molecular ion levels for these masses. Based on these results, this ammonium formate-methanol (pH 7.4) elution system was used in all subsequent LC- HRMS analyses.

RT: 0.00 - 30.00 SM: 7G 17.19 NL: 1.61E8 100 A TIC TIC F: FTMS - p APCI corona Full ms 80 [120.00-700.00] MS 150318_M_1ug_APCI

60 13.05

40 2.91 26.92 6.72 8.29 10.71 20.99 14.99 24.30 28.76 20

0 13.05 NL: 9.27E5 100 m/z= 358.50-359.50 F: B 359 m/z FTMS - p APCI corona 80 Full ms [120.00-700.00] MS 150318_M_1ug_APCI 60

40 21.01

RelativeAbundance 20 26.86 2.94 4.71 7.36 10.09 14.54 17.76 22.96 25.18 28.59 0 17.18 NL: 3.83E6 100 m/z= 592.50-593.50 F: FTMS - p APCI corona 80 Full ms C 593 m/z [120.00-700.00] MS 150318_M_1ug_APCI 60

0.72 2.68 4.78 6.59 9.96 11.73 13.35 14.93 19.69 21.56 23.47 25.10 27.21 28.86 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 Time (min)

Figure 23. LC-HRMS Total ion current (TIC) chromatogram (A) of M. australis using atmospheric pressure chemical ionisation with mass range between m/z 120-700. Mobile phase acetonitrile and acetic acid were used. The extracted ion chromatogram: [M-H]- m/z of 359 (B) and 593 (C) were found to correspond to the major peaks from the TIC chromatogram.

134

C:\Users\...\140521_M_1ug_a_neg 21/05/2014 4:11:43 PM

RT: 0.00 - 20.01 SM: 7G 13.36 NL: 8.19E7 100 TIC F: FTMS - p APCI corona Full 90 ms [120.00-700.00] MS 80 140521_M_1ug_a_ neg 70

RelativeAbundance 30 1.98 20

10 1.11 3.54 5.88 7.35 8.85 10.36 11.89 15.28 17.53 18.76 0 0 2 4 6 8 10 12 14 16 18 20 C:\Users\...\140521_M_1ug_a_neg 21/05/2014 4:11:43 PM Time (min) 140521_M_1ug_a_negRT: 0.00 - 20.01 SM: #876-9247G RT: 13.15-13.84 AV: 24 NL: 1.36E7 F: FTMS - p APCI corona Full ms [120.00-700.00] 13.36 NL: 8.19E7 100 593.1854 TIC F: FTMS - p 100 A APCI corona Full 90 ms [120.00-700.00] 90 MS 80 140521_M_1ug_a_ 80 neg 70 70 60 60 50 50 40

40 285.0759 RelativeAbundance RelativeAbundance 3030 1.98 2020 639.1906 1010 1.11 3.54 5.88 7.35 10.36 11.89 15.28 17.53 18.76 353.10778.85 663.1848 151.0033 198.0723 259.0967 403.5844 460.1206 516.1076 563.1748 00 0 150 2 200 4250 3006 3508 400 10 450 12 500 14 550 16600 65018 70020 C:\Users\...\140521_M_1ug_a_neg 21/05/2014 4:11:43 PM Timem/z (min) RT:140521_M_1ug_a_neg0.00 - 20.01 SM: 7G #123-138 #876-924RT: 1.83-2.04RT: 13.15-13.84AV: 8 NL:AV:3.39E624 NL: 1.36E7 F: FTMSFTMS - p- pAPCI APCI corona corona Full Fullms [120.00-700.00] ms [120.00-700.00] 13.36 NL: 8.19E7 100 359.0760 TIC593.1854 F: FTMS - p 100100 APCI corona Full 90 ms [120.00-700.00] 9090 B MS 80 140521_M_1ug_a_ 80 80 neg 70 7070 60 6060 50 5050

4040 285.0759

RelativeAbundance RelativeAbundance RelativeAbundance 3030 1.98 2020 179.0344 20 316.0941 207.2928 639.1906 1010 1.11 3.54 5.88 7.35 8.85 353.107710.36 11.89 15.28 17.53 18.76 663.1848 151.0033 198.0723 270.0887259.0967 427.0630403.5844 494.1201460.1206541.1228516.1076593.1850563.1748653.5089 00 0 1501502 2002004 250 2506 300 8300 350 10350400 12 400450 14 500450 16 550500 18 60055020 650 600 700 650 700 Time (min) m/z m/z 140521_M_1ug_a_neg #876-924 #123-138RT: 13.15-13.84RT: 1.83-2.04AV: 24AV:NL:8 1.36E7NL: 3.39E6 F: FTMSFTMS - p- pAPCI APCI corona corona Full Fullms [120.00-700.00] ms [120.00-700.00] 359.0760 593.1854 100100

9090 C

8080

7070

6060

5050

4040 285.0759 RelativeAbundance RelativeAbundance 3030 179.0344 2020 316.0941 207.2928 639.1906 1010 353.1077 494.1201 663.1848 151.0033 198.0723 259.0967270.0887 403.5844 460.1206427.0630516.1076 563.1748541.1228 593.1850 653.5089 00 150150 200200 250 250 300 300 350 350400 400450 500450 550500 600550 650 600 700 650 700 m/z m/z 140521_M_1ug_a_neg #123-138 RT: 1.83-2.04 AV: 8 NL: 3.39E6 F: FTMS - p APCI corona Full ms [120.00-700.00] 359.0760 Figure100 24. LC-HRMS Total ion current (TIC) chromatogram (A) of M. australis using

negative90 atmosphere pressure chemical ionisation with mass range between m/z 120-

700.80 Mobile phase ammonium formate in water and ammonium formate in 90% - methanol70 were used. The extracted ion chromatogram: [M-H] of 359 (B) and 593 (C) corresponded60 to peaks at retention time 2 and 13.4 min, respectively. 50

40 135

RelativeAbundance 30 179.0344 20 316.0941 207.2928 10 270.0887 427.0630 494.1201 541.1228 593.1850 653.5089 0 150 200 250 300 350 400 450 500 550 600 650 700 m/z 5.2.3.2. Identification of phenolic compounds in Mentha australis by Liquid Chromatography-High Resolution Mass Spectrometry and Tandem Mass Spectrometry

The obtained mass spectra of the unknown peaks were compared with those generated by the Xcalibur software to produce empirical formulae whose chemical structures and names were obtained from the online database, Chemspider. The empirical formulae of

C:\Users\...\140526_m_nonspiked 26/05/2014 4:37:05 PM predicted compounds were also integrated into the software to simulate a mass spectrum RT: 0.00 - 20.02 SM: 7G of the compound. An example of the comparison betweenRT: 13.49 the mass spectraNL: 9.83E6 of an 100 TIC F: FTMS - p ESI Full unknown compound80 in M. australis (A) and the simulated rosmarinic acid (B) msis shown[120.00-1000.00] MS ICIS - in Figure 2560. The simulation software generates a predicted spectrum of the [M140526_m_nonspiked-H] of RT: 15.02 rosmarinic acid,40 thus indicating that the m/z 473 and 523 on the MRT:. australis 16.99 spectrum is RT: 12.18 20 RT: 18.08 Relative Relative Abundance not part of rosmarinicRT: 2.02 acid.RT: Thus,5.84 simulatedRT: 10.50 mass spectrum can provide more 0 RT: 5.84 NL: 2.29E5 information100 than just the accurate mass. Other comparisons are given in Appendices 3 m/z= 359.0762-359.0782 80 F: FTMS - p ESI Full ms to 5. The unknowns were found to have consistent masses to the simulated formulae.[120.00-1000.00] MS 60 ICIS Using this method, the following phenolic compounds were identified in M. australis140526_m_nonspiked: 40 chlorogenic acid, caffeic acid, narirutin, rosmarinic acid, biochanin A, apigenin, 20 RT: 4.01 RT: 6.99 naringenin, hesperetin and neoponcirin/poncirinRT: 9.39 (Figure 26). 0 0 2 4 6 8 10 12 14 16 18 20 Time (min)

359.0764 NL: 100 M. australis 1.78E5 A 80 140526_m_nonspiked#2 87-320 RT: 5.46-6.07 60 366.9337 AV: 17 F: FTMS - p ESI Full ms [120.00-1000.00] 40 395.0527 20 473.0685 Relative Relative Abundance 422.0716 525.1020 287.0553 341.0868 573.0618 0 359.0772 NL: 100 8.07E5 80 C18 H16 O8 +H: Simulated data C18 H15 O8 60 B pa Chrg -1

20 0 300 350 400 450 500 550 600 m/z

Figure 25. Comparison of molecular ion [M-H]- (m/z 359) of rosmarinic acid

(C18H16O8) in M. australis (A) and the simulated data (C18H16O8 – H) generated from Xcalibur (B) by LC-HRMS.

136

Chlorogenic acid Caffeic acid

Narirutin Rosmarinic acid

Biochanin A Apigenin

Naringenin Hesperetin

Neoponcirin Poncirin

Figure 26. Chemical structure of compounds identified in purified M. australis extract.

137

For peak 6, which consisted of a molecular ion [M-H]- mass of m/z 593 and [M+H]+ mass of m/z 595, two possible compounds were generated, neoponcirin and poncirin. These two compounds are flavanone glycosides with identical empirical formula and hence mass. Neoponcirin and poncirin consist of the same flavonoid isosakuranetin backbone and glucose sugar attachment, but differ in the rhamnose attachment. The rhamnose molecule is attached to glucose at different positions in the two compounds (Figure 27). For neoponcirin, the rhamnose unit is attached to the fifth carbon on the glucose ring while the poncirin rhamnose is attached to the second carbon of the glucose.

Figure 27. Glucose and rhamnose sugar attachment in neoponcirin (left) and poncirin (right).

The likelihood that the unknown compound of peak 6 contained isosakuranetin as its backbone was tentatively identified by the presence of a fragment ion [M-H]- of m/z 285 in the mass spectrum (Figure 28). The results were consistent when the sample was analysed in the cation form, where a fragment ion [M+H]+ of m/z 287 was identified.

This fragment ion corresponded to the molecular formula C16H14O5, which consists of several possible compounds that could be glycosylated, and isosakuranetin was one of the compounds generated from the Chemspider library. Along with the molecular and fragment ion information displayed in the spectrum of peak 6, it was deduced that the compound could be neoponcirin or poncirin. However, LC-HRMS was unable to

138

C:\Users\...\140526_m_nonspiked 26/05/2014 4:37:05 PM

RT: 0.00 - 20.02 SM: 7G RT: 13.49 NL: 9.83E6 100 TIC F: FTMS - p ESI Full 80 ms [120.00-1000.00] MS ICIS 60 140526_m_nonspiked determine whether the compound was neoponcirin or poncirin. RT:While 15.02 LC-HRMS provided40 accurate mass spectra for identification, LC-MS/MS was also implementedRT: 16.99 to RT: 12.18 further20 confirm these results. LC-MS/MS would not only enable observation of RT: the 18.08 Relative Relative Abundance RT: 2.02 RT: 5.84 RT: 10.50 molecular0 ions but also the product ions to provide a more definite result. RT: 13.49 NL: 1.18E6 100 m/z= 593.1859-593.1886 80 F: FTMS - p ESI Full ms [120.00-1000.00] MS 60 ICIS 140526_m_nonspiked 40

20 RT: 14.87 0 0 2 4 6 8 10 12 14 16 18 20 Time (min)

593.1859 NL: 100 1.46E6 A 140526_m_nonspiked#700- 713 RT: 13.36-13.55 AV: 7 50 F: FTMS - p ESI Full ms 285.0760 563.1755 [120.00-1000.00] 356.1292 435.1711 470.3136 524.0463 0 593.1876 NL: 100 7.12E5 B C28 H34 O14 +H: 50 C28 H33 O14 pa Chrg -1

0 Relative Relative Abundance 285.0768 NL: 100 8.30E5 C C16 H14 O5 +H: 50 C16 H13 O5 pa Chrg -1

0 300 350 400 450 500 550 600 m/z

Figure 28. Negative HESI spectrum of peak 6 (A) obtained from LC-HRMS Total ion current (TIC) chromatography and comparison with simulated spectra of molecular ion [M-H]- spectra of m/z 593 (B) and fragment ion [M-H]- of m/z 285 (C).

139

The multiple reaction monitoring (MRM) method is highly specific than LC-HRMS. This technique was utilised for confirmation of the chemical structures obtained by LC- HRMS by examining the product ions of reference standards and screening out the unwanted ions. The sensitivity and resolution of the analyte in the triple quadrupole mass spectrometer was optimised using an automatic tuning procedure. The MRM parameters of the target compounds were optimised by running full scan of reference standards by syringe infusion and the m/z values of precursor (Q1) ions were selected. The auto-tuning software enabled acquisition of four product (Q3) ions from the precursor masses and their respective collision energy voltages. These parameters were set into the instrumental method to maximise sample analysis. The results are presented in Table 20.

Both the negative ion HESI and the APCI modes were applied in the analysis of the samples because it was found that certain compounds, such as rosmarinic acid, chlorogenic acid and naringenin showed a higher signal-to-noise ratio on the mass spectra with the former, while others, e.g., hesperetin, apigenin and caffeic acid, exhibited a greater signal-to-noise ratio under the latter mode. Standard spiking was also performed on LC-MS/MS as well as LC-HRMS to further facilitate the confirmation process. Higher intensities of product ion were also observed when chlorogenic acid, caffeic acid, rosmarinic acid, apigenin, naringenin and hesperetin were “spiked” into the sample than the sample without spiking.

As shown in Table 20, peak 1 had a m/z of 353 in negative mode and 355 in positive mode; peak 4 had a m/z of 359 in negative mode and 361 in positive mode. Based on these parameters and co-chromatography with reference standards, it was confirmed that peak 1 was chlorogenic acid and peak 4 was rosmarinic acid. Four minor compounds, namely caffeic acid (peak 2), apigenin and naringenin (peak 7) and hesperetin (peak 8), were also identified and confirmed with reference standards.

140

Table 20. Phenolic compounds identified in methanol-based lyophilised extract of Mentha australis

HPLC Retention time Molecular mass Peak Compound Empirical formula (min, λmax) [M-H]- (MS/MS) (m/z) [M+H]+ (MS/MS) (m/z)

1,2,3 1 Chlorogenic acid C16H18O9 3.9 (325 nm) 353 (191.1) 355(180.0)

2,3 2 Caffeic acid C9H8O4 5.2 (322 nm) 179 (135.1) 181 (163.0)

3 3 Narirutin C27H32O14 8.6 (278 nm) 579 581

1,2,3 4 Rosmarinic acid C18H16O8 10.8 (321 nm) 359 (161.2) 361 (175.3)

3 5 Biochanin A C16H12O5 16.7 (268, 333 nm) 283 285

3 6 Neoponcirin C28H34O14 17.4 (281 nm) 593 595

2,3 7 Apigenin C15H10O5 23.4 (267, 335 nm) 269 (117.4) 271 (153.2)

2,3 7 Naringenin C15H12O5 23.4 (288 nm) 271 (119.4) 273 (147.0)

2,3 8 Hesperetin C16H14O6 23.7 (285 nm) 301 (164.1) 303 (152.8)

1Confirmed by HPLC with retention time, spectra comparison and co-elution with reference standard. 2Confirmed by LC-MS/MS. Cations are expressed as [M-H]- and anions are expressed as [M+H]+. Product ions expressed in brackets (MS/MS). HESI: rosmarinic acid, chlorogenic acid and naringenin; APCI: hesperidin, hesperetin, apigenin and caffeic acid. 3Confirmed by LC-HRMS with accuracy up to four decimal places.

139

141

However, peaks 3, 5 and 6 could not be determined by LC-MS/MS; therefore LC- HRMS was employed. In LC-HRMS analysis, peak 3 had a m/z of 579 in negative mode, while peak 5 had a m/z of 285 in negative mode. In addition, in the negative mode of peak 3, a fragment mass of m/z 271 was also found, which corresponded to naringenin. Based on the Chemspider database, peaks 3 and 5 were predicted to have an empirical formula of C27H32O14 and C16H12O5, respectively, with narirutin or naringin given as the most likely compounds for peak 3 and biochanin A for peak 5. However, peak 3 did not co-elute with the naringin reference standard on HPLC and LC-MS. Hence, narirutin was a more likely candidate for peak 3. The lambda max value of peaks 3, as obtained from the HPLC-PDA data, was 278 nm, which corresponded with that of naringenin glycosides, for which narirutin is a member. In the MRM of peak 3 in LC-MS/MS, product ions of naringenin were found to be present, which further demonstrated that peak 3 was a naringenin derivative.

Peak 5 had two absorption maxima, 268 and 333 nm, which corresponded with those of O-methylated isoflavones, of which biochanin A is a member. Based on these pieces of evidence, peak 3 was concluded to be most likely narirutin while peak 5 was most likely biochanin A, although further evidence was required to confirm their chemical structures. Due to these two peaks being rather small, fraction collection was not successful, which prevented NMR analysis to be conducted. In agreement with these findings, the presence of narirutin has been reported in Mentha species (Inoue et al., 2002, Guedon and Pasquier, 1994), while naringin is mainly found in citrus fruits and is responsible for their bitterness (Yusof et al., 1990).

Peak 6 could not be identified with LC-MS/MS. However, LC-HRMS analysis showed the presence of a molecular ion of m/z of 593 in negative mode with a fragment peak at m/z 285 (Figure 28). From the mass spectra, based on the Chemspider database, two possible compounds, poncirin or neoponcirin, were suggested. Both compounds have identical empirical formulae and hence molecular masses, which makes it challenging to be identified with chromatographic/mass spectrometric techniques. Moreover, standards of these two compounds were not available commercially to enable a comparison. In order to elucidate the chemical identity of peak 6, NMR analysis was performed (Section 5.2.4).

142

5.2.4. Confirmation by Nuclear Magnetic Resonance

5.2.4.1. 1H and 13C NMR analysis

NMR analysis provides information about the chemical nature of nuclei present in the molecule by giving them a characteristic peak with specific chemical shift (δ), whilst the coupling constant (J) of the peak can be used to deduce the interaction between the nucleus and their neighbouring partner (Field et al., 2012). Due to their universal occurrence in organic compounds, proton (1H) and carbon-13 (13C) nuclei are usually used as the targets of NMR analysis. In this study, only fractions corresponding to peaks

4 and 6 could be collected by preparative HPLC.

For peak 4, the 1H and 13C NMR spectra were obtained and compared to the reference standard as well as the literature to further confirm its identity as rosmarinic acid (Kim 1 et al., 2007, Mehrabani et al., 2010, Lu and Foo, 1999). A typical H NMR spectrum of reference standard rosmarinic acid is displayed in Figure 29. The chemical structure of rosmarinic acid consisted of two aromatic rings and a conjugated double bond. These two features of rosmarinic acid produced characteristic proton chemical shifts between 6-7 ppm (Table 21) and carbon-13 (13C) chemical shifts between 110-150 ppm (Table 22). Furthermore, position 7′ of rosmarinic acid consists of two protons attached to the carbon. The chemical shifts of the two protons (7′a and 7′b) varied slightly, exhibiting chemical shifts of 2.9 ppm and 3.1 ppm, respectively.

A simulation tool on the Mnova software was also used to compare rosmarinic acid (peak 4) with the predicted spectrum of the compound generated by the software. This software can generate the 1H and 13C NMR spectra by providing it with a drawn 2-D chemical structure of the predicted compound. Mnova Verify was another software used to confirm the chemical structure. This software is an automated structure verification step, which evaluates analytical data of the sample (including 1H, 13C and 2-D NMR spectra) and judges their compatibility with that of the proposed structure. The verification analysis confirmed the 1H NMR spectra of peak 4 as rosmarinic acid. In addition, it also demonstrated that peak 6 was neoponcirin.

143

Solvent peaks

Aromatic ring

7 8 8′ 7′a, 7′b

Figure 29. Chemical structure of rosmarinic acid indicating the protons at positions 3, 4, 3′, 4′, which cannot be detected by NMR (above). 1H NMR spectra showing the corresponding proton chemical shifts (below).

144

Table 21. Comparison of 1H NMR data of peak 4 with rosmarinic acid reference standard and literature

1H NMR δ, ppm (J, Hz)3

2 Position1 Rosmarinic acid Rosmarinic acid M. australis standard (peak 4) CD3OD CD3OD, CD3OD, 300 MHz 600 MHz 600 MHz

2 1H, d 7.17 (1.9) 7.04 (2.0) 7.04 (1.4)

5 1H, d 6.88 (8.8) 6.78 (8.2) 6.77 (8.2)

6 1H, dd 7.00 (2.0, 8.8) 6.95 (2.0, 8.3) 6.94 (1.2, 8.1)

7 1H, d 7.53 (16.0) 7.55 (15.9) 7.53 (15.8)

8 1H, d 6.34 (16.0) 6.27 (15.9) 6.26 (15.7)

2′ 1H, s 6.89 6.75 6.75

5′ 1H, d 6.75 (8.0) 6.69 (7.9) 6.69 (7.9)

6′ 1H, dd 6.71 (1.7, 8.0) 6.61 (1.9, 7.9) 6.61 (7.7)

7′a 1H, dd 2.95 (8.8, 13.8) 3.00 (8.4, 14.4) 2.93 (9.6, 14.2)

7′b 1H, d 3.16 (13.8) 3.10 (14.4) 3.09 (13.2)

8′ 1H, d 5.07 (8.8) 5.18 (8.5) 5.08 (7.3)

1Protons in hydroxyl groups cannot be detected. Multiplicity: singlet (s), doublet (d), doublet of doublet (dd). 2Data published by Lu and Foo (1999) 3Coupling constants are displayed in brackets.

145

Table 22. Comparison of 13C NMR data of peak 4 with rosmarinic acid reference standard and literature

13C NMR δ, ppm (J, Hz)

1 Position Rosmarinic acid Rosmarinic acid M. australis standard (peak 4) CD3OD, CD3OD, CD3OD, 75 MHz 150 MHz 150 MHz

1 129.5 127.7 128.0 2 117.3 115.2 115.1 3 143.4 145.3 144.8 4 145.6 146.2 146.0 5 118.5 116.3 115.7 6 124.1 121.8 121.7 7 148.7 149.7 149.4 8 117.0 114.4 115.0 9 171.4 168.5 168.5 1′ 133.1 129.3 129.1 2′ 119.6 117.6 117.5 3′ 146.8 146.8 146.6 4′ 147.5 147.7 146.7 5′ 118.7 116.5 116.4 6′ 125.2 123.2 122.9 7′ 39.9 37.9 38.8 8′ 79.3 74.6 74.7 9′ 179.5 173.5 173.3 1Protons in hydroxyl groups cannot be detected. 2Data published by Lu and Foo (1999) 146

The 1H NMR spectrum of reference standard rosmarinic acid was compared with the purified fraction of peak 4 from M. australis as displayed in Figure 30. This provided tentative evidence that the identity of peak 4 is rosmarinic acid. However, identification of peak 6 was more difficult due to the structural similarity between poncirin and neoponcirin. Typical 1H and 13C NMR spectra of the purified fraction of peak 6 are given in Appendices 8 to 9, respectively. As mentioned previously, both compounds contain the flavonoid backbone isosakuranetin, with the main difference being the sugar attachment as illustrated in Figure 31. Poncirin differs from neoponcirin in the attachment of rhamnose to the glucose, where in poncirin rhamnose is attached to glucose at position 2′′, whereas in neoponcirin rhamnose is attached to 5′′. This difference is believed to be more influential to the chemical environment of the sugar moieties than to that of the isosakuranetin backbone. Positions 2′′, 6′′ and 1′′′ of the sugar unit represented the main structural differences between poncirin and neoponcirin, while the changes to the isosakuranetin backbone were minimal. As shown on Table 23, difference in 1H chemical shifts occurred in positions 6′′ and 1′′′ of the sugar unit. However, the 1H chemical shift of position 2′′ was not reported in literature. Hence, the 13C NMR spectrum of the mint extract was acquired, which provided the additional information required to identify neoponcirin as the compound present in the mint (Table 24). The integration of 1H and 13C for peak 6 agreed with the molecular formula

(C28H34O14), which suggests that only one compound was present, but additional NMR experiments are required to determine whether it is poncirin or neoponcirin.

Further structural information of peak 6 was obtained from 1H and 13C NMR spectra and comparison of which was made to the literature (Kim et al., 2007, Chacón Morales et al., 2013, Kuo et al., 2000). From the 1H NMR data published by Kuo et al. (2000), the chemical shifts of neoponcirin were found to be consistent with those of peak 6. In particular, position 6a′′, 6b′′ and 1′′′ had a chemical shift of 3.60 ppm (4.9, 11.5 Hz), 3.75 ppm (2.7, 11.5 Hz), 4.69 ppm (s), while peak 6 had 3.67 ppm (3.5, 9.6 Hz), 3.87 ppm (1.6, 3.4 Hz) and 4.69 ppm (s), respectively (Table 23). Although the different types of solvent used caused a drift in chemical shifts, the identified peaks were sufficient to elucidate the chemical structure of peak 6. Based on the 13C NMR spectra published by Kim et al. (2007), position 2′′, 6′′ and 1′′′ have a chemical shift of 77.2, 60.4 and 97.4 ppm in poncirin and 73.0, 66.0 and 99.4 ppm in neoponcirin,

147 respectively (Table 24). The results of peak 6 – 73.2, 66.0 and 99.4 ppm for 2′′, 6′′ and 1′′′ respectively – implied its identity as neoponcirin.

148

Figure 30. 1H NMR spectra comparing reference standard rosmarinic acid (A) against the purified fraction of peak 4 from M. australis (B).

146

149

Figure 31. Structural similarity of the flavonoid backbone and differences in position 2′′ and 5′′ of the sugar component in poncirin, naringin, neoponcirin and hesperidin.

150

To confirm the structure of peak 6, NMR evidence of the sugar moiety was also needed. Since the disaccharide units of poncirin and neoponcirin were identical to naringin and hesperidin respectively, an alternative approach was conducted by employing naringin and hesperidin as reference materials to determine which disaccharide is attached to the flavonoid in peak 6. When NMR spectra of peak 6 were compared with those of naringin, the chemical shift and the coupling constant of the disaccharide section did not match well, which ruled out the possibility of peak 6 being poncirin. The NMR spectra of hesperidin were in agreement with those published in literature and showed a good match with those of peak 6 in the disaccharide section (Shi et al., 2008).

5.2.4.2. Two-dimensional NMR spectroscopy analysis

To obtain more information on the structure, three types of 2-D NMR analysis were conducted: Heteronuclear Single Quantum Coherence (HSQC), Heteronuclear Multiple- Bond Correlation (HMBC) and Correlation Spectroscopy (COSY). The identity of peak 4 was further confirmed as rosmarinic acid from HSQC, HMBC and COSY analysis, which provided information on the structure of a molecule. Simulated data was generated from the Mnova software and the 2-D analysis of the reference standard was also conducted and showed matching spectra between peak 4 and rosmarinic acid. Typical 2-D NMR spectra of peak 4 and peak 6 are given in Appendices 10 to 12.

5.2.4.3. Heteronuclear Single Quantum Coherence analysis

The HSQC spectrum provided information on the protons within the molecule and the carbon partner which it is attached to. Although there was no reference standard for neoponcirin or poncirin for comparison, HSQC analysis for peak 6 demonstrated that the chemical shifts of proton and carbon positions 2′′, 6′′ and 1′′′ of peak 6 corresponded to neoponcirin (Figure 32). In the 2-D diagram, the two hydrogens that connected to the carbon in position 6′′ are indicated by green colour. These signals distinctively indicate the structural difference present in the linkage between the two sugar units of neoponcirin as compared to that of poncirin. Furthermore, the diagram also indicates a

151 correlation between the proton and carbon at positions 2′′ and 1′′′, respectively. An example of a whole 2-D HSQC spectrum is shown in Appendix 10.

152

Table 23. Comparison of 1H NMR data of purified fraction of peak 6 with poncirin, neoponcirin and hesperidin

1H NMR δ, ppm (J, Hz)1

Position Poncirin Neoponcirin Hesperidin M. australis (Peak 6) DMSO-d6, CD3OD, DMSO-d6, CD3OD, 400 MHz2 300 MHz3 400 MHz4 600 MHz

2′, 6′ 2H, d 7.45 (8.7) 7.43 (8.8) 6.97 (2.0), 7.44 (8.6) 6.83 (2.0, 8.0) 3′, 5′ 2H, d 6.98 (8.7) 6.95 (8.8) 6.88 (8.0) 6.97 (8.8) 6, 8 2H, d 6.09 ( 2.1), 6.18 (2.3), 6.13 (2.0), 6.19 (2.3), 6.13 ( 2.1) 6.20 (2.3) 6.14 (2.0) 6.21 (2.3) 2 1H, dd 5.60 (2.9, 12.6 ) 5.46 (3.2, 12.6) 5.50 (5.0, 11.0) 5.47 (3.0, 12.7) 1′′ 1H, d 5.14 (7.2 ) 4.93 (7.3) 4.97 (7.2) 4.95 (7.4) 1′′′ 1H, s 4.59 4.69 4.54 4.69 OCH3 3H, s 3.77 3.80 3.78 3.81 6b′′ 1H, dd N.A. 3.75 (2.7, 11.5) - 3.87 (1.6, 3.4) 6a ′′ 1H, dd N.A. 3.60 (4.9, 11.5) - 3.67 (3.5, 9.6) 2′′-5′′ - N.A. N.A. 3.20-3.60 (m) 3.46-3.89 2′′′-5′′′ - N.A. N.A. 3.20-3.60 (m) 3.35-3.65 5′′′ 1H, d N.A. N.A. 2.51 (6.0) 3.65 (6.3) 3a 1H, dd 3.18 (12.6, 17.2 ) 3.19 (12.6, 17.2 ) 3.11 (11.0, 17.0) 3.18 (12.6, 17.2) 3b 1H, dd 2.78 (2.6, 17.2 ) 2.79 (3.2, 17.2 ) 2.78 (5.0, 17.0) 2.79 (3.1, 16.9) 6′′′ 3H, d 1.16 (6.2 ) 1.18 (6.3 ) 1.09 (6.0) 1.19 (6.2) 1N.A. = not available. Main difference between the structure in poncirin and neoponcirin is highlighted in grey. 2Data published by Kim et al. 3Data published by Kuo et al. 4 149 Data published by Shi et al.

153

Table 24. Comparison of 13C NMR data of purified fraction of peak 6 with neoponcirin and hesperidin

13C NMR δ, ppm (J, Hz)

Position Poncirin Neoponcirin Hesperidin M. australis DMSO-d6 DMSO-d6 DMSO-d6 (Peak 6) CD3OD 100 MHz1 100 MHz1 100 MHz2 150 MHz3

4 197.1 197.1 197.2 197.0 7 164.9 165.1 165.5 165.5 5 163.0 163.0 163.2 163.6 9 162.7 162.6 162.7 163.0 4′ 159.6 159.5 148.1 160.1 1′ 130.4 130.3 131.1 130.7 2′, 6′ 128.4 128.5 114.3, 118.1 127.7 3′, 5′ 113.9 113.9 146.6, 112.2 113.7 10 103.4 103.3 103.5 103.6 1′′ 100.4 100.6 99.6 100.7 1′′′ 97.4 99.4 100.8 99.7 6 96.4 96.5 96.5 96.6 8 95.2 95.5 95.7 95.7 2 78.4 78.3 78.6 79.0 5′′ 76.9 76.2 76.5 76.4 3′′ 76.1 75.6 75.7 75.7 2′′ 77.2 73.0 72.2 73.2 4′′′ 71.8 72.1 73.2 72.7 2′′′ N.A. N.A. 70.5 70.9 3′′′ 70.4 70.7 70.9 70.6 4′′ 69.6 69.6 69.7 69.9 5′′′ 68.3 68.3 68.5 68.4 6′′ 60.4 66.0 66.2 66.0

OCH3 55.2 55.2 55.8 54.4 3 42.1 41.7 42.2 42.6 6′′′ 18.1 17.8 18.1 16.5 1Data published by Kim et al. 2Data published by Shi et al. 3Main difference between the structure in poncirin and neoponcirin is highlighted in grey.

154

5.2.4.4. Heteronuclear Multiple Bond Coherence analysis

HMBC analysis enables identification of correlations between protons through bonds within the molecule and interactions of the protons with its neighbouring bonds that are separated by two or three bonds (Pinheiro and Justino, 2010). With this information, the structure could be assigned by connecting the proton to its carbon partner. This analysis further confirmed that peak 4 was rosmarinic acid (Table 25). Peak 6, which was more difficult to identify, was confirmed by HMBC analysis as neoponcirin. HMBC, in a similar fashion to HSQC, was performed to distinguish neoponcirin from poncirin by deducing the structural differences through the characteristic correlations between the proton and carbon signals of the respective compounds. That is, the protons at position 6′′ had the neighbouring carbons 1′′, 5′′, 1′′′ and 2′′′, and position 1′′′ had the neighbouring carbons 6′′, 2′′′, 3′′′ and 5′′′ (Table 26).

5.2.4.5. Correlation Spectroscopy analysis

COSY analysis shows the position of the proton and its neighbouring protons through its coupling pattern. This technique was used in conjunction with HSQC and HMBC result to deduce the structure of the unknown fraction of the mint extract that was obtained from HPLC. On the COSY spectrum, peak 4 showed consistent correlation with reference standard rosmarinic acid, in which the double bond at position 7 and 8 showed interactions. For peak 6, the protons at positions 3′-5′ and 2′-6′ in the aromatic ring were found to interact with each other while the protons at position 6′′′ were found to interact with those from positions 4′′′ and 5′′′. The protons from position 3α connected with 3β and also with protons at position 2. An example of the COSY spectrum is presented in Figure 33, where the boxes formed on the spectrum indicate interaction. As indicated by the collective results obtained from LC-HRMS and LC- MS/MS, peak 4 was identified as rosmarinic acid, while the NMR results of the fraction provided additional evidence to support this finding. For peak 6, in which no reference standard was available, NMR analyses were required to provide more concrete evidence in addition to the LC-HRMS and LC-MS/MS results in order to identify this fraction containing neoponcirin.

155

6′′

2′′

1′′′

Figure 32. HSQC spectrum of peak 6 of M. australis in CD3OD showing carbon position 2′′, 6′′ and 1′′′.

152

156

Table 25. Heteronuclear Single Quantum Coherence and Heteronuclear Multiple-Bond Correlation data of purified fraction of peak 4 (i.e., rosmarinic acid)

M. australis (Peak 4) HSQC δ, ppm, HMBC δ, ppm, CD3OD CD3OD Position 1H NMR 13C NMR 13C NMR 600 MHz 150 MHz 150 MHz

2 7.03 115.1 122.9, 146.4 5 6.77 116.4 38.8, 121.8, 128.0, 146.4, 149.4 6 6.92 121.7 115.1, 146.4 7 7.55 146.4 115.1, 122.9, 128.0, 169.0 8 6.27 115.7 38.8, 128.0, 169.0 2′ 6.76 117.5 38.8, 122.9, 128.0, 146.4, 149.4 5′ 6.66 116.2 146.0 6′ 6.61 122.9 38.8, 117.5 7′a 2.93 38.8 77.7, 117.5, 121.77 7′b 3.09 38.8 117.5, 121.77 8′ 5.08 77.7 38.8, 169.0

157

Table 26. Heteronuclear Single Quantum Coherence and Heteronuclear Multiple-Bond Correlation data of purified fraction of peak 6 (i.e., neoponcirin)

M. australis (Peak 6) HSQC δ, ppm (J, Hz), HMBC δ, ppm (J, Hz),

CD3OD CD3OD Position 1H NMR 13C NMR 13C NMR 600 MHz 150 MHz 150 MHz

2′, 6′ 7.43 129.1 80.4, 115.0, 129.1, 161.5 3′, 5′ 6.96 115.0 115.0, 132.1, 161.5 1′′ 4.95 101.0 166.9 1′′′ 4.69 102.1 67.4, 69.8, 71.3, 72.1 6 6.19 97.1 98.0, 105.0, 164.4, 166.9 8 6.21 98.0 97.1,105.0, 165.0, 166.9 2 5.47 80.4 129.1, 132.1 5′′ 3.45 77.9 71.3, 74.7 3′′ 3.59 77.1 67.4, 71.3 2′′ 3.45 74.7 77.9 4′′′ 3.33 74.1 69.8, 72.1 2′′′ 3.67 72.4 74.1 3′′′ 3.87 72.1 72.4, 74.1 4′′ 3.37 71.3 77.9 5′′′ 3.63 69.8 74.1 6b′′ 3.99 67.4 71.3, 102.1 6a ′′ 3.62 67.4 77.1, 102.1

OCH3 3.81 55.8 161.5 3a 3.17 44.0 80.4, 132.1, 198.4 3b 2.79 44.0 105.0, 198.4 6′′′ 1.19 17.9 69.8, 71.3, 74.1

158

2′, 6′ 3′, 5′ 7 8

Rosmarinic acid

3′, 5′

Neoponcirin 2′, 6′

Figure 33. COSY spectrum of crude extract of M. australis in CD3OD showing the protons relationship in the double bond of rosmarinic acid (peak 4) and benzene ring of neoponcirin (peak 6).

159

5.2.5. Quantification of phenolic compounds by High Performance Liquid Chromatography-Photodiode Array Detector analysis

The photodiode array detector (PDA) is a widely used detection technique in the identification of organic compounds, thus has been employed in this study. The phenolic compounds, observed under 280 nm and 320 nm wavelengths, in the purified extract were quantified using HPLC-PDA. For those phenolics where reference standards were not available, gallic acid was used as an alternative and the amount was expressed as milligram gallic acid equivalent per gram purified extract detected at 280 nm (Table 27). The main compounds in M. australis were rosmarinic acid (30.4%), neoponcirin (29.5%), chlorogenic acid (7.7%), narirutin (5.6%) and biochanin A (1.9%), while the minor compounds were caffeic acid, apigenin, hesperetin and naringenin, which were present at trace levels (<1%). The levels of minor compounds were too low to be accurately quantified by HPLC.

Table 27. Quantification of phenolic compounds identified in methanol-based lyophilised extract of Mentha australis by HPLC-PDA

Retention time Concentration Peak Compound 2 (min, λmax ) (mg per g purified extract)

1 Chlorogenic acid, 3.9 (325 nm) 15.4 ± 0.05

2 Caffeic acid 5.2 (322 nm) trace

3 Narirutin1 8.6 (278 nm) 27.2 ± 0.02

4 Rosmarinic acid 10.8 (321 nm) 160.4 ± 0.85

5 Biochanin A1 16.7 (268, 333 nm) 9.6 ± 0.06

6 Neoponcirin1 17.4 (281 nm) 145.0 ± 0.42

7 Apigenin 23.4 (267, 335 nm) trace

7 Naringenin 23.4 (288 nm) trace

8 Hesperetin 23.7 (285 nm) trace

1Values are expressed as mg gallic acid equivalent (GE) per g purified extract for compounds detected at 280 nm. The data represent the mean ± standard deviation of three replicates. 2HPLC-PDA: High Performance Liquid Chromatography-Photodiode Array Detector.

160

The peaks on the HPLC chromatogram have been integrated where the width value and peak detection are adjusted to select only the target peak eliminating overlapping peaks, such as biochanin A. However, the quantification of peak 1 may be slightly overestimated due to a small overlapping of this peak.

As mentioned previously, the phenolic compounds mostly reported in Mentha include rosmarinic acid, caffeic acid, chlorogenic acid, eriocitrin, luteolin and its derivatives luteolin-O-glucoside and luteolin-O-rutinoside (Dorman et al., 2003, Inoue et al., 2002, Guedon and Pasquier, 1994, Kosar et al., 2004). In the present study, luteolin and eriocitrin were not identified in M. australis, as reference standards of these two compounds did not match any peaks in the HPLC and GC-MS analyses. Neoponcirin has been found in a number of species in the Lamiaceae family (Kuo et al., 2000), but this is the first time it is identified in the Mentha genus. This compound represents 30.4% of total phenolics in the leaves of M. australis. This flavanone glycoside is one of the main constituents found in citrus fruit, particularly in trifoliate orange (Poncirus trifoliata) and is widely used as remedial medicine for inflammation, allergy, digestive ulcers and gastritis (Kim et al., 2007). Studies have shown that neoponcirin possesses antioxidant properties (Hung et al., 2010) and induces apoptosis in tumour cells (Hung et al., 2010, Singhal et al., 2012). In vitro and in vivo studies had reported this compound to be a chemotherapeutic agent for lung cancer (Hung et al., 2010). The very high level of neoponcirin present in the river mint would be expected to contribute significantly to the strong antioxidant activity, and possibly other biological activities, of this native Australian herb, as reported in Chapter 4.

High levels of rosmarinic acid have been reported in the Mentha genus in a number of studies (Dorman et al., 2003, McKay and Blumberg, 2006) as well as in the subfamily Nepetoidae of Lamiaceae (Petersen and Simmonds, 2003) and in the Boraginaceae family. Dorman et al. (2003) have reported in their study that most Mentha species contain rosmarinic acid at levels between 27.5-47.8%, while another study investigating 40 varieties of Mentha x piperita found that these species contain rosmarinic acid at levels between 14.2-28.3% (Guedon and Pasquier, 1994). The amount of rosmarinic acid in the Australian native mint, M. australis, obtained in the present study was comparable with those reported in other Mentha species.

161

The high levels of rosmarinic acid in Mentha species make a significant contribution to the total phenolic content and play an important role in the antioxidant activity of these herbs (Tepe et al., 2007). This is reflected in the present study. As shown in Chapter 4, both of the Mentha species (river mint and spearmint), which were rich in rosmarinic acid, exhibited higher antioxidant activity than the mint bush. Furthermore, rosmarinic acid is also reported to be largely responsible for a number of other biological activities, including anti-inflammatory (Shen et al., 2011, Petersen and Simmonds, 2003, Tepe et al., 2007, Pearson et al., 2010), antihepatitis (Tepe et al., 2007) and antitumor (Tepe et al., 2007) properties. This is also partially reflected in this study as the native mint, as well as the common spearmint, showed significant inhibition against the inflammation- related enzyme, hyaluronidase, as reported in Chapter 4.

In some Mentha species caffeic acid at levels below 1.2% and trace amount of apigenin have been reported (Dorman et al., 2003), whereas narirutin was found at levels below 1% (Guedon and Pasquier, 1994). This is also the first time biochanin A, which is usually found in red clover (Trifolium pratense L.), to be reported in the Mentha genus. Low levels of naringenin, hesperetin, apigenin and caffeic acid and their derivatives had also been reported in some Mentha species (Kosar et al., 2004, Guedon and Pasquier, 1994). These components, although relatively minor compared with rosmarinic acid and neoponcirin, would nevertheless expected to contribute to the biological activities of the native mint.

5.3. Conclusion

A combination of chromatographic, mass spectrometry and NMR techniques enabled the identification of nine phenolic acids and flavanone glycosides in M. australis. While HPLC-PDA and GC-MS provided partial information, the increased mass accuracy and resolution of LC-HRMS and the highly sensitive and specific LC-MS/MS, produced a more complete depiction of the identity of the phenolic compounds. The chemical structures of the major compounds were further confirmed by 1H and 13C NMR, including 2-D NMR analyses. The main compounds identified in this Australian native mint species were rosmarinic acid, neoponcirin, biochanin A, narirutin and chlorogenic acid, while the minor compounds found were caffeic acid, apigenin, 162 hesperetin and naringenin. Neoponcirin and biochanin A were identified for the first time in the Mentha genus. These compounds, especially rosmarinic acid and neoponcirin, are expected to contribute significantly to the biological activities, including antioxidant and hyaluronidase inhibitory activities, of the native mint, as reported in Chapter 4.

163

Chapter 6 Identification and quantification of phenolic compounds in Prostanthera rotundifolia R. Br.

6.1. Introduction

The two largest genera of the Lamiaceae family that are utilised as food and medicine are Mentha and Prostanthera. Unlike the Mentha genus, which can be found in many areas around the world, Prostanthera species are more uniquely Australian with at least 100 species endemic to Australia (Palá-Paúl et al., 2006). Most of the Prostanthera mint bushes are ornamental, though two of these species (Prostanthera rotundifolia, Prostanthera striatiflora) have been recorded to be used by the indigenous people as medicine as well as ointment to treat sores, skin diseases, aches and pains (Lassak and McCarthy, 2001), respiratory infections and malaise (Barr, 1993). Fresh leaves of Prostanthera incisa were also used as flavouring (Dupont et al., 2006, Agboola and Radovanovic-Tesic, 2002). Currently, P. rotundifolia (mint bush) and P. incisa (cut leaf) are commercially available as a culinary herb for seasoning and flavouring.

Mint bushes are known to contain a number of different types of phytochemicals, including essential oils and phenolics, among others. The essential oils from the Prostanthera genus are well studied. Their main component is cineole, which constitutes over 40% of steam-distilled essential oil in most species (Lassak and McCarthy, 2001, Fulton, 2000, Palá-Paúl et al., 2006, Southwell and Tucker, 1996). However, there is little information reported on the phenolic components of the Prostanthera genus. As mints from the Lamiaceae family have shown a history of being used for their health-benefiting properties (as discussed in Chapter 2), native Australian mint bushes of Prostanthera from the same family became an area worth studying, and the information generated would not only improve our knowledge on the phytochemistry of this native Australian food plant, but also provide a foundation for understanding their health-promoting bioactivities.

The main objective of this chapter was to identify and quantify the phenolic compounds in the Australian native species P. rotundifolia (mint bushes). The identification and quantification of the unknown chemical constituents were carried out by implementing a series of instrumental analytical techniques as described in Chapter 5. Briefly, HPLC- 164

PDA as well as GC-MS was used to establish well defined peaks for comparison with reference standards, where available, and generate preliminary structural information of the unknown compounds. This was followed by a combination of LC-HRMS and LC- MS/MS analyses to confirm the phenolic compounds identified by HPLC-PDA and GC- MS, and to elucidate the structures of those that were unable to be identified by those techniques. Finally, the major peaks identified by HPLC-PDA were collected using a preparative column for verification of the chemical structure by NMR.

165

6.2. Results and discussion 6.2.1. Determination of phenolic compounds in Prostanthera rotundifolia by HPLC- PDA analysis

The extracted polyphenolics of P. rotundifolia, purified by adsorbent material XAD-7 Amberlite® resins, were initially analysed by HPLC-PDA. To achieve good resolution of phenolic compounds, the same optimisation methods applied to Mentha australis (Section 5.2.1.1) for HPLC-PDA, as well as LC-HRMS, LC-MS/MS, were used for this species. Preliminary analysis was conducted on a 60 min elution program time at 1 mL min-1, which resulted in an elution profile of the purified extract (Figure 34). As can be seen, the target analytes started to elute from the column after 15 min at a gradient of 15% solvent A (pure acetonitrile) and 85% B (2.5% acetic acid in water, v/v). Therefore, in the modified elution procedure, this was used as the starting point of the gradient, and the elution program was shortened to 34 min with the flow rate reduced to 0.8 mL min-1. The flowrate was reduced from 1.0 mL min-1 as it provided a better resolution. In contrast to the case with M. australis, where the peaks were displayed most predominantly at 320 nm wavelength, in case of the P. rotundifolia extracts, the peaks were visible predominantly at 280 nm wavelength, suggesting a different composition of phenolic compounds in this genus. These modifications did not show a significantly improved peak resolution of the purified P. rotundifolia extract as expected, however this gradient method was used to collect fractions from the major peaks (Figure 35).

Identification of the major phenolic compounds in P. rotundifolia extract followed the same five-step identification and confirmation process as in the case of M. australis described in the previous chapter (Section 5.2.1.2). Preliminary identification was based on HPLC-PDA analysis, which included comparison of peak retention times with those of phenolic reference standards, co-elution with reference standards, and UV spectra analysis. To date, the composition of phenolic compounds in P. rotundifolia has not been reported. Information available in the literature indicates that several common phenolic compounds, such as rosmarinic acid, chlorogenic acid, quercetin, luteolin, eriocitrin and syringic acid, are often present in herbs from the Lamiaceae family. However, when phenolic standards of those compounds were applied to the HPLC-PDA system, none but one appeared to match any peaks of P. rotundifolia. One compound

166

(Peak 1) was successfully identified by this method as caffeic acid (retention time 5.1

min, λmax = 322 nm). The retention time, UV spectra and co-chromatography of peak 1 were all consistent with caffeic acid standard. The chemical identity of caffeic acid was further confirmed by GC-MS, LC-HRMS and LC-MS/MS. Subsequently, other techniques were employed to identify the remaining phenolic components in P. rotundifolia.

Absorbance (mAU) Absorbance

Figure 34. Initial HPLC chromatogram at 280 nm of purified P. rotundifolia extract.

167

Absorbance (mAU) Absorbance

Figure 35. HPLC chromatogram at 280 nm of purified P. rotundifolia extract: caffeic acid (1), p-coumaric acid (2), hesperidin (3), verbascoside (4), 1-O-β-ᴅ-glucopyranosyl sinapate (5), 4-methoxycinnamic acid (6), glucose ester of p-coumaric acid (7), and naringenin (8).

6.2.2. Identification and confirmation of phenolic compounds in P. rotundifolia by LC-HRMS and LC-MS/MS

Using LC-HRMS and simulation software in Xcalibur and Chemspider database, the eight peaks visible on the HPLC chromatogram (Figure 35) representing phenolic compounds of P. rotundifolia were initially tentatively identified as caffeic acid (peak 1), p-coumaric acid (peak 2), hesperidin (peak 3), verbascoside or isoverbascoside (peak 4), 1-O-β-ᴅ-glucopyranosyl sinapate (peak 5), 4-methoxycinnamic acid (peak 6), glucose ester of p-coumaric acid (peak 7) and naringenin (peak 8). Subsequently, four of the eight compounds, namely caffeic acid, p-coumaric acid, hesperidin and naringenin were successfully confirmed with reference standards by mass spectrometry.

The LC-HRMS chromatogram shows two major peaks, which eluted at retention times 13.0 min and 19.2 min (Figure 23). These two peaks corresponded, respectively, to peak 4 and peak 6 on the HPLC-PDA chromatogram. The molecular ion [M-H]- 623 168

m/z was most dominant at 13.0 min, while the molecular ion [M-H]- 177 m/z was most intense at 19.2 min (Figure 23). The identity of these two compounds is later confirmed later by NMR spectroscopy.

C:\Users\...\150318_P_negative_APCI 18/03/2015 1:44:04 PM

RT: 0.0 - 28.0 SM: 7G 13.0 NL: 1.03E8 100 A. TIC TIC F: FTMS - p APCI C:\Users\...\150318_P_negative_APCI corona Full ms [120.00-700.00] MS 80 19.2 RT: 0.0 - 28.0 SM: 7G 150318_P_negative_AP 13.0 100 19.2 26.1 CI 60 5.1 7.7 10.8 15.2 0 26.1 13.0 100 23.6 5.1 40 0 2.8 7.2 8.6 10.8 21.0 0.7 19.2 100 13.0 14.2 16.9 21.8 Relative Abundance 3.6 26.4 20 0 0 5 10 15 20 25 Time (min) 0 623.1950 13.0 NL: 1.09E6 100 100 B. 623 m/z m/z= 623.1951-623.1971 C:\Users\...\150318_P_negative_APCI80 F: FTMS - p APCI corona 80 60 Full ms [120.00-700.00] RT: 0.0 - 28.0 SM: 7G MS 40 13.0 100 19.2 26.1 150318_P_negative_AP 5.1 7.7 10.8 654.2299 60 20 15.2 CI Relative Abundance 0 504.6221 678.4046 541.2481 13.0583.5144 1000 500 550 600 650 40 0 m/z 19.2 100 13.0 21.8 Relative Abundance 3.6 26.4 RelativeAbundance 20 0 0 5 10 15 20 25 Time (min) 11.2 0 19.2 NL: 1.05E5 100 177.0188 100 m/z= 177.0547-177.0567 80 177.0551 C. 177 m/z F: FTMS - p APCI corona 80 60 Full ms [120.00-700.00] MS 40 13.0 150318_P_negative_AP 60 20

Relative Abundance 177.0776 CI 176.5296 176.9495 0 40 176.5 177.0 m/z 21.8 20 3.6 16.0 23.4 11.6 21.0 26.4 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 Time (min)

Figure 36. LC-HRMS total ion current (TIC) chromatogram (A) of P. rotundifolia using atmosphere pressure chemical ionisation with mass range between m/z 120-700 (mobile phase: acetonitrile and acetic acid). A [M-H]- mass of 623 (B) and 177 (C) corresponded to the major peaks from the TIC chromatogram.

169

On the simulated mass spectra generated by the Xcalibur software, the [M-H]- mass of 623 (peak 4) showed a consistent mass to that produced by the molecular formula

C29H36O15 (Figure 37), which corresponded to severeal chemical structures. While, the [M-H]+ mass of 179 of the second major peak (peak 6) displayed a mass matching that of the molecular formula C10H10O3 (Figure 38), which corresponded to at least five other compounds. At this stage, these two compounds can only be tentatively identified, however these structures will later be elucidated by NMR analyses.

Table 28 shows the molecular ions for each compound as identified by LC-HRMS and the corresponding peaks on the HPLC chromatogram. However, for compound represented by peak 5, LC-HRMS could only identify it as 1-O-β-ᴅ-glucopyranosyl sinapate by negative ionisation using both the HESI and APCI modes. As for peak 4, there were two possibilities provided by LC-HRMS analysis – verbascoside or isoverbascoside. The latter compound is an isomer of the former and, therefore, the two compounds were difficult to separate with chromatographic methods. Further analysis was conducted using tandem mass spectrometry and gas chromatography to ascertain the chemical structures of the suspected compounds.

Table 28 also shows the fragment anions and cations of the phenolic compounds in P. rotundifolia obtained from LC-MS/MS. With the LC-MS method, some compounds are more sensitive with one type of ionisation, such as HESI or APCI, while not as well by another type (Cuyckens and Claeys, 2004). In most studies on herbs, positive ion mode was applied; however, it was reported that negative ion mode using both APCI and electrospray ionisation was more sensitive in flavonoid analysis (Cuyckens and Claeys, 2004, Justesen, 2000, Shen et al., 2011). This was also found in the present study, where it was observed that under negative ion HESI mode, standard naringenin showed a higher signal-to-noise ratio on the mass spectra, while under negative ion APCI mode, hesperetin, p-coumaric acid and caffeic acid exhibited a greater signal-to-noise ratio. Using both ionisation modes, as well as co-chromatography with the reference standards on LC-MS/MS and LC-HRMS, compounds represented by peaks 1, 2, 3, and 8 were identified to be, respectively, caffeic acid, p-coumaric acid, hesperidin and naringenin.

170

623.1963 NL: 100 2.23E5 P. rotundifolia 140526_p_nonspike 80 d#521 RT: 9.93 AV: 1 F: FTMS - p ESI Full ms 60 [120.00-1000.00]

C:\Users\...\150318_P_positive_APCI40 18/03/2015 2:44:22 PM

625.2020 RelativeAbundance RT: 0.0020 - 28.00 19.14 24.79 NL: 100 377.1805 621.1812 335.1125 461.1649 509.2224 549.1958 653.2064 7.04E7 0 623.1981 NL: TIC MS 10090 7.03E5 150318_P_ C 29 H36 O15 +H: positive_AP 80 80 C 29 H35 O15 CI Simulated data pa Chrg -1 70 60 23.20 60 4.67 23.04 25.07 40 50 22.46 25.51 12.99 20 22.37 26.31 40 12.87 Relative Relative Abundance 3.55 5.11 15.99 628.2125 300 300 350 400 450 500 550 16.75600 650 m/z 13.12 12.39 20 7.96 11.14 1.62 2.48 6.80 10

0 - Figure 37.0 Comparison2 4 of molecular6 8 ion10 [M-1]12 (m/z 14623) of16 peak18 4 (C2029H36O2215) in 24P. 26 rotundifolia and the simulated data generated from TimeXcalibur (min) by LC-HRMS.

179.0691 NL: 100 2.04E6 P. rotundifolia 80 150318_P_positive_APCI#13 01-1317 RT: 19.01-19.24 60 AV: 17 T: FTMS + p APCI corona Full ms 178.0851 40 [120.00-700.00] 178.0609 20 Relative Relative Abundance 177.0899 180.0724 169.1212 172.1321 175.0742 182.0800 184.0954 0 179.0703 NL: 100 Simulated data 8.90E5 80 C10 H11 O3: C10 H11 O3 60 pa Chrg 1

20 180.0736 181.0745 183.0812 0 170 172 174 176 178 180 182 184 m/z

+ Figure 38. Comparison of molecular ion [M-1] (m/z 179) of peak 6 (C10H10O3) in P. rotundifolia and the simulated data generated from Xcalibur by LC-HRMS.

171

Table 28. Phenolic compounds identified in methanol-based lyophilised extract of Prostanthera rotundifolia

Molecular mass

2 Empirical HPLC Retention time Peak Compound formula (min, λ ) - + max [M-H] (MS/MS) (m/z) [M+H] (MS/MS) (m/z)

1,2 1 Caffeic acid C9H8O4 5.1 (322 nm) 179 (135.1) 181 (163.0)

1,2 2 p-coumaric acid C9H8O3 6.4 (310 nm) 163 (119.2) 165 (91.2)

1,2 3 Hesperidin C28H34O15 10.4 (283 nm) 609 (301.0) 611 (302.9)

2 4 Verbascoside C29H36O15 15.4 (329 nm) 623 625

2 5 1-O-β-ᴅ-glucopyranosyl sinapate C17H22O10 17.0 (312 nm) 385 ND

2 6 4-methoxycinnamic acid C10H10O3 18.6 (312 nm) 177 179

2 7 Glucose ester of p-coumaric acid C15H18O8 19.6 (312 nm) 325 327

1,2 8 Naringenin C15H12O5 23.0 (288 nm) 271 (119.4) 273 (147.0)

1Confirmed by LC-MS/MS. Cations are expressed as [M-H]- and anions are expressed as [M+H]+. Product ions expressed in brackets (MS/MS). HESI: naringenin; APCI: hesperidin, p-coumaric acid and caffeic acid. ND = not detected. 2Confirmed by LC-HRMS with accuracy up to four decimal places.

168

172

The structures of the remaining compounds (peaks 4, 5, 6 and 7, Figure 35) could not be determined by LC-MS/MS due to the unavailability of reference standards. Peak 4 and peak 6 on the HPLC chromatogram represented the major compounds, and hence their fractions were collected by preparative HPLC for NMR analysis to determine their chemical identities (Section 6.2.3). However, peaks 5 and 7 were minor peaks and the fractions of these peaks could not be collected for NMR analysis. The corresponding identity of these two peaks was determined by LC-HRMS and GC-MS.

In the negative mode of LC-HRMS under APCI and HESI ionisation method, compounds represented by peaks 5 and 7 produced, respectively, m/z of 385 and m/z

325, with the predicted empirical formula of C17H22O10 and C15H18O8. The maximum detection wavelength (λmax) of peak 5, as obtained from the HPLC-PDA data, was 312 nm, which corresponded with those of sinapic acid derivatives, for which 1-O-β-ᴅ- glucopyranosyl sinapate is a member. Peak 7 had an absorption maximum at 312 nm, which matched with the spectra of p-coumaric acid derivatives. These results are in agreement with the GC-MS analysis data, which showed the fragments of the two phenolic acids in the mass spectra. Subsequently, peak 5 was concluded to be most likely 1-O-β-ᴅ-glucopyranosyl sinapate while peak 7 was most likely a p-coumaric acid derivative. Furthermore, the product ion of p-coumaric acid in the mint extract produced very intense response signals on LC-MS/MS. This indicated the presence of a glucose ester of p-coumaric acid, the fragmentation of which generated the high response of p- coumaric acid. However, further evidence is required to confirm their chemical structures.

6.2.3. Determination of phenolic compounds by Gas Chromatography-Mass Spectrometry

GC-MS analysis provided complementary information to the LC-HRMS and LC- MS/MS data. Figure 39 displays the total ion current GC-MS chromatogram of non- hydrolysed (A) and hydrolysed (B) crude extracts of P. rotundifolia. It can be seen that, after hydrolysis, more peaks appeared on the chromatogram as a result of the fragmentation of larger compounds or compounds that were bound to sugar molecules. Table 29 presents the identified molecular and fragmented ions. The obtained mass spectra of the silylated phenolic compounds were compared with those of the reference 173 standards and the NIST 2.0 library, which further confirmed the identities of caffeic acid (peak 1), p-coumaric acid (peak 2) and 4-methoxycinnamic acid (peak 6). However, naringenin (peak 8) was only tentatively identified by retention time and molecular mass comparison. The NIST library was limited to only certain silylated compounds, which was the reason naringenin could not be identified using the NIST library, hence the reverse match factor could not be obtained.

For verbascoside (peak 4), 1-O-β-ᴅ-glucopyranosyl sinapate (peak 5) and glucose ester of p-coumaric acid (peak 7), only their aglycones were identified. The mass spectra of the hydrolysed extract showed higher intensities for caffeic acid and p-coumaric acid than those of the non-hydrolysed extract. The increased level of caffeic acid might have resulted from the breaking down of verbascoside which comprises of caffeic acid, while p-coumaric acid may be produced from the fragmentation of glucose ester of p- coumaric acid. Sinapic acid and cinnamic acid ions were also found in the GC mass spectra, which may be derived from 1-O-β-ᴅ-glucopyranosyl sinapate and 4- methoxycinnamic acid, respectively. Furthermore, using the Xcalibur software, empirical formula were generated from the GC-MS spectra, which were found to match these silylated phenolic compounds on the Chemspider database, thus confirming the identities of the compounds obtained by GC-MS.

Recently, silyl derivatisation of non-volatile extracts have proven to be useful for identifying phenolic acids and flavonoids by GC-MS analysis (Zhang and Zuo, 2004, Proestos and Komaitis, 2013). However, this may not be the case when more complex phenolic compounds are involved such as flavonoid glycosides. The fragmentation patterns of these silyl derivatised compounds are often difficult to interpret and unavailable in the NIST library for comparison (Cuyckens and Claeys, 2004). More research is required for identification of complex silyl derivatives (Proestos et al., 2006). As a result, few phenolic compounds could be identified and confirmed using this method.

174

C:\Users\...\2013-10-24_KittyTang\P 24/10/2013 8:35:31 PM

RT: 0.00 - 109.07 56.54 NL: 100 6.67E9 TIC MS P 90 A 80 53.77 70

64.91 40

RelativeAbundance 63.34 30 91.67 20 88.36 31.77 51.99 10 32.42 69.44 85.27 92.48 7.82 28.67 47.28 72.29 26.16 95.93 0 0 10 20 30 40 50 60 70 80 90 100 Time (min) C:\Users\...\2013-10-24_KittyTang\Pa 24/10/2013 10:29:31 PM

RT: 0.00 - 109.06 64.94 NL: 100 6.87E9 B TIC MS Pa 90 55.85

60 54.89 50 60.82

40 53.42 RelativeAbundance 30

20 65.64 51.98 69.44 88.25 15.66 91.62 10 34.39 35.91 38.36 20.56 73.34 82.02 92.83 7.89 100.29 0 0 10 20 30 40 50 60 70 80 90 100 Time (min)

Figure 39. Total ion current (TIC) GC-MS chromatogram of non-hydrolysed (A) and hydrolysed (B) crude extract of P. rotundifolia

175

Table 29. Molecular ions present in the mass spectra of silylated phenolic compounds1 identified by Gas Chromatography-Mass Spectrometry of P. rotundifolia

Retention Molecular ion Reverse match Peak Silylated Compound MW2 Fragment ions (%) Time (min) [M]+ factor

1 Caffeic acid 63.08 396 396 873 191(19), 219(100), 396(89)

2 p-Coumaric acid 57.59 308 308 934 219(89), 249(84), 293(100), 308(65)

6 4-Methoxycinnamic acid 53.78 250 250 850 161(100), 235(82), 191(74), 250(52), 133(23), 118(21), 121(20), 236(15) 8 Naringenin 86.91 488 488 - 28(39), 32(17), 45(14), 473(12), 296(11), 44(10), 147(9) - Cinnamic acid 35.09 220 220 718 103 (63.4), 131 (83.7), 161 (98.4), 205(100), 220 (31.8) - Sinapic acid 56.40 368 368 846 161 (39), 249(25), 338(100), 353(57) , 368 (64) 1Identified as trimethylsilyl derivatives. 2MW: molecular weight of silylated compound

172

176

6.2.4. Confirmation by Nuclear Magnetic Resonance Spectroscopy 6.2.4.1. 1H and 13C NMR analysis

Peaks 4, 6 and 7 of the HPLC chromatogram represented compounds in sufficient amount that could be collected by preparative HPLC for NMR analysis. Prior to NMR analysis, the isolated fractions of peaks 4 and 6 were injected into LC-MS/MS to verify their [M-H]- masses, which were, respectively, 623, 177 and 325.

1H and 13C NMR spectra were obtained for compounds represented by peaks 4 and 6 and were compared with literature information to confirm or disapprove their identities being verbascoside (peak 4), 4-methoxycinnamic acid (peak 6) and glucose ester of p- coumaric acid. The chemical structure of the verbascoside and its isomer is displayed in Figure 40 to facilitate discussion. To distinguish verbascoside from isoverbascoside, their 1H and 13C NMR data was compared with the literature. As shown in Table 30, peak 4 showed consistent coupling constants and splitting patterns to those of verbascoside as reported by Gómez-Aguirre et al. (2012). Although verbascoside and isoverbascoside have the same molecular formula, the two compounds exhibit distinguishable spectral differences which were shown primarily through the 1H NMR results. This is evident by comparing a previous NMR study conducted by Gómez- Aguirre et al. (2012) of the two isomers. Although the 1H at positions 6′′′ and 8 are chemically identical, their chemical shifts, as suggested by the authors, show adequate differences to help identify the compound that exist in the samples. For instance, in terms of chemical shift, the proton arrangement of verbascoside differs from its isomer in the aglycone attachment at position 8, the glucose attachment at positions 4′ and 6′, and the rhamnose at position 6′′′ as indicated in Figure 40. The main difference in the carbon arrangement is highlighted at position 6′ with verbascoside (62.4 ppm) exhibiting a lower chemical shift than that of isoverbascoside (65.1 ppm).

As for peak 6, its 1H and 13C NMR data matched those of 4-methoxycinnamic acid (Table 32) as reported in the literature (Sobolev et al., 2006), providing another piece of evidence of confirmation for its identity. While the 1H NMR data of peak 7 matched those of glucose ester of (E)-p-coumaric acid (Zhang et al., 2008), 13C NMR data of peak 7 was unavailable in literature. However, as displayed on Table 33, a close resemblance was observed when peak 7 was compared with the 13C NMR data of p-

177 coumaric acid (Sobolev et al., 2006). With the combination of LC-MS and NMR information, peak 7 was identified as glucose ester of (E)-p-coumaric acid.

A) Verbascoside

B) Isoverbascoside

Figure 40. Chemical shift differences as given in Table 30 between verbascoside (A) and its isomer isoverbascoside (B).

178

Table 30. Comparison of 1H NMR data of peak 4 with verbascoside and isoverbascoside

1H NMR δ, ppm (J, Hz) Verbascoside2 Isoverbascoside2 P. rotundifolia 1 CD OD (Peak 4) CD OD Position CD3OD 3 3 500 MHz 300 MHz 600 MHz

Aglycone 2 6.69 (2.0) 6.68 (2.1) 6.70 (2.0) 5 6.66 (8.0) 6.66 (7.5) 6.67 (8.0) 6 6.55 (2.0,8.0) 6.55 (2.4, 7.8) 6.57 (2.0,8.0) 7a 2.76 (3.5,7.0,14.0) 2.78 m 2.80 m 7b 2.80 (3.0,7.0,14.0) 2.78 m 2.80 m 8a 3.71 (6.5,7.0,8.0) 3.68 m 3.73 m 8b 4.04 (3.0,6.5,8.0) 3.72 m 4.05 m Glucose 1′ 4.36 (8.0) 4.36 (8.1) 4.37 (8.0) 2′ 3.38 (9.5,7.5) 3.35 m 3.39 3′ 3.81 (9.3) 3.81 (3.0, 10.3) 3.81 4′ 4.91 (9.5) 3.4-3.5 m 4.91 (9.5) 5′ 3.52 m 3.60 m 3.54 m 6′a 3.54 m 4.46 m 3.55 m 6′b 3.60 (10) 4.28 m 3.62 Caffeic acid 2′′ 7.04 (2.0) 7.04 (2.1) 7.05 (2.1) 5′′ 6.77 (8.0) 6.76 (8.1) 6.78 (8.1) 6′′ 6.95 (2.0,8.0) 6.96 (2.7, 8.1) 6.96 (2.1,8.1) 7′′ 7.58 (16.0) 7.59 (16.2) 7.59 (16.0) 8′′ 6.26 (16.0) 6.26 (15.9) 6.27 (15.9) Rhamnose 1′′′ 5.18 (2.0) 5.17 s 5.19 (1.9) 2′′′ 3.91 (2.0,3.0) 3.90 m 3.92 m 3′′′ 3.56 (3.0,9.5) 3.57 m 3.55 m 4′′′ 3.28 (9.5) 3.5-3.8 m 3.29 m 5′′′ 3.55 m 4.00 m 3.58 m 6′′′ 1.08 (6.0) 1.17 (6.0) 1.09 (6.2) 1See Figure 40 for numbering system of the structure. The major difference in chemical shifts between verbascoside and isoverbascoside are in bold. 2Data published by Gómez-Aguirre et al. (2012).

179

Table 31. Comparison of 13C NMR data of peak 4 with verbascoside

13C NMR δ, ppm (J, Hz)

Verbascoside2 Isoverbascoside2 P. rotundifolia 1 CD OD (Peak 4) CD OD Position 3 CD3OD 3 125 MHz 75 MHz 150 MHz Aglycone 1 131.5 131.4 131.5 2 117.1 117.1 117.1 3 146.1 146.1 146.1 4 144.6 144.7 144.7 5 116.3 116.3 116.3 6 121.2 121.2 121.2 7 36.5 36.6 36.6 8 72.2 72.2 72.2 Glucose 1′ 104.2 104.3 104.2 2′ 76.2 76.3 76.0 3′ 81.6 78.1 81.6 4′ 70.6 70.4 70.6 5′ 76 76.0 76.0 6′ 62.3 65.1 62.4 Caffeic acid 1′′ 127.6 127.6 127.6 2′′ 115.2 115.2 115.3 3′′ 146.8 146.8 146.8 4′′ 149.7 149.8 149.8 5′′ 116.5 116.5 116.5 6′′ 123.1 123.2 123.2 7′′ 148 148.0 148.0 8′′ 114.7 114.7 114.7 9′′ 168.2 167.9 168.3 Rhamnose 1′′′ 103.0 102.2 103.0 2′′′ 72.3 72.3 72.3 3′′′ 70.4 70.4 70.4 4′′′ 73.8 73.8 73.8 5′′′ 72 72.0 72.0 6′′′ 18.4 18.0 18.4 1See Figure 40 for numbering system of the structure. The major difference in chemical shifts between verbascoside and isoverbascoside is in bold. 2Data published by Gómez-Aguirre et al. (2012).

180

Table 32. Comparison of 1H and 13C NMR data of peak 6 with 4-methyoxycinnamic acid

1H and 13C NMR δ, ppm (J, Hz)

trans-4-methoxycinnamic acid1 P. rotundifolia Position CD3OD (Peak 6) CD3OD 1H NMR 13C NMR 1H NMR 13C NMR 400 MHz 100 MHz 600 MHz 150 MHz

1 - - 128.6 - 128.38 2, 6 d 6.96 (8.8) 115.6 6.98 (8.8) 115.53 3, 5 d 7.54 (8.8) 131.0 7.62 (8.7) 131.21 4 - - 163.2 - 163.26 7 d 7.63 (16.0) 146.3 7.69 (16.0) 146.31

8 d 6.33 (16.0) 116.8 6.50 (15.9) 116.37 9 - - 171.0 - 169.08

OCH3 s 3.83 56.0 3.84 55.88

1Data published by Sobolev et al. (2006).

181

Table 33. Comparison of 1H-NMR and 13C-NMR data of peak 7 with glucose ester of p-coumaric acid

1 1′

1H- and 13-C-NMR δ (J, Hz) P. rotundifolia CD3OD 1 13 DMSO-d6 CD3OD H-NMR C-NMR Position 1 H-NMR 13C-NMR 600MHz 150MHz 400MHz1 100MHz2 p-coumaric acid 1 128.6 128.38 2, 6 d 7.58 (8.6) 131.0 7.62 (8.6) 131.21 3, 5 d 6.80 (8.6) 115.6 6.98 (8.7) 115.53 4 163.2 163.26 7 d 7.64 (15.9) 146.3 7.69 (15.9) 146.31 8 d 6.40 (15.9) 116.8 6.50 (15.9) 116.37 COOH 171.0 169.08 Glucose 1′ d 5.46 (8.0) - 4.96 (8.0) 100.34 2′ - 3.35 74.42 3′ 3.37-3.48 - 3.38 78.65 m 4′ - 3.27 71.97 5′ - 3.74 74.66 6′α dd 3.65 (5.6, - 3.68 (6.9, 11.9) 12.0) 63.09 6′β dd 3.44 (5.6, - 3.96 (2.2, 11.9) 12.0)

1Data from Zhang et al. (2008) - glucose ester of (E)-p-coumaric acid 2Data published by Sobolev et al. (2006) – only showing p-coumaric acid.

182

6.2.4.2. Heteronuclear Single Quantum Coherence analysis

As per chapter 5, HSQC was used to elucidate the chemical structure of the compound represented by peak 4 of the HPLC chromatogram. As highlighted previously, there are several distinct areas in positions 4′ and 6′ in the structure of verbascoside that differ from isoverbascoside (Figure 40), and these differences would be reflected in their NMR spectra. Table 34 presents the HSQC results for peak 4. The results showed that the chemical shifts of proton and carbon and their connectivities at positions 8, 4′, 6′ and 6′′′ corresponded to those of verbascoside when compared with literature (Gómez- Aguirre et al., 2012). Figure 41 illustrates an example of position 6′ of the compound in peak 4 on the HSQC spectra. On the spectra, the two highlighted points indicated that there were two protons with chemical shifts of 3.55 and 3.62 ppm being attached to the same carbon (62.36 ppm), which are characteristics of position 6′ or the CH2 group in the structure of verbascoside (Gómez-Aguirre et al., 2012). In comparison, the two protons (4.46 and 4.28 ppm) and carbon (65.1 ppm) of isoverbascoside at position 6′ exhibited higher chemical shifts as highlighted earlier. This essentially ruled out the compound represented by peak 4 to be the isomer isoverbascoside. A full HSQC spectrum of the identified verbascoside is displayed in Appendix 14.

Similarly, the identity of peak 6 was confirmed by HSQC analysis (Table 35). The 2-D spectrum showed the connectivities within the benzene ring at positions 3 and 5 (1H = 7.62 ppm; 13C = 131.21 ppm) and positions 2 and 6 (1H = 6.98 ppm; 13C = 115.53 ppm) (Figure 42). The double bond at position 7 (1H = 7.69 ppm; 13C = 146.31 ppm) and position 8 (1H = 6.50 ppm; 13C = 116.37 ppm) was also found to have connectivities that is matching to those reported by Sobolev et al. (2006). The protons within the

OCH3 group were found to have connectivity with the carbon in the same group, displaying a chemical shift of 1H = 3.84 ppm and 13C = 55.88 ppm. These results agreed with LC-HRMS data, confirming the tentatively identified compound to be consistent with 4-methoxycinnamic acid.

183

Figure 41. HSQC spectrum of peak 4 indicating position 6′ of verbascoside.

184

2, 6 8

3, 5

Figure 42. HSQC spectrum of peak 6 indicating benzene ring positions 3 and 5, positions 2 and 6, position 7 and position 8.

185

6.2.4.3. Heteronuclear Multiple Bond Coherence (HMBC) analysis

Although HSQC could determine which protons are bound to certain carbons in a compound, HMBC could provide further insight into the relationship between a proton and its neighbouring carbons. Usually a proton can connect up to four carbons (Field et al., 2012). The combined information from HSQC and HMBC analysis would thus enable the full structural form of the compound to be constructed. Results of the HMBC analysis of peak 4 and peak 6 are respectively displayed in Figure 34 and Figure 35 along with HSQC information. For example, the results for peak 4 illustrated that the protons at position 4′ (1H = 4.91) were connected to the neighbouring carbons 2′ (13C = 74.65), 6′ (13C = 166.87) and 3′ (13C = 80.23). While the two protons at position 6′ (1H = 3.55 and 3.62) had the neighbouring carbons 1′ (13C = 102.81) and 4′ (13C = 69.17). The same analysis was conducted on positions 6′′′ and 8, which are the regions that differ between verbascoside and its isomer. The results show that position 6′′′ (1H = 1.09) had the neighbouring carbons 3′′′ (13C = 69.02) and 4′′′ (13C = 72.38). As for the two protons at position 8 (1H = 3.73 and 4.05), the neighbouring carbons were at positions 7 (13C = 35.17), 1′ (13C = 102.81) and 1 (13C = 130.07).

As for peak 6, the HMBC data show the connectivities within the benzene ring, the double bond and the carbon at the COOH region. Within the benzene ring, the protons at position 2 and 6 (1H = 6.98) had the neighbouring carbons 2 and 6 (13C = 115.53), 1 (13C = 128.38) and 4 (13C = 163.26). Whereas, protons at positions 3 and 5 (1H = 7.62) showed connectivities with carbon at positions 3 and 5 (13C = 131.21), 7 (13C = 146.31) and 4 (13C = 163.26). The protons at the double bond region of positions 7 (1H = 7.69) displayed connectivities with positions 2 (13C = 116.37), 1 (13C = 128.38), benzene region – 3 and 5 (13C = 131.21) and position 9 (13C = 169.08), and at position 8 (1H = 6.50), neigbouring carbons 1 (13C = 128.38) and 9 (13C = 169.08) were identified. 1 Furthermore, the three protons at the OCH3 region ( H = 3.84) showed connectivities with carbon 4 (13C = 163.26). These results further indicated that peak 4 was verbascoside and not isoverbascoside, and peak 6 was 4-methoxycinnamic acid.

186

Table 34. Heteronuclear Single Quantum Coherence and Heteronuclear Multiple-Bond Correlation data of purified fraction of peak 4 (i.e., verbascoside)1

P. rotundifolia (Peak 4) HSQC (δ, ppm) HMBC (δ, ppm) CD3OD CD3OD Position 1H NMR 13C NMR 13C NMR 600 MHz 150 MHz 150 MHz Aglycone 2 6.70 115.70 35.17, 119.84, 130.07, 143.29, 144.74 5 6.67 114.89 143.29, 144.74 6 6.57 119.84 35.17, 143.29 7a 2.80 35.17 70.87, 115.70, 119.85, 130.07 7b 2.80 8a 3.73 70.87 35.17, 102.81, 130.07 8b 4.05, Glucose 1′ 4.37 102.81 70.87 2′ 3.39 74.81 80.23, 102.81 3′ 3.81 80.23 69.17, 74.81, 101.64 4′ 4.91 69.17 74.65, 80.23, 166.87 5′ 3.93 76.41 69.17 6′a 3.55 62.36 69.17, 102.81 6′b 3.62 Caffeic acid 2′′ 7.05 113.80 121.82, 145.46, 146.63, 148.41 5′′ 6.78 115.10 126.25, 145.46, 148.41 6′′ 6.96 121.81 113.80, 146.63, 148.41 7′′ 7.59 146.61 113.82, 121.81, 126.25, 166.87 8′′ 6.27 113.29 126.25, 166.87 Rhamnose 1′′′ 5.19 101.64 69.02, 70.64, 80.23 2′′′ 3.92 70.95 101.64 3′′′ 3.55 69.02 69.17, 74.65, 102.81 4′′′ 3.29 72.38 69.17, 70.64 5′′′ 3.54 74.65 17.05, 72.38 6′′′ 1.09 17.05 69.02, 72.38 1Full HMBC spectrum is displayed on Appendix 15. 187

Table 35. Heteronuclear Single Quantum Coherence and Heteronuclear Multiple-Bond Correlation data of purified fraction of peak 6 (i.e., 4-methoxycinnamic acid)1

P. rotundifolia (Peak 6) HSQC (δ, ppm) HMBC (δ, ppm) CD3OD CD3OD Position 1H NMR 13C NMR 13C NMR 600 MHz 150 MHz 150 MHz

2 6.98 (8.8) 115.53 115.53, 128.38, 163.26 3 7.62 (8.7) 131.21 131.21, 146.31, 163.26 5 7.62 (8.7) 131.21 131.21, 146.31, 163.26 6 6.98 (8.8) 115.53 115.53, 128.38, 163.26 7 7.69 (16.0) 146.31 116.37, 128.38, 131.21, 169.08 8 6.50 (15.9) 116.37 128.38, 169.08

OCH3 3.84 s 55.88 163.26

1Full HMBC spectrum is displayed on Appendix 15.

6.2.4.4. Correlation Spectroscopy analysis

In addition to HSQC and HMBC, COSY was used to confirm the structure of peak 4 and 6 by observing the interactions of a proton with its neighbouring proton as presented on Table 36. Regarding peak 4, interaction between protons within the aromatic ring was evident at positions 5 (1H = 6.67) and 6 (1H = 6.57) of the aglycone in verbascoside, and positions 5′′ (1H = 6.78) and 6′′ (1H = 6.96) of the caffeic acid region. The connectivities within the rhamnose structure of verbascoside were demonstrated by the interactions between protons at position 6′′′ (1H = 1.09), 3′′′ (1H = 3.55) and 4′′′ (1H = 188

3.29). The protons interaction within the double ring was evident at positions 7′′ (1H = 1 1 7.59) and 8′′ ( H = 6.27), and also within the CH2 region at positions 8 ( H = 3.73 and 4.05).

Regarding peak 6, proton interactions were observed at positions 3-5 (1H = 7.54) and 2- 6 (1H = 6.98) in the aromatic ring of 4-methyoxycinnamic acid. Proton interactions were also observed within the double bond at positions 7 (1H = 7.63) and 8 (1H = 6.33). An illustration of COSY analysis is shown in Figure 43. The boxes drawn show correlations between protons in the structure of both verbascoside and 4- methyoxycinnamic acid. Thus, combining all the evidences obtained by LC-HRMS, LC-MS/MS and NMR, it was concluded that peak 4 represented verbascoside and peak 6 represented 4-methyoxycinnamic acid. The structures identified are displayed in Figure 44.

Table 36. COSY data of peak 4 and 6 in purified P. rotundifolia extract

1H NMR (δ, ppm) Positions (CD3OD) Peak 4 (i.e., verbascoside)

5-6 6.67 6.57 8a-8b 3.73 4.05 7′′-8′′ 7.59 6.27 5′′-6′′ 6.78 6.96 6′′′, 3′′′, 4′′′ 1.09 3.55, 3.29

Peak 6 (i.e., 4-methoxycinnamic acid)

5-6, 2-3 7.54 6.98 7-8 7.63 6.33

189

Verbascoside

4-methoxycinnamic acid

7′′ 8′′ Verbascoside 8 7 2, 6 3, 5 4-methoxycinnamic acid

8′′ 7

3, 5

2, 6

7′′ 8

Figure 43. COSY spectrum of crude extract P. rotundifolia (CD3OD) showing the protons relationship in the double bond of verbascoside (peak 4) and the benzene ring of 4-methoxycinnamic acid (peak 6).

190

Caffeic acid p-coumaric acid

Hesperidin

Verbascoside

1-O-β-ᴅ-glucopyranosyl sinapate 4-methoxycinnamic acid

Glucose ester of (E)-p-coumaric acid Naringenin

Figure 44. Chemical structure of compounds identified in purified P. rotundifolia extract.

191

6.2.5. Quantification of phenolic compounds by HPLC-PDA

The main compounds identified in P. rotundifolia extract were verbascoside (48.8%), 4- methoxycinnamic acid (36.4%), glucose ester of p-coumaric acid (9.2%) and 1-O-β-ᴅ- glucopyranosyl sinapate (5.6%), while the minor compounds identified in trace amounts were caffeic acid, p-coumaric acid, hesperidin and naringenin (Table 37). The levels of minor compounds were too low to be accurately quantified by HPLC. The overlapping of peaks has been taken into consideration in this study. The integration of the peaks has been adjusted to select only for the peak of interest and eliminating overlapping peaks.

Verbascoside, also known as acteoside, has been reported in a number of medicinal plants. Studies have found this compound to possess antioxidant (Bilia et al., 2008), antimicrobial (Pardo et al., 1993), antinociceptive (Akdemir et al., 2011) and anti- inflammatory (Akdemir et al., 2011) activities. Thus, this compound is expected to have a significant contribution to the antioxidant capacity as well as the inhibitory effect on the inflammation-related hyaluronidase of the mint bush, as reported in Chapter 4. The presence of verbascoside has been previously identified in Prostanthera melissifolia F. Muell along with a number of other phenolic compounds including apigenin, ursolic acid, martynoside, isomartynoside, isoverbascoside, betonyoside F and isobetonyoside F. (Kisiel and Piozzi, 1999). Of these phenolic compounds, however, only verbascoside was identified in the native Australian mint bush P. rotundifolia. Furthermore, 4- methoxycinnamic acid and 1-O-β-ᴅ-glucopyranosyl sinapate were identified for the first time in Prostanthera in the present study. These phenolic compounds were all present in significant amounts in the native Australian mint bush, but especially 4- methoxycinnamic acid, which was the second large phenolic component in this herb.

Apart from verbascoside, which has been shown to exert antioxidant effect (Bilia et al., 2008), the remaining major compounds, 4-methyoxycinnamic acid and 1-O-β-ᴅ- glucopyranosyl sinapate, have not been commonly found as antioxidants in herbs, although as phenolic compounds, they would be expected to have some antioxidant capacities. Nevertheless, the lack of high levels of phenolic compounds with large antioxidant capacities, such as rosmarinic acid, in the mint bush probably explains its relatively low antioxidant capacities compared with the two Mentha herbs, as reported in Chapter 4.

192

Table 37. Quantification of phenolic compounds identified in methanol-based extract of Prostanthera rotundifolia by HPLC-PDA

Retention time Concentration

1 Peak Compound (min, λmax) (mg per g purified extract )

1 Caffeic acid 5.1 (322 nm) trace

2 p-coumaric acid 6.4 (310 nm) trace

3 Hesperidin 10.4 (283 nm) trace

4 Verbascoside 15.4 (329 nm) 127.1 ± 5.7

5 1-O-β-ᴅ-glucopyranosyl sinapate 17.0 (312 nm) 14.6 ± 2.9

6 4-methoxycinnamic acid 18.6 (312 nm) 94.7 ± 1.0

7 glucose ester of p-coumaric acid 19.6 (312 nm) 24.0 ± 2.7

8 Naringenin 23.0 (288 nm) trace

1Values are expressed as mg gallic acid equivalent (GE) per g purified extract for compounds detected at 280 nm. The data represent the mean ± standard deviation of three replicates. HPLC- PDA: High Performance Liquid Chromatography-Photodiode Array Detector.

6.3. Conclusion

The identity of eight phenolic acids and flavanone glycosides in the leaves of P. rotundifolia was elucidated using a combination of chromatographic techniques and mass spectrometry. The structural features of verbascoside and 4-methoxycinnamic acid were confirmed by 1-D and 2-D NMR analyses. The main compounds identified were verbascoside, 4-methoxycinnamic acid, glucose ester of p-coumaric acid and 1-O-β-ᴅ- glucopyranosyl sinapate. Verbascoside was reported previously in Prostanthera; however the latter three phenolic compounds were identified for the first time in this genus. The minor compounds found were caffeic acid, p-coumaric acid, hesperidin and naringenin, which are commonly present in the Lamiaceae family. The lack of high levels of phenolic compounds with strong antioxidant activities, such as rosmarinic acid, in the Australian native mint bush is probably responsible for its relatively low antioxidant capacity.

193

Chapter 7 Conclusions and Recommendations

This study was carried out with the aim to expand the knowledge on Australian native edible plants by exploring the potential health promoting properties of two native Australian mints in the Lamiaceae family, Mentha australis R. Br. (river mint) and Prostanthera rotundifolia R. Br. (mint bush), which have not been studied previously. Furthermore, this study is also aimed at identifying the phenolic compounds of the two native Australian mints using cutting edge analytical techniques. The following section presents some key conclusions in achieving these goals along with suggestions for future research.

The experimental components of the project began with the determination of total phenolic and flavonoid contents in native Australian mints with the common spearmint used as a reference. The crude methanolic extracts of the river mint contained a high level of total phenolic content (76.3 ± 3.8 mg GE/g, DW) and total flavonoid content (14.7 ± 0.6 mg RE/g, DW), which were higher or comparable to those in many common herbs such as spearmint and rosemary. The crude extracts of the mint bush had a considerably lower level of total phenolics (42.1 ± 4.3 mg GE/g, DW), but higher total flavonoid content (24.8 ± 1.2 mg RE/g, DW) than the river mint.

The project then moved to the evaluation of the Australian mints for their biological activities, namely antioxidant activities and the capacity to inhibit digestive and inflammation-related enzymes (α-glucosidase, α-amylase, pancreatic lipase and hyaluronidase). The river mint and mint bush extracts possessed high antioxidant capacities as assessed by four different assays. The crude extracts of river mint exhibited high ABTS radical scavenging, DPPH radical scavenging, ferric reducing and ORAC antioxidant activities comparable to that of crude spearmint extracts. The crude mint bush extracts exhibited lower antioxidant capacities than the Mentha species. Furthermore, antioxidant capacity results obtained from the different antioxidant assays were positively correlated with total phenolic content, but the total flavonoid content was significantly correlated only with the results of ABTS and DPPH assays.

The river mint and mint bush were also found to exhibit inhibitory activities against key health relevant digestive and inflammation-related enzymes. Overall, the Australian mints were stronger inhibitors of pancreatic lipase than the other enzymes. Crude 194 extracts of river mint possessed significant inhibitory activity against pancreatic lipase

(IC50 = 0.61 mg/mL), hyaluronidase (IC50 = 1.83 ± 0.04 mg/mL) and α-glucosidase

(IC50 = 5.6 ± 0.5 mg/mL). The crude mint bush extracts also exhibited a strong pancreatic lipase-inhibitory effect (IC50 = 0.39 ± 0.01 mg/mL); however, it was found to be a poor inhibitor of the other enzymes. The river mint and mint bush were found to be more potent inhibitors of pancreatic lipase than the common spearmint (IC50 = 1.10 ± 0.01 mg/mL), but a less potent inhibitor of this enzyme compared to another common herb, rosemary (IC50 = 0.01 mg/mL). Generally, both the Australian mints were also found to have higher α-glucosidase inhibitory activities than that of spearmint, although only the river mint exhibited greater inhibitory activity against the hyaluronidase enzyme. The hyaluronidase inhibitory activity of the river mint was also higher than other herbs from the Lamiaceae family, including rosemary and thyme. None of the evaluated mints showed much inhibitory activity towards α-amylase.

Purification of the crude extracts by adsorbent material XAD-7 Amberlite® resin resulted in significant decreases in total phenolic content, total flavonoid content and antioxidant capacities for all three herbal plants examined. Similarly, the purified mint extracts exhibited lower inhibitory effects on most of the health-relevant enzymes compared to their crude counterparts. Specifically, purification resulted in a lower pancreatic lipase, hyaluronidase and α-amylase inhibitory activities. The α-amylase inhibitory activities of the purified extracts were especially low in that they could not be detected at the lower concentration range (<5 mg/mL). It was therefore concluded that the non-phenolic components in the crude extracts contributed significantly to the measured “phenolic” contents, as well as biological activities of the herbs. These non- phenolic components included lipids, reducing sugars and pigments, which were mostly removed by purification. On the other hand, the inhibitory activity of the river mints against α-glucosidase was concluded to be mostly due to phenolic compounds, which were greatly enriched by purification with resultant increases in the activity. The purified, polyphenolic-rich extracts of river mint exhibited a higher inhibition activity compared to spearmint against hyaluronidase and comparable inhibition effect against α-glucosidase, while purified mint bush extracts showed lower activities than the two Mentha species, in respect to the inhibition of the two enzymes.

195

In the next major phase of the project, the phenolic compounds in the purified extracts of river mint and mint bush were identified by the combination of a series of analytical techniques. Firstly, HPLC-PDA and GC-MS analysis provided tentative information about the chemical structures of the phenolic compounds. LC-HRMS and LC-MS/MS were then used to produce information on the molecular mass of the suspect compounds, and special chemical structure software was employed to provide an empirical formula for the compound. The combination of these two techniques, coupled with the structural information obtained by HPLC-PDA and GC-MS, enabled the identification and confirmation of most of the phenolic compounds. Finally, the application of 1H NMR, 13C NMR and 2-D NMR techniques enabled the elucidation and confirmation of the chemical structure for compounds whose concentration was high enough to permit collection by preparative HPLC.

Nine phenolic compounds were identified and quantified in purified river mint extracts (Chapter 5). The major compounds identified were neoponcirin (30.4%), rosmarinic acid (29.5%), chlorogenic acid (7.7%), narirutin (5.6%) and biochanin A (1.9%), while caffeic acid, apigenin, naringenin and hesperetin existed in trace amounts. For the purified mint bush extract, eight phenolic compounds were identified (Chapter 6). The major phenolic compounds were verbascoside (48.8%), 4-methoxycinnamic acid (36.4%), glucose ester of p-coumaric acid (9.2%) and 1-O-b-D-glucopyranosyl sinapate (5.6%), while caffeic acid, p-coumaric acid, hesperidin and naringenin were the minor components in mint bush.

The structural features of neoponcirin and rosmarinic acid in river mint and verbascoside and 4-methoxycinnamic acid in mint bush were confirmed by 1D- and 2D- NMR analyses. Neoponcirin and biochanin A were identified for the first time in the Mentha genus. Furthermore, 4-methoxycinnamic acid and 1-O-β-ᴅ-glucopyranosyl sinapate were also identified for the first time in the genus of Prostanthera. The minor compounds found in the two native Australian mint samples, caffeic acid, apigenin, hesperetin, p-coumaric acid, hesperidin and naringenin, are commonly present in the Lamiaceae family. The main constituents (rosmarinic acid, neoponcirin, verbascoside) found in the mints have been widely reported to possess antioxidant capacity. Rosmarinic acid, verbascoside and 4-methoxycinnamic acid have also been reported to exhibit to inhibitory activity against hyaluronidase and α-glucosidase. These

196 components were concluded to be the main contributors to the respective biological activities of the Australian native mints.

Several analytical conclusions are also worth noting. First, HPLC-PDA is a simple and powerful technique for identifying phenolic compounds, but only when the general phenolic composition of the sample is mostly known (e.g., when such information is available for similar matrices). When analysing a relatively new matrix without prior knowledge of the constituents, however, its usefulness is greatly diminished because it is difficult to decide what reference standards to acquire. Second, the combination of LC-HRMS and LC-MS/MS using different ionisation methods is a powerful approach to identify unknown phenolic compounds in plant matrices. The molecular and fragment mass information obtained, coupled with analysis by structural prediction software, enabled the construction of empirical formulae of the analyte, which could guide the acquisition of reference standards for comparison. Finally, NMR and especially its various 2D analysis models are extremely powerful in elucidating the structure of unknown compounds, even when reference standards are unavailable for confirmation. The combination of LC-HRMS, LC-MS/MS and NMR analyses should be possible to elucidate the structure of most phenolic compounds in plant matrices.

With regards to future research directions in this area, several recommendations can be made. Relatively high antioxidant capacity was observed in the two Australian native mints by reagent based assays. However, it is also important to understand how they would perform under conditions more closely resembling the human physiological conditions. Future work therefore could evaluate the cellular antioxidant activity of the phenolic-rich mint extracts using tissue cell lines such as the HepG2 cells. Also, antioxidant activities of phenolic compounds have been shown to be associated with a number of chemopreventive effects, including protection against cellular damage induced by hydrogen peroxide and anti-proliferative and pro-apoptotic activities against human cancer cells (Tan et al., 2011b, Sakulnarmrat et al., 2015). Research has also shown that the main phenolic compounds in mints can exhibit other significant health benefiting properties, such as rosmarinic acid exhibiting anti-inflammatory, antihepatitis and antitumour activities, and verbascoside exhibiting anti-inflammatory, antimicrobial and antinociceptive effects. Considering that these phenolic compounds are present in high levels in the two native Australian mints, the potential health

197

beneficial effects of consuming these herbs warrant investigation, possibly using animal model or human volunteer studies.

Phenolic-rich extracts of the two native Australian mints showed significant α- glucosidase activities, demonstrating that the mints could potentially play a role in the management of diabetes. However, because the results are obtained by in vitro enzymatic methods, the conclusion can only be regarded as preliminary. Animal model and human clinical studies are needed to verify these results before health benefit claims can be made and this could be another major direction for future study.

Finally, Australia is home to a wide range of unique native flora, which has traditionally served as a source of food and medicine for the indigenous people. Modern scientific studies have shown that these plants possess numerous health promoting properties attributable to their diverse range of bioactive compounds. However, systematic scientific investigations in this area are still in the early stage. Significantly greater research effort is needed in order to fully unravel and rediscover the health beneficial properties of this bountiful source of food and medicinal plants and phytochemicals. Furthermore, collaboration with the indigenous communities may provide further insights on the traditional knowledge of Australian native plants.

198

REFERENCES

AACC (2000). Approved Methods of the American Association of Cereal Chemists, AACC. Adhvaryu, A., Erhan, S. Z., Liu, Z. S. & Perez, J. M. (2000). Oxidation kinetic studies of oils derived from unmodified and genetically modified vegetables using pressurized differential scanning calorimetry and nuclear magnetic resonance spectroscopy. Thermochimica Acta, 364, 87-97. Adisakwattana, S., Sookkongwaree, K., Roengsumran, S., Petsom, A., Ngamrojnavanich, N., Chavasiri, W., Deesamer, S. & Yibchok-Anun, S. (2004). Structure–activity relationships of trans-cinnamic acid derivatives on α- glucosidase inhibition. Bioorganic & Medicinal Chemistry Letters, 14, 2893- 2896. Aehle, E., Raynaud-Le Grandic, S., Ralainirina, R., Baltora-Rosset, S., Mesnard, F., Prouillet, C., Mazière, J. C. & Fliniaux, M. A. (2004). Development and evaluation of an enriched natural antioxidant preparation obtained from aqueous spinach (Spinacia oleracea) extracts by an adsorption procedure. Food Chemistry, 86, 579-585. Agboola, S., Ee, K. Y. & Huhn, A. (2012). Functional properties of processed Australian wattle (Acacia victoriae Bentham) seed extracts. Food Chemistry, 133, 990-997. Agboola, S., Ee, K. Y., Mallon, L. & Zhao, J. (2007). Isolation, characterization, and emulsifying properties of wattle seed (Acacia victoriae Bentham) extracts. Journal of Agricultural and Food Chemistry, 55, 5858-5863. Agboola, S. O. & Radovanovic-Tesic, M. (2002). Influence of Australian native herbs on the maturation of vacuum-packed cheese. LWT-Food Science and Technology, 35, 575-583. Ahmed, A. K. & Johnson, K. A. (2000). Horticultural development of Australian native edible plants. Australian Journal of Botany, 48, 417-426. Akdemir, Z., Kahraman, C., Tatlı, I., Akkol, E., Süntar, I. & Keles, H. (2011). Bioassay- guided isolation of anti-inflammatory, antinociceptive and wound healer glycosides from the flowers of Verbascum mucronatum Lam. Journal of Ethnopharmacology, 136, 436-443. ALA (2015). Atlas of Living Australia [Online]. Australia. Available: http://www.ala.org.au/ [Accessed 05/04/2012]. Andon, N., Hollingworth, S., Koller, A., Greenland, A., Yates, J. & Haynes, P. (2002). Proteomic characterization of wheat amyloplasts using identification of proteins by tandem mass spectrometry. Proteomics, 2, 1156-1168. ANPS (2015). Australian Native Plants Society (Australia) [Online]. Available: http://anpsa.org.au/ [Accessed 5/06/2012]. Apostolidis, E., Kwon, Y. & Shetty, K. (2006). Potential of cranberry-based herbal synergies for diabetes and hypertension management. Asia Pacific Journal of Clinical Nutrition, 15, 433-441. Arumugam, P., Priya, N. G., Subathra, M. & Ramesh, A. (2008). Anti-inflammatory activity of four solvent fractions of ethanol extract of Mentha spicata L. investigated on acute and chronic inflammation induced rats. Environmental Toxicology and Pharmacology, 26, 92-95.

199

Arumugam, P., Ramamurthy, P. & Ramesh, A. (2010). Antioxidant and cytotoxic activities of lipophilic and hydrophilic fractions of Mentha spicata L. (Lamiaceae). International Journal of Food Properties, 13, 23-31. Asgar, A. (2013). Anti-diabetic potential of phenolic compounds: A review. International Journal of Food Properties, 16, 91-103. Azqueta, A. & Collins, A. R. (2012). Carotenoids and DNA damage. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis, 733, 4-13. Bahadoran, Z., Mirmiran, P. & Azizi, F. (2013). Dietary polyphenols as potential nutraceuticals in management of diabetes: a review. Journal of Diabetes and Metabolic Disorders, 12, 1. Bajpai, V. K., Rahman, A., Choi, U. K., Youn, S. J. & Kang, S. C. (2007). Inhibitory parameters of the essential oil and various extracts of Metasequoia glyptostroboides Miki ex Hu to reduce food spoilage and food-borne pathogens. Food Chemistry, 105, 1061-1066. 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, 191-203. Bannerman, C. (2006). Indigenous food and cookery books: Redefining aboriginal cuisine. Journal of Australian Studies, 30, 19-36. Barr, A. (1993). Traditional Aboriginal medicines in the Northern Territory of Australia, Darwin, Conservation Commission of the Northern Territory of Australia. Barr, A., Knight, T., Andrews, M. & Alexander, V. (1988). Traditional bush medicines: an Aboriginal Pharmacopoeia, Richmond, Greenhouse Publications. Bellik, Y., Boukraâ, L., Alzahrani, H. A., Bakhotmah, B. A., Abdellah, F., Hammoudi, S. M. & Iguer-Ouada, M. (2012). Molecular mechanism underlying anti- inflammatory and anti-allergic activities of phytochemicals: an update. Molecules, 18, 322-353. Berger, S. & Braun, S. (2004). 200 and More NMR Experiments: A Practical Course, Weinheim, Wiley and Sons. Bergman, M., Varshavsky, L., Gottlieb, H. E. & Grossman, S. (2001). The antioxidant activity of aqueous spinach extract: chemical identification of active fractions. Phytochemistry, 58, 143-152. Bilia, A. R., Giomi, M., Innocenti, M., Gallori, S. & Vincieri, F. F. (2008). HPLC– DAD–ESI–MS analysis of the constituents of aqueous preparations of verbena and lemon verbena and evaluation of the antioxidant activity. Journal of Pharmaceutical and Biomedical Analysis, 46, 463-470. Birari, R. B. & Bhutani, K. K. (2007). Pancreatic lipase inhibitors from natural sources: unexplored potential. Drug Discovery Today, 12, 879-889. Blewitt, M. & Southwell, I. A. (2000). Backhousia anisata Vickery, an alternative source of (E)-anethole. Journal of Essential Oil Research, 12, 445-454. Blokhina, O., Virolainen, E. & Fagerstedt, K. V. (2003). Antioxidants, oxidative damage and oxygen deprivation stress: a review. Annals of Botany, 91, 179-194. Boschin, G., Scigliuolo, G. M., Resta, D. & Arnoldi, A. (2014). ACE-inhibitory activity of enzymatic protein hydrolysates from lupin and other legumes. Food Chemistry, 145, 34-40. Bowen-Forbes, C. S., Zhang, Y. & Nair, M. G. (2010). Anthocyanin content, antioxidant, anti-inflammatory and anticancer properties of blackberry and raspberry fruits. Journal of Food Composition and Analysis, 23, 554-560.

200

Brand-Miller, J. C. & Holt, S. H. A. (1998). Australian Aboriginal plant foods: a consideration of their nutritional composition and health implications. Nutrition Research Reviews, 11, 5-23. Brand, J. C., Cherikoff, V., Lee, A. & Truswell, A. S. (1982). An outstanding food source of vitamin C. The Lancet, 320, 873. Britton, G., Liaaen-Jensen, S. & Pfander, H. (2009). Carotenoids Volume 5: nutrition and health, Basel, Birkhäuser Basel. Brophy, J. J., Goldsack, R. J., Craven, L. A. & Ford, A. J. (2007). Leaf oil of Backhousia enata (Myrtaceae). Journal of Essential Oil Research, 19, 26-27. Brophy, J. J., Goldsack, R. J., Fookes, C. J. R. & Forster, P. I. (1995). Leaf oils of the genus Backhousia (Myrtaceae). Journal of Essential Oil Research, 7, 237-254. Brophy, J. J., Goldsack, R. J. & Forster, P. I. (2004). Leaf essential oils of the Australian species of Acronychia (Rutaceae). Journal of Essential Oil Research, 16, 597-607. Burt, S. (2004). Essential oils: their antibacterial properties and potential applications in foods—a review. International Journal of Food Microbiology, 94, 223-253. Bustanji, Y., Issa, A., Mohammad, M., Hudaib, M., Tawah, K., Alkhatib, H., Almasri, I. & Al-Khalidi, B. (2010). Inhibition of hormone sensitive lipase and pancreatic lipase by Rosmarinus officinalis extract and selected phenolic constituents. Journal of Medicinal Plants Research, 4, 2235-42. Butsat, S. & Siriamornpun, S. (2016). Effect of solvent types and extraction times on phenolic and flavonoid contents and antioxidant activity in leaf extracts of Amomum chinense C. International Food Research Journal, 23. Calo, J. R., Crandall, P. G., O'bryan, C. A. & Ricke, S. C. (2015). Essential oils as antimicrobials in food systems – A review. Food Control, 54, 111-119. Capecka, E., Mareczek, A. & Leja, M. (2005). Antioxidant activity of fresh and dry herbs of some Lamiaceae species. Food Chemistry, 93, 223-226. Cappiello, A., Famiglini, G., Palma, P., Pierini, E., Termopoli, V. & Trufelli, H. (2008). Overcoming matrix effects in liquid chromatography − mass spectrometry. Analytical Chemistry, 80, 9343-9348. Cazzola, R., Camerotto, C. & Cestaro, B. (2011). Anti-oxidant, anti-glycant, and inhibitory activity against α-amylase and α-glucosidase of selected spices and culinary herbs. International Journal of Food Sciences and Nutrition, 62, 175- 184. Chacón Morales, P., Amaro, J. M., Bahsas, B. & Alí, H. (2013). Isolation and characterization of (+)-mellein, the first isocoumarin reported in Stevia genus. Avances en Química, 8, 145-151. Cherikoff, V. & Isaacs, J. (1989). Bush food handbook: how to gather, grow, process & cook Australian wild foods, South Yarra, Ti Tree Press. Clark, C. J., Hockings, P. D., Joyce, D. C. & Mazucco, R. A. (1997). Application of magnetic resonance imaging to pre-and post-harvest studies of fruits and vegetables. Postharvest Biology and Technology, 11, 1-21. Cleary, J., Grey-Gardner, R. & Josif, P. (2009). Hands across the Desert: Linking desert Aboriginal Australians to each other and to the bush foods industry, Alice Springs, Report to Department of Agriculture, Fisheries and Forestry by Desert Knowledge Cooperative Research Centre. . Cock, I. E. (2008). Antibacterial activity of selected Australian native plant extracts. Internet Journal of Microbiology, 4, 1-8. Cock, I. E. 2011. Medicinal and Aromatic Plants-Australia. Encyclopedia of Life Support Systems (EOLSS). 201

Collins, T. L., Jones, G. L. & Sadgrove, N. J. (2014). Volatiles from the rare australian desert plant Prostanthera centralis BJ Conn (Lamiaceae): Chemical composition and antimicrobial activity. Agriculture, 4, 308-316. Corradini, E., Foglia, P., Giansanti, P., Gubbiotti, R., Samperi, R. & Laganà, A. (2011). Flavonoids: chemical properties and analytical methodologies of identification and quantitation in foods and plants. Natural Product Research, 25, 469-495. Cushnie, T. T. & Lamb, A. J. (2005). Antimicrobial activity of flavonoids. International Journal of Antimicrobial Agents, 26, 343-356. Cuyckens, F. & Claeys, M. (2004). Mass spectrometry in the structural analysis of flavonoids. Journal of Mass Spectrometry, 39, 1-15. Dai, D., He, J., Sun, R., Zhang, R., Aisa, H. A. & Abliz, Z. (2009). Nuclear magnetic resonance and liquid chromatography–mass spectrometry combined with an incompleted separation strategy for identifying the natural products in crude extract. Analytica Chimica Acta, 632, 221-228. Dai, J. & Mumper, R. J. (2010). Plant phenolics: extraction, analysis and their antioxidant and anticancer properties. Molecules, 15, 7313-7352. Dandona, P., Aljada, A., Chaudhuri, A., Mohanty, P. & Garg, R. (2005). Metabolic syndrome a comprehensive perspective based on interactions between obesity, diabetes, and inflammation. Circulation, 111, 1448-1454. Danh, L. T., Triet, N. D. A., Zhao, J., Mammucari, R. & Foster, N. (2012). Antioxidant activity, yield and chemical composition of lavender essential oil extracted by supercritical CO2. The Journal of Supercritical Fluids, 70, 27-34. De La Rosa, L. A., Alvarez-Parrilla, E. & Gonzalez-Aguilar, G. A. (2010). Fruit and Vegetable Phytochemicals: Chemistry, Nutritional Value and Stability, Iowa, Wiley-Blackwell. De Rijke, E., Out, P., Niessen, W. M. A., Ariese, F., Gooijer, C. & Udo, A. T. (2006). Analytical separation and detection methods for flavonoids. Journal of Chromatography A, 1112, 31-63. Dellar, J., Cole, M., Gray, A., Gibbons, S. & Waterman, P. (1994). Antimicrobial sesquiterpenes from Prostanthera aff. melissifolia and P. rotundifolia. Phytochemistry, 36, 957-960. Delort, E. & Jaquier, A. (2009). Novel terpenyl esters from Australian finger lime (Citrus australasica) peel extract. Flavour and Fragrance Journal, 24, 123-132. Deman, J. M. (2013). Principles of Food Chemistry, New York, Springer. Dimitrios, B. (2006). Sources of natural phenolic antioxidants. Trends in Food Science & Technology, 17, 505-512. Dorman, D., Kosar, M., Kahlos, K., Holm, Y. & Hiltunen, R. (2003). Antioxidant properties and composition of aqueous extracts from Mentha species, hybrids, varieties, and cultivars. Journal of Agricultural and Food Chemistry, 51, 4563- 4569. Dorman, H. J. D. & Deans, S. G. (2000). Antimicrobial agents from plants: antibacterial activity of plant volatile oils. Journal of Applied Microbiology, 88, 308-316. Dupont, S., Caffin, N., Bhandari, B. & Dykes, G. A. (2006). In vitro antibacterial activity of Australian native herb extracts against food-related bacteria. Food Control, 17, 929-932. Ee, K. Y., Agboola, S., Rehman, A. & Zhao, J. (2011). Characterisation of phenolic components present in raw and roasted wattle (Acacia victoriae Bentham) seeds. Food Chemistry, 129, 816-821.

202

Ee, K. Y., Agboola, S., Rehman, A. & Zhao, J. (2012). In vitro antioxidant and bioactive properties of raw and roasted wattle (Acacia victoriae Bentham) seed extracts. International Journal of Food Science & Technology, 47, 2000-2008. Ee, K. Y., Rehman, A., Agboola, S. & Zhao, J. (2009a). Influence of heat processing on functional properties of Australian wattle seed (Acacia victoriae Bentham) extracts. Food Hydrocolloids, 23, 116-124. Ee, K. Y. & Yates, P. (2013). Nutritional and antinutritional evaluation of raw and processed Australian wattle (Acacia saligna) seeds. Food Chemistry, 138, 762- 769. Ee, K. Y., Zhao, J., Rehman, A. & Agboola, S. (2008). Characterisation of trypsin and α-chymotrypsin inhibitors in Australian wattle seed (Acacia victoriae Bentham). Food Chemistry, 107, 337-343. Ee, K. Y., Zhao, J., Rehman, A. & Agboola, S. (2009b). Purification and characterization of a Kunitz-type trypsin inhibitor from Acacia victoriae Bentham seeds. Journal of Agricultural and Food Chemistry, 57, 7022-7029. El-Beshbishy, H. A. & Bahashwan, S. A. (2012). Hypoglycemic effect of basil (Ocimum basilicum) aqueous extract is mediated through inhibition of α- glucosidase and α-amylase activities: an in vitro study. Toxicology and Industrial Health, 28, 42-50. Fiamegos, Y., Nanos, C., Vervoort, J. & Stalikas, C. (2004). Analytical procedure for the in-vial derivatization—extraction of phenolic acids and flavonoids in methanolic and aqueous plant extracts followed by gas chromatography with mass-selective detection. Journal of Chromatography A, 1041, 11-18. Field, L. D., Sternhell, S. & Kalman, J. R. (2012). Organic Structures from Spectra, Chichester, West Sussex, John Wiley & Sons. Figueroa-PéRez, M. G., Rocha-GuzmáN, N. E., Pérez-Ramírez, I., Mercado-Silva, E. & Reynoso-Camacho, R. (2014). Metabolite Profile, Antioxidant Capacity, and Inhibition of Digestive Enzymes in Infusions of Peppermint (Mentha piperita) Grown under Drought Stress. Journal of Agricultural and Food Chemistry, 62, 12027-12033. Forbes-Smith, M. & Paton, J. E. (2002). Innovative products from Australian native foods, Barton, Report by Rural Industries Research and development Corporation (RIRDC) Fulton, A. (2000). Food safety of three species of native mint, Barton, Rural Industries Research & Development Corporation (RIRDC). Garcia-Salas, P., Morales-Soto, A., Segura-Carretero, A. & Fernández-Gutiérrez, A. (2010). Phenolic-compound-extraction systems for fruit and vegetable samples. Molecules, 15, 8813-8826. Garza, A. L., Milagro, F., Boque, N., Campión, J. & Martinez, J. A. (2011). Natural inhibitors of pancreatic lipase as new players in obesity treatment. Planta Medica, 77, 773-785. Giardi, M. T., Rea, G. & Berra, B. (2011). Bio-Farms for Nutraceuticals: Functional Food and Safety Control by Biosensors, New York, Springer Gilles, M., Zhao, J., An, M. & Agboola, S. (2010). Chemical composition and antimicrobial properties of essential oils of three Australian Eucalyptus species. Food Chemistry, 119, 731-737. Gómez-Aguirre, Y. A., Zamilpa, A., González-Cortazar, M. & Trejo-Tapia, G. (2012). Adventitious root cultures of Castilleja tenuiflora Benth. as a source of phenylethanoid glycosides. Industrial Crops and Products, 36, 188-195.

203

Gott, B. (2008). Indigenous use of plants in south-eastern Australia. Telopea, 12, 215- 226. Gropper, S. S. & Smith, J. L. (2012). Advanced Nutrition and Human Metabolism, Belmont, CA, Cengage Learning. Guedon, D. J. & Pasquier, B. P. (1994). Analysis and distribution of flavonoid glycosides and rosmarinic acid in 40 Mentha x piperita clones. Journal of Agricultural and Food Chemistry, 42, 679-684. Guillén, M. a. D. & Ruiz, A. (2001). High resolution 1 H nuclear magnetic resonance in the study of edible oils and fats. Trends in Food Science & Technology, 12, 328- 338. Gulluce, M., Orhan, F., Yanmis, D., Arasoglu, T., Guvenalp, Z. & Demirezer, L. O. (2015). Isolation of a flavonoid, apigenin 7-O-glucoside, from Mentha longifolia (L.) Hudson subspecies longifolia and its genotoxic potency. Toxicology and Industrial Health, 31, 831-840. Gulluce, M., Sahin, F., Sokmen, M., Ozer, H., Daferera, D., Sokmen, A., Polissiou, M., Adiguzel, A. & Ozkan, H. (2007). Antimicrobial and antioxidant properties of the essential oils and methanol extract from Mentha longifolia L. ssp. longifolia. Food Chemistry, 103, 1449-1456. Guo, Y., Sakulnarmrat, K. & Konczak, I. (2014). Anti-inflammatory potential of native Australian herbs polyphenols. Toxicology Reports, 1, 385-390. Gutierrez, J., Rodriguez, G., Barry-Ryan, C. & Bourke, P. (2008). Efficacy of plant essential oils against foodborne pathogens and spoilage bacteria associated with ready-to-eat vegetables: antimicrobial and sensory screening. Journal of Food Protection, 71, 1846-1854. Halket, J. M., Waterman, D., Przyborowska, A. M., Patel, R. K. P., Fraser, P. D. & Bramley, P. M. (2005). Chemical derivatization and mass spectral libraries in metabolic profiling by GC/MS and LC/MS/MS. Journal of Experimental Botany, 56, 219-243. Hansen, S., Pedersen-Bjergaard, S. & Rasmussen, K. (2011). Introduction to Pharmaceutical Chemical Analysis, Chichester, West Sussex, John Wiley & Sons. Hegarty, M. P., Hegarty, E. E. & Wills, R. B. H. (2001). Food safety of Australian plant bushfoods, Barton, Rural Industries Research Development Corporation (RIRDC). Hodgson, J. M. & Wahlqvist, M. L. (1993). Nutrition and health of Victorian Aborigines (Kooris). Asia Pacific Journal of Clinical Nutrition, 2, 43-57. Huang, W. Y., Cai, Y. Z. & Zhang, Y. (2009). Natural phenolic compounds from medicinal herbs and dietary plants: potential use for cancer prevention. Nutrition and Cancer, 62, 1-20. Hung, J. Y., Hsu, Y. L., Ko, Y. C., Tsai, Y. M., Yang, C. J., Huang, M. S. & Kuo, P. L. (2010). Didymin, a dietary flavonoid glycoside from citrus fruits, induces Fas- mediated apoptotic pathway in human non-small-cell lung cancer cells in vitro and in vivo. Lung Cancer, 68, 366-374. Hurst, W. J. (2008). Methods of Analysis for Functional Foods and Nutraceuticals, Second Edition, CRC Press. Hussain, A. I., Anwar, F., Shahid, M., Ashraf, M. & Przybylski, R. (2010). Chemical composition, and antioxidant and antimicrobial activities of essential oil of spearmint (Mentha spicata L.) from Pakistan. Journal of Essential Oil Research, 22, 78-84.

204

Hvattum, E. & Ekeberg, D. (2003). Study of the collision‐induced radical cleavage of flavonoid glycosides using negative electrospray ionization tandem quadrupole mass spectrometry. Journal of Mass Spectrometry, 38, 43-49. Ignat, I., Volf, I. & Popa, V. I. (2011). A critical review of methods for characterisation of polyphenolic compounds in fruits and vegetables. Food Chemistry, 126, 1821-1835. Ikarashi, N., Takeda, R., Ito, K., Ochiai, W. & Sugiyama, K. (2011). The inhibition of lipase and glucosidase activities by acacia polyphenol. Evidence-Based Complementary and Alternative Medicine, 2011, 1-8. Ikigai, H., Nakae, T., Hara, Y. & Shimamura, T. (1993). Bactericidal catechins damage the lipid bilayer. Biochimica et Biophysica Acta (BBA)-Biomembranes, 1147, 132-136. Inoue, T., Sugimoto, Y., Masuda, H. & Kamei, C. (2002). Antiallergic effect of flavonoid glycosides obtained from Mentha piperita L. Biological and Pharmaceutical Bulletin, 25, 256-259. Ippoushi, K., Yamaguchi, Y., Itou, H., Azuma, K. & Higashio, H. (2000). Evaluation of Inhibitory Effects of Vegetables and Herbs on Hyaluronidase and Identification of Rosmarinic Acid as a Hyaluronidase Inhibitor in Lemon Balm (Melissa officinalis L.). Food Science and Technology Research, 6, 74-77. Ito, H., Miyazaki, T., Ono, M. & Sakurai, H. (1998). Antiallergic activities of rabdosiin and its related compounds: chemical and biochemical evaluations. Bioorganic & Medicinal Chemistry, 6, 1051-1056. Jacob, J. K., Tiwari, K., Correa-Betanzo, J., Misran, A., Chandrasekaran, R. & Paliyath, G. (2012). Biochemical basis for functional ingredient design from fruits. Annual Review of Food Science and Technology, 3, 79-104. Javanmardi, J., Stushnoff, C., Locke, E. & Vivanco, J. M. (2003). Antioxidant activity and total phenolic content of Iranian Ocimum accessions. Food Chemistry, 83, 547-550. Johnson, I. T. & Williamson, G. (2003). Phytochemical Functional Foods, Cambridge, Woodhead Publishing. Joubert, E., Manley, M. & Botha, M. (2008). Evaluation of spectrophotometric methods for screening of green rooibos (Aspalathus linearis) and green honeybush (Cyclopia genistoides) extracts for high levels of Bio‐active compounds. Phytochemical Analysis, 19, 169-178. Justesen, U. (2000). Negative atmospheric pressure chemical ionisation low-energy collision activation mass spectrometry for the characterisation of flavonoids in extracts of fresh herbs. Journal of Chromatography A, 902, 369-379. Kanatt, S. R., Chander, R. & Sharma, A. (2007). Antioxidant potential of mint (Mentha spicata L.) in radiation-processed lamb meat. Food Chemistry, 100, 451-458. Kaufmann, A., Butcher, P., Maden, K., Walker, S. & Widmer, M. (2010). Comprehensive comparison of liquid chromatography selectivity as provided by two types of liquid chromatography detectors (high resolution mass spectrometry and tandem mass spectrometry):“Where is the crossover point?”. Analytica Chimica Acta, 673, 60-72. Khoddami, A., Wilkes, M. A. & Roberts, T. H. (2013). Techniques for analysis of plant phenolic compounds. Molecules, 18, 2328-2375. Kim, C. Y., Lee, H. J., Lee, M. K., Ahn, M. J. & Kim, J. (2007). One step purification of flavanone glycosides from Poncirus trifoliata by centrifugal partition chromatography. Journal of Separation Science, 30, 2693-2697.

205

Kırca, A. & Arslan, E. (2008). Antioxidant capacity and total phenolic content of selected plants from Turkey. International Journal of Food Science & Technology, 43, 2038-2046. Kisiel, W. & Piozzi, F. (1999). Phenylethanoid glycosides from Prostanthera melissifolia. Phytochemistry, 51, 1083-1085. Konczak, I. (2009). Health Benefits of Australian Native Foods: An Evaluation of Health-enhancing Compounds, Barton, Rural Industries Research and Development Corporation (RIRDC). Konczak, I. & Roulle, P. (2011). Nutritional properties of commercially grown native Australian fruits: Lipophilic antioxidants and minerals. Food Research International, 44, 2339-2344. Konczak, I., Zabaras, D., Dunstan, M. & Aguas, P. (2010a). Antioxidant capacity and hydrophilic phytochemicals in commercially grown native Australian fruits. Food Chemistry, 123, 1048-1054. Konczak, I., Zabaras, D., Dunstan, M. & Aguas, P. (2010b). Antioxidant capacity and phenolic compounds in commercially grown native Australian herbs and spices. Food Chemistry, 122, 260-266. Kosar, M., Dorman, H. D., Can Baser, K. H. & Hiltunen, R. (2004). Screening of free radical scavenging compounds in water extracts of Mentha samples using a postcolumn derivatization method. Journal of Agricultural and Food Chemistry, 52, 5004-5010. Kratchanova, M., Denev, P., Ciz, M., Lojek, A. & Mihailov, A. (2010). Evaluation of antioxidant activity of medicinal plants containing polyphenol compounds. Comparison of two extraction systems. Acta Biochimica Polonica, 57, 229-234. Krauss, M., Singer, H. & Hollender, J. (2010). LC–high resolution MS in environmental analysis: from target screening to the identification of unknowns. Analytical and Bioanalytical Chemistry, 397, 943-951. Kuo, Y. H., Lee, S. M. & Lai, J. S. (2000). Constituents of the whole herb of Clinoponium laxiflorum. Journal of the Chinese Chemical Society, 47, 241-246. Kwon, Y.-I. I., Vattem, D. A. & Shetty, K. (2006). Evaluation of clonal herbs of Lamiaceae species for management of diabetes and hypertension. Asia Pacific Journal of Clinical Nutrition, 15, 107-118. Lacaille-Dubois, M., Franck, U. & Wagner, H. (2001). Search for potential angiotensin converting enzyme (ACE)-inhibitors from plants. Phytomedicine, 8, 47-52. Lassak, E. V. & Mccarthy, T. (2001). Australian Medicinal Plants, Sydney, Reed New Holland. Lawrence, B. M. (2007). Mint: The Genus Mentha, Boca Raton, FL, Taylor & Francis Group, CRC Press. Leco (2006). TruSpec CN Carbon/Nitrogen Determinator: Instruction Manual, St. Joseph, Michigan, LECO Corporation. Lee, S. Y., Mediani, A., Ah, N. A., Abu Bakar Sajak, A. & Abas, F. (2014). Antioxidant and α-glucosidase inhibitory activities of the leaf and stem of selected traditional medicinal plants. International Food Research Journal, 21, 165-172. Li, Y. Q., Zhou, F. C., Gao, F., Bian, J. S. & Shan, F. (2009). Comparative evaluation of quercetin, isoquercetin and rutin as inhibitors of α-glucosidase. Journal of Agricultural and Food Chemistry, 57, 11463-11468. Liu, R. H. (2003). Health benefits of fruit and vegetables are from additive and synergistic combinations of phytochemicals. The American Journal of Clinical Nutrition, 78, 517S-520S.

206

Lu, Y. & Foo, L. Y. (1999). Rosmarinic acid derivatives from Salvia officinalis. Phytochemistry, 51, 91-94. Macdonald‐Wicks, L. K., Wood, L. G. & Garg, M. L. (2006). Methodology for the determination of biological antioxidant capacity in vitro: a review. Journal of the Science of Food and Agriculture, 86, 2046-2056. Madhavi, D. L., Deshpande, S. S. & Salunkhe, D. K. (1996). Food Antioxidants: Technological, Toxicological, and Health Perspectives, New York, Marcel Dekker. Maen, A. & Cock, I. (2015). Inhibitory activity of high antioxidant Australian native fruits against the bacterial triggers of selected autoimmune diseases. Pharmacognosy Communications, 5, 52-62. Mares-Perlman, J. A., Millen, A. E., Ficek, T. L. & Hankinson, S. E. (2002). The body of evidence to support a protective role for lutein and zeaxanthin in delaying chronic disease. Overview. The Journal of Nutrition, 132, 518S-524S. Marrelli, M., Loizzo, M. R., Nicoletti, M., Menichini, F. & Conforti, F. (2014). In vitro investigation of the potential health benefits of wild Mediterranean dietary plants as anti‐obesity agents with α‐amylase and pancreatic lipase inhibitory activities. Journal of the Science of Food and Agriculture, 94, 2217-2224. Marschner, P. (2012). Marschner's Mineral Nutrition of Higher Plants, London, Academic Press. Mcdonald, J., Caffin, N., Sommano, S. & Cocksedge, R. (2006). The effect of post harvest handling on selected native food plants, Barton, Rural Industries Research and Development Corporation (RIRDC). Mcdougall, G. J., Shpiro, F., Dobson, P., Smith, P., Blake, A. & Stewart, D. (2005). Different polyphenolic components of soft fruits inhibit α-amylase and α- glucosidase. Journal of Agricultural and Food Chemistry, 53, 2760-2766. Mckay, D. L. & Blumberg, J. B. (2006). A review of the bioactivity and potential health benefits of peppermint tea (Mentha piperita L.). Phytotherapy Research, 20, 619-633. Mcmaster, M. C. (2005). LC/MS: A Practical User's Guide, Wiley. Mehrabani, M., Shams-Ardakani, M., Ghannadi, A., Ghassemi-Dehkordi, N. & Sajjadi- Jazi, S. (2010). Production of rosmarinic acid in Echium amoenum fisch. and CA mey. cell cultures. Iranian Journal of Pharmaceutical Research, 111-115. Merken, H. M. & Beecher, G. R. (2000). Measurement of food flavonoids by high- performance liquid chromatography: a review. Journal of Agricultural and Food Chemistry, 48, 577-599. Miliauskas, G., Venskutonis, P. R. & Van Beek, T. A. (2004). Screening of radical scavenging activity of some medicinal and aromatic plant extracts. Food Chemistry, 85, 231-237. Mimica-Dukic, N. & Bozin, B. (2008). Mentha L. species (Lamiaceae) as promising sources of bioactive secondary metabolites. Current Pharmaceutical Design, 14, 3141-3150. Mkaddem, M., Bouajila, J., Ennajar, M., Lebrihi, A., Mathieu, F. & Romdhane, M. (2009). Chemical composition and antimicrobial and antioxidant activities of Mentha (longifolia L. and viridis) essential oils. Journal of Food Science, 74, M358-M363. Moon, S. H., Kim, K. T., Lee, N. K., Han, Y. S., Nah, S. Y., Cho, S. G., Park, Y. S. & Paik, H. D. (2009). Inhibitory effects of naringenin and its novel derivatives on hyaluronidase. Food Science and Biotechnology, 18, 267-270.

207

Mori, A., Nishino, C., Enoki, N. & Tawata, S. (1987). Antibacterial activity and mode of action of plant flavonoids against Proteus vulgaris and Staphylococcus aureus. Phytochemistry, 26, 2231-2234. Morteza-Semnani, K., Saeedi, M. & Akbarzadeh, M. (2006). The Essential Oil Composition of Mentha aquatica L. Journal of Essential Oil Bearing Plants, 9, 283-286. Muckenschnabel, I., Bernhardt, G., Spruss, T. & Buschauer, A. (1998). Pharmacokinetics and tissue distribution of bovine testicular hyaluronidase and vinblastine in mice: an attempt to optimize the mode of adjuvant hyaluronidase administration in cancer chemotherapy. Cancer Letters, 131, 71-84. Mustafa, R. A., Hamid, A. A., Mohamed, S. & Bakar, F. A. (2010). Total phenolic compounds, flavonoids, and radical scavenging activity of 21 selected tropical plants. Journal of Food Science, 75, C28-C35. Mutanen, M. & Pajari, A. M. (2010). Vegetables, Whole Grains, and Their Derivatives in Cancer Prevention, New York, Springer. Nampoothiri, S. V., Prathapan, A., Cherian, O. L., Raghu, K. G., Venugopalan, V. & Sundaresan, A. (2011). In vitro antioxidant and inhibitory potential of Terminalia bellerica and Emblica officinalis fruits against LDL oxidation and key enzymes linked to type 2 diabetes. Food and Chemical Toxicology, 49, 125- 131. Naushad, M. & Khan, M. R. (2014). Ultra Performance Liquid Chromatography Mass Spectrometry: Evaluation and Applications in Food Analysis, Boca Raton, FL, Taylor & Francis Group, CRC Press. Netzel, M., Netzel, G., Tian, Q., Schwartz, S. & Konczak, I. (2006). Sources of antioxidant activity in Australian native fruits. Identification and quantification of anthocyanins. Journal of Agricultural and Food Chemistry, 54, 9820-9826. Netzel, M., Netzel, G., Tian, Q., Schwartz, S. & Konczak, I. (2007). Native Australian fruits-a novel source of antioxidants for food. Innovative Food Science and Emerging Technologies, 8, 339-346. Nollet, L. M. L. (2004). Handbook of Food Analysis: Physical characterization and nutrient analysis, CRC PressI Llc. Noumi, E., Snoussi, M., Hajlaoui, H., Trabelsi, N., Ksouri, R., Valentin, E. & Bakhrouf, A. (2011). Chemical composition, antioxidant and antifungal potential of Melaleuca alternifolia (Tea Tree) and Eucalyptus globulus essential oils against oral Candida species. Journal of Medicinal Plants Research, 5, 4147-56. Nunes, S., Madureira, R., Campos, D., Sarmento, B., Gomes, A. M., Pintado, M. & Reis, F. (2015). Therapeutic and Nutraceutical Potential of Rosmarinic Acid- Cytoprotective Properties and Pharmacokinetic Profile. Critical Reviews in Food Science and Nutrition. Nuutila, A., Kammiovirta, K. & Oksman-Caldentey, K.-M. (2002). Comparison of methods for the hydrolysis of flavonoids and phenolic acids from onion and spinach for HPLC analysis. Food Chemistry, 76, 519-525. Orata, F. 2012. Derivatization reactions and reagents for gas chromatography analysis. In: MOHD, M. A. (ed.) Advanced Gas Chromatography-Progress in Agricultural, Biomedical and Industrial Applications. Rijeka: InTech. Özben, T. (2013). Free Radicals, Oxidative Stress, and Antioxidants: Pathological and Physiological Significance, New York, Springer. Padmini, E., Prema, K., Vijaya Geetha, B. & Usha Rani, M. (2008). Comparative study on composition and antioxidant properties of mint and black tea extract. International Journal of Food Science & Technology, 43, 1887-1895. 208

Palá-Paúl, J., Copeland, L. M., Brophy, J. J. & Goldsack, R. J. (2006). Essential oil composition of two variants of Prostanthera lasianthos Labill. from Australia. Biochemical Systematics and Ecology, 34, 48-55. Pardo, F., Perich, F., Villarroel, L. & Torres, R. (1993). Isolation of verbascoside, an antimicrobial constituent of Buddleja globosa leaves. Journal of Ethnopharmacology, 39, 221-222. Park, J. B. (2011). Identification and quantification of a major anti-oxidant and anti- inflammatory phenolic compound found in basil, lemon thyme, mint, oregano, rosemary, sage, and thyme. International Journal of Food Sciences and Nutrition, 62, 577-584. Pearson, W., Fletcher, R. S., Kott, L. S. & Hurtig, M. B. (2010). Protection against LPS- induced cartilage inflammation and degradation provided by a biological extract of Mentha spicata. BMC Complementary and Alternative Medicine, 10, 19. Peng, J., Elias, J. E., Thoreen, C. C., Licklider, L. J. & Gygi, S. P. (2003). Evaluation of multidimensional chromatography coupled with tandem mass spectrometry (LC/LC-MS/MS) for large-scale protein analysis: the yeast proteome. Journal of Proteome Research, 2, 43-50. Pereira, O. R. & Cardoso, S. M. (2013). Overview on Mentha and Thymus polyphenols. Current Analytical Chemistry, 9, 382-396. Peter, K. V. (2012). Handbook of Herbs and Spices, Cambridge, Woodhead Publishing Ltd. Petersen, M. & Simmonds, M. S. (2003). Rosmarinic acid. Phytochemistry, 62, 121- 125. Phelps, D. G. (1997). Bush food industry in Western Queensland, Barton, Rural Industries Research and Development Corporation (RIRDC). Pietrzyk, D. J. & Chu, C. H. (1977). Amberlite XAD copolymers in reversed phase gravity flow and high pressure liquid chromatography. Analytical Chemistry, 49, 757-764. Pietta, P. (2000). Flavonoids as antioxidants. Journal of Natural Products, 63, 1035- 1042. Pinheiro, P. F. & Justino, G. C. 2010. Structural analysis of flavonoids and related compounds-a review of spectroscopic applications. In: RAO, V. (ed.) Phytochemicals—A global perspective of their roles in nutrition health. Boca Raton: InTech. CRC press. Pitt, J. J. (2009). Principles and applications of liquid chromatography-mass spectrometry in clinical biochemistry. Clin Biochem Rev, 30, 19-34. Polettini, A. (2006). Applications of LC-MS in Toxicology, London, Pharmaceutical Press. Popova, M., Bankova, V., Butovska, D., Petkov, V., Nikolova‐Damyanova, B., Sabatini, A. G., Marcazzan, G. L. & Bogdanov, S. (2004). Validated methods for the quantiﬁcation of biologically active constituents of poplar‐type propolis. Phytochemical Analysis, 15, 235-240. Proestos, C. & Komaitis, M. (2006). Ultrasonically assisted extraction of phenolic compounds from aromatic plants: comparison with conventional extraction technics. Journal of Food Quality, 29, 567-582. Proestos, C. & Komaitis, M. (2013). Analysis of Naturally Occurring Phenolic Compounds in Aromatic Plants by RP-HPLC Coupled to Diode Array Detector (DAD) and GC-MS after Silylation. Foods, 2, 90-99. Proestos, C., Sereli, D. & Komaitis, M. (2006). Determination of phenolic compounds in aromatic plants by RP-HPLC and GC-MS. Food Chemistry, 95, 44-52. 209

Pyrzynska, K. & Biesaga, M. (2009). Analysis of phenolic acids and flavonoids in honey. Trends in Analytical Chemistry, 28, 893-902. Raaman, N. (2006). Phytochemical Techniques, New India Publishing Agency. Rai, R. (2016). Advances in Food Biotechnology, Chichester, West Sussex, John Wiley & Sons. Ramassamy, C. (2006). Emerging role of polyphenolic compounds in the treatment of neurodegenerative diseases: a review of their intracellular targets. European Journal of Pharmacology, 545, 51-64. Rao, A. V. & Rao, L. G. (2007). Carotenoids and human health. Pharmacological Research, 55, 207-216. Reuter, S., Gupta, S. C., Chaturvedi, M. M. & Aggarwal, B. B. (2010). Oxidative stress, inflammation, and cancer: how are they linked? Free Radical Biology and Medicine, 49, 1603-1616. Rios, J. & Recio, M. (2005). Medicinal plants and antimicrobial activity. Journal of Ethnopharmacology, 100, 80-84. Robards, K. (2003). Strategies for the determination of bioactive phenols in plants, fruit and vegetables. Journal of Chromatography A, 1000, 657-691. Russell, M., Southwell, I. A. & Waterman, P. G. (2001). α-Phellandren-8-ol, a Rare Monoterpenene Alcohol from the Oil of Prostanthera staurophylla F. Muell. Journal of Essential Oil Research, 13, 446-447. Sakulnarmrat, K., Fenech, M., Thomas, P. & Konczak, I. (2013). Cytoprotective and pro-apoptotic activities of native Australian herbs polyphenolic-rich extracts. Food Chemistry, 136, 9-17. Sakulnarmrat, K. & Konczak, I. (2012). Composition of native Australian herbs polyphenolic-rich fractions and in vitro inhibitory activities against key enzymes relevant to metabolic syndrome. Food Chemistry, 134, 1011-1019. Sakulnarmrat, K., Srzednicki, G. & Konczak, I. (2014). Composition and inhibitory activities towards digestive enzymes of polyphenolic-rich fractions of Davidson's plum and quandong. LWT-Food Science and Technology, 57, 366- 375. Sakulnarmrat, K., Srzednicki, G. & Konczak, I. (2015). Bioprospecting Davidson's plum and quandong: Cytoprotective and proapoptotic activities. LWT-Food Science and Technology, 61, 622-629. Samee, H., Li, Z. X., Lin, H., Khalid, J. & Guo, Y. C. (2009). Anti-allergic effects of ethanol extracts from brown seaweeds. Journal of Zhejiang University Science B, 10, 147-153. Schaich, K., Tian, X. & Xie, J. (2015). Hurdles and pitfalls in measuring antioxidant efficacy: a critical evaluation of ABTS, DPPH, and ORAC assays. Journal of Functional Foods, 14, 111-125. Scherer, R., Lemos, M. F., Lemos, M. F., Martinelli, G. C., Martins, J. D. L. & Da Silva, A. G. (2013). Antioxidant and antibacterial activities and composition of Brazilian spearmint (Mentha spicata L.). Industrial Crops and Products, 50, 408-413. Schieber, A., Keller, P. & Carle, R. (2001). Determination of phenolic acids and flavonoids of apple and pear by high-performance liquid chromatography. Journal of Chromatography A, 910, 265-273. Sergent, T., Vanderstraeten, J., Winand, J., Beguin, P. & Schneider, Y. J. (2012). Phenolic compounds and plant extracts as potential natural anti-obesity substances. Food Chemistry, 135, 68-73.

210

Shahidi, F. (1997). Natural Antioxidants: Chemistry, Health Effects, and Applications, Champaign, Illinois, AOCS Press. Shahrbandy, K. & Hosseinzadeh, R. (2007). In vitro Antioxidant Activity of Polygonium hyrcanicum, Centaurea depressa, Sambucus ebutus, Mentha spicata and Phytolacca americana. Pakistan Journal of Biological Sciences, 10, 637-640. Shan, B., Cai, Y. Z., Sun, M. & Corke, H. (2005). Antioxidant capacity of 26 spice extracts and characterization of their phenolic constituents. Journal of Agricultural and Food Chemistry, 53, 7749-7759. Sharafi, S. M., Rasooli, I., Owlia, P., Taghizadeh, M. & Astaneh, S. (2010). Protective effects of bioactive phytochemicals from Mentha piperita with multiple health potentials. Pharmacognosy Magazine, 6, 147. Sharma, P., Jha, A. B., Dubey, R. S. & Pessarakli, M. (2012). Reactive oxygen species, oxidative damage, and antioxidative defense mechanism in plants under stressful conditions. Journal of Botany, 2012, 26. Shen, D., Pan, M. H., Wu, Q. L., Park, C. H., Juliani, H. R., Ho, C. T. & Simon, J. E. (2011). A rapid LC/MS/MS method for the analysis of nonvolatile antiinflammatory agents from Mentha spp. Journal of Food Science, 76, C900- C908. Shi, S., Huang, K., Zhang, Y. & Liu, S. (2008). Preparative isolation and purification of two flavonoid glycosides from Taraxacum mongolicum by high-speed counter- current chromatography. Separation and Purification Technology, 60, 81-85. Shoji, T., Mutsuga, M., Nakamura, T., Kanda, T., Akiyama, H. & Goda, Y. (2003). Isolation and structural elucidation of some procyanidins from apple by low- temperature nuclear magnetic resonance. Journal of Agricultural and Food Chemistry, 51, 3806-3813. Simpson, B. K., Nollet, L. M. L., Toldrã¡, F., Benjakul, S., Paliyath, G. & Hui, Y. H. (2012). Food Biochemistry and Food Processing, Chichester, West Sussex, John Wiley & Sons. Singhal, J., Nagaprashantha, L. D., Vatsyayan, R., Awasthi, S. & Singhal, S. S. (2012). Didymin induces apoptosis by inhibiting N-Myc and upregulating RKIP in neuroblastoma. Cancer Prevention Research, 5, 473-483. Skoog, D. A., Holler, F. J. & Crouch, S. R. (2007). Principles of Instrumental Analysis, Belmont, CA, Thomson Brooks/Cole, Cengage Learning. Soberón, J. R., Sgariglia, M. A., Sampietro, D. A., Quiroga, E. N. & Vattuone, M. A. (2010). Free radical scavenging activities and inhibition of inflammatory enzymes of phenolics isolated from Tripodanthus acutifolius. Journal of Ethnopharmacology, 130, 329-333. Sobolev, V. S., Horn, B. W., Potter, T. L., Deyrup, S. T. & Gloer, J. B. (2006). Production of stilbenoids and phenolic acids by the peanut plant at early stages of growth. Journal of Agricultural and Food Chemistry, 54, 3505-3511. Soković, M. D., Vukojević, J., Marin, P. D., Brkić, D. D., Vajs, V. & Van Griensven, L. J. (2009). Chemical Composition of Essential Oilsof Thymus and Mentha Species and Their Antifungal Activities. Molecules, 14, 238-249. Sommano, S., Caffin, N., Mcdonald, J. & Cocksedge, R. (2011). The impact of thermal processing on bioactive compounds in Australian native food products (bush tomato and Kakadu plum). Food Research International, 50, 557-561. Soto, M. L., Moure, A., Domínguez, H. & Parajó, J. C. (2011). Recovery, concentration and purification of phenolic compounds by adsorption: a review. Journal of Food Engineering, 105, 1-27. 211

Southwell, I. & Tucker, D. (1996). cis-Dihydroagarofuran, A Novel Sesquiterpene Ether, from Prostanthera sp. aff. ovalifolia. Journal of Essential Oil Research, 8, 585-586. Southwell, I. A. & Brophy, J. J. (1992). Differentiation within the australian Tasmannia by essential oil comparison. Phytochemistry, 31, 3073-3081. Southwell, I. A., Russell, M., Smith, R. L. & Archer, D. W. (2000). Backhousia citriodora F. Muell.(Myrtaceae), a superior source of citral. Journal of Essential Oil Research, 12, 735-741. Sulieman, A. M. E., Abdelrahman, S. E. & Rahim, A. M. A. (2011). Phytochemical Analysis of Local Spearmint (Mentha spicata) Leaves and Detection of the Antimicrobial Activity of its Oil. Journal of Microbiology Research, 1, 1-4. Sultana, B., Anwar, F. & Ashraf, M. (2009). Effect of extraction solvent/technique on the antioxidant activity of selected medicinal plant extracts. Molecules, 14, 2167-2180. Sultanbawa, Y. & Sultanbawa, F. (2016). Australian Native Plants: Cultivation and Uses in the Health and Food Industries, CRC Press. Tadera, K., Minami, Y., Takamatsu, K. & Matsuoka, T. (2006). Inhibition of α- Glucosidase and α-Amylase by Flavonoids. Journal of Nutritional Science and Vitaminology, 52, 149-153. Tan, A. C., Hou, D. X., Konczak, I., Tanigawa, S., Ramzan, I. & Sze, D. M. Y. (2011a). Native Australian fruit polyphenols inhibit COX-2 and iNOS expression in LPS- activated murine macrophages. Food Research International, 44, 2362-2367. Tan, A. C., Konczak, I., Ramzan, I. & Sze, D. M. Y. (2011b). Antioxidant and cytoprotective activities of native Australian fruit polyphenols. Food Research International, 44, 2034-2040. Tan, A. C., Konczak, I., Ramzan, I. & Sze, D. M. Y. (2011c). Native Australian fruit polyphenols inhibit cell viability and induce apoptosis in human cancer cell lines. Nutrition and Cancer, 63, 444-455. Tan, A. C., Konczak, I., Ramzan, I., Zabaras, D. & Sze, D. M. (2011d). Potential antioxidant, antiinflammatory, and proapoptotic anticancer activities of Kakadu plum and Illawarra plum polyphenolic fractions. Nutrition and Cancer, 63, 1074-84. Tepe, B., Eminagaoglu, O., Akpulat, H. A. & Aydin, E. (2007). Antioxidant potentials and rosmarinic acid levels of the methanolic extracts of Salvia verticillata (L.) subsp. verticillata and S. verticillata (L.) subsp. amasiaca (Freyn & Bornm.) Bornm. Food Chemistry, 100, 985-989. Toldrá, F. (2009). Safety of Meat and Processed Meat - Food Microbiology and Food Safety, New York, Springer Tucker, G. & Robards, K. (2008). Bioactivity and structure of biophenols as mediators of chronic diseases. Critical Reviews in Food Science and Nutrition, 48, 929- 966. Uribe, E., Marín, D., Vega-Gálvez, A., Quispe-Fuentes, I. & Rodríguez, A. (2015). Assessment of vacuum-dried peppermint (Mentha piperita L.) as a source of natural antioxidants. Food Chemistry. Uttara, B., Singh, A. V., Zamboni, P. & Mahajan, R. (2009). Oxidative stress and neurodegenerative diseases: a review of upstream and downstream antioxidant therapeutic options. Current neuropharmacology, 7, 65-74. Van Den Ende, W. & Peshev, D. 2013. Sugars as antioxidants in plants. In: TUTEJA, N. A. G., S. S. (ed.) Crop improvement under adverse conditions. New York: Springer. 212

Velioglu, Y., Mazza, G., Gao, L. & Oomah, B. (1998). Antioxidant activity and total phenolics in selected fruits, vegetables, and grain products. Journal of Agricultural and Food Chemistry, 46, 4113-4117. Vuong, Q. V., Hirun, S., Chuen, T. L., Goldsmith, C. D., Bowyer, M. C., Chalmers, A. C., Phillips, P. A. & Scarlett, C. J. (2014). Physicochemical composition, antioxidant and anti-proliferative capacity of a lilly pilly (Syzygium paniculatum) extract. Journal of Herbal Medicine, 4, 134-140. Wang, H., Provan, G. J. & Helliwell, K. (2004). Determination of rosmarinic acid and caffeic acid in aromatic herbs by HPLC. Food Chemistry, 87, 307-311. Weerakkody, N. S., Caffin, N., Turner, M. S. & Dykes, G. A. (2010). In vitro antimicrobial activity of less-utilized spice and herb extracts against selected food-borne bacteria. Food Control, 21, 1408-1414. Williams, C. (2010). Medicinal Plants in Australia Volume 1: Bush Pharmacy, NSW, Rosenberg Publishing Pty, Limited. Winnett, V., Boyer, H., Sirdaarta, J. & Cock, I. E. (2014). The potential of Tasmannia lanceolata as a natural preservative and medicinal agent: Antimicrobial activity and toxicity. Pharmacognosy Communications, 4, 42-52. Wojdyło, A., Oszmiański, J. & Czemerys, R. (2007). Antioxidant activity and phenolic compounds in 32 selected herbs. Food Chemistry, 105, 940-949. Wong, S. P., Leong, L. P. & Koh, J. H. W. (2006). Antioxidant activities of aqueous extracts of selected plants. Food Chemistry, 99, 775-783. Wongsa, P., Chaiwarit, J. & Zamaludien, A. (2012). In vitro screening of phenolic compounds, potential inhibition against α-amylase and α-glucosidase of culinary herbs in Thailand. Food Chemistry, 131, 964-971. Wright, M. H., Matthews, B., Greene, A. C. & Cock, I. E. (2015). Growth inhibition of the zoonotic bacteria Bacillus anthracis by high antioxidant Australian plants: New leads for the prevention and treatment of anthrax. Pharmacognosy Communications, 5, 173-189. Wu, A. H., Gerona, R., Armenian, P., French, D., Petrie, M. & Lynch, K. L. (2012). Role of liquid chromatography–high-resolution mass spectrometry (LC-HR/MS) in clinical toxicology. Clinical Toxicology, 50, 733-742. Wu, X., Beecher, G. R., Holden, J. M., Haytowitz, D. B., Gebhardt, S. E. & Prior, R. L. (2004). Lipophilic and hydrophilic antioxidant capacities of common foods in the United States. Journal of Agricultural and Food Chemistry, 52, 4026-4037. Wu, X. & Prior, R. L. (2005). Systematic identification and characterization of anthocyanins by HPLC-ESI-MS/MS in common foods in the United States: fruits and berries. Journal of Agricultural and Food Chemistry, 53, 2589-2599. Xu, Z. & Howard, L. R. (2012). Analysis of Antioxidant-Rich Phytochemicals, Chichester, West Sussex, John Wiley & Sons. Yang, R. Y., Lin, S. & Kuo, G. (2008). Content and distribution of flavonoids among 91 edible plant species. Asia Pacific Journal of Clinical Nutrition, 17, 275-279. Yang, S. A., Jeon, S. K., Lee, E. J., Shim, C. H. & Lee, I. S. (2010). Comparative study of the chemical composition and antioxidant activity of six essential oils and their components. Natural Product Research, 24, 140-151. Yeoh, H. H. & Wee, Y. C. (1994). Leaf protein contents and nitrogen-to-protein conversion factors for 90 plant species. Food Chemistry, 49, 245-250. Yosr, Z., Hnia, C., Rim, T. & Mohamed, B. (2013). Changes in essential oil composition and phenolic fraction in Rosmarinus officinalis L. var. typicus Batt. organs during growth and incidence on the antioxidant activity. Industrial Crops and Products, 43, 412-419. 213

Yusof, S., Ghazali, H. M. & King, G. S. (1990). Naringin content in local citrus fruits. Food Chemistry, 37, 113-121. Zhang, K. & Zuo, Y. (2004). GC-MS determination of flavonoids and phenolic and benzoic acids in human plasma after consumption of cranberry juice. Journal of Agricultural and Food Chemistry, 52, 222-227. Zhang, Y., Seeram, N. P., Lee, R., Feng, L. & Heber, D. (2008). Isolation and identification of strawberry phenolics with antioxidant and human cancer cell antiproliferative properties. Journal of Agricultural and Food Chemistry, 56, 670-675. Zhang, Y., Xiao, G., Sun, L., Wang, Y., Wang, Y. & Wang, Y. (2013). A new flavan-3- ol lactone and other constituents from Euonymus alatus with inhibitory activities on α-glucosidase and differentiation of 3T3-L1 cells. Natural Product Research, 27, 1513-1520. Zhao, J. & Agboola, S. (2007). Functional Properties of Australian Bushfoods. Rural Industries Research and Development Corporation, Australia. Zheng, J., Gao, H.-Y., Chen, G.-T., Yang, X.-K., Wu, B. & Wu, L.-J. (2006). Chemical constituents of the active parts of Mentha spicata L.(Ⅱ). Journal of Shenyang Pharmaceutical University, 4, 004. Zheng, W. & Wang, S. Y. (2001). Antioxidant activity and phenolic compounds in selected herbs. Journal of Agricultural and Food Chemistry, 49, 5165-5170. Zuo, Y., Wang, C. & Zhan, J. (2002). Separation, characterization, and quantitation of benzoic and phenolic antioxidants in American cranberry fruit by GC-MS. Journal of Agricultural and Food Chemistry, 50, 3789-3794.

214

APPENDICES

Assay Equation R2

Folin-Ciocalteu assay y = 0.0016x + 0.2011 0.9992 Total flavonoids y = 0.0021x + 0.0205 0.9990 ABTS radical scavenging capacity y = -0.001x + 1.1291 0.9988 DPPH radical scavenging capacity y = -0.0051x + 1.0797 0.9997 Ferric reducing antioxidant power y = 0.0007x + 0.1033 0.9994 Oxygen radicals absorbance capacity y = 606.46x + 6818 0.9999 α-Glucosidase inhibitory activity y = -0.0007x + 1.1022 0.9867 α-Amylase inhibitory activity y = -0.0011x + 0.6531 0.9743 Pancreatic lipase inhibitory y = -305.58x + 395.74 0.9687

Appendix 1. Regression line equation for antioxidant and enzyme-based assays

215

A B

C D

E F

Appendix 2. Spectra of (a) chlorogenic acid, rosmarinic acid, caffeic acid (b) ferulic, cinnamic acid, (c) rutin hydrate, (d) naringin, naringenin, hesperetin, hesperidin, neoponcirin and (e), apigenin, (f) p-coumaric acid and (g) sinapic acid.

216

C:\Users\...\140526_m_nonspiked 26/05/2014 4:37:05 PM

RT: 12.46 - 15.23 SM: 7G RT: 13.49 NL: 9.83E6 100 TIC F: FTMS - p ESI Full 80 ms [120.00-1000.00] MS ICIS 60 140526_m_nonspiked

20 RT: 14.57 RT: 15.02 Relative Relative Abundance RT: 12.64 0 RT: 14.16 NL: 1.91E4 100 m/z= 283.0591-283.0622 80 F: FTMS - p ESI Full ms [120.00-1000.00] MS 60 ICIS 140526_m_nonspiked 40

0 12.5 13.0 13.5 14.0 14.5 15.0 Time (min)

283.0603 NL: 100 M. australis 1.45E4 80 140526_m_nonspiked#74 1-754 RT: 14.05-14.27 60 AV: 7 F: FTMS - p ESI 285.0760 Full ms [120.00-1000.00] 40

20 281.9330 284.0638 Relative Relative Abundance 286.0793 276.0338 279.0112 280.9821 287.6057 0 283.0612 NL: 100 Simulated data 8.31E5 80 C16 H12 O5 +H: C16 H11 O5 60 pa Chrg -1

20 284.0646 285.0679 287.0722 0 276 278 280 282 284 286 288 m/z

Appendix 3. Comparison of molecular ion [M-1]- (m/z 283) of biochanin A in M. australis and the simulated data generated from Xcalibur by LC-HRMS.

217

C:\Users\...\140526_m_nonspiked 26/05/2014 4:37:05 PM

RT: 0.00 - 20.02 SM: 7G RT: 13.49 NL: 9.83E6 100 TIC F: FTMS - p ESI Full 80 ms [120.00-1000.00] MS ICIS 60 140526_m_nonspiked RT: 15.02 40 RT: 16.99 RT: 12.18 20 RT: 18.08 Relative Relative Abundance RT: 2.02 RT: 5.84 RT: 10.50 0 RT: 10.42 NL: 1.78E4 100 m/z= 579.1700-579.1720 80 F: FTMS - p ESI Full ms [120.00-1000.00] MS 60 ICIS 140526_m_nonspiked 40

0 0 2 4 6 8 10 12 14 16 18 20

Time (min)

579.1701 NL: 100 M. australis 1.44E4 80 140526_m_nonspiked#53 9-557 RT: 10.27-10.61 60 AV: 10 F: FTMS - p ESI Full ms [120.00-1000.00] 40 581.1763

20 572.3714 585.5578 Relative Relative Abundance 549.1959 567.2066 595.2014 609.1445 0 579.1719 NL: 100 Simulated data 7.20E5 80 C27 H32 O14 +H: C27 H31O14 60 pa Chrg -1

20 584.1838 0 550 560 570 580 590 600 610 m/z

Appendix 4. Comparison of molecular ion [M-1]- (m/z 579) of narirutin in M. australis and the simulated data generated from Xcalibur by LC-HRMS.

218

C:\Users\...\140526_m_nonspiked 26/05/2014 4:37:05 PM

RT: 0.00 - 20.02 SM: 7G RT: 13.49 NL: 9.83E6 100 TIC F: FTMS - p ESI Full 80 ms [120.00-1000.00] MS ICIS 60 140526_m_nonspiked RT: 15.02 40 RT: 16.99 RT: 12.18 20 RT: 18.08 Relative Relative Abundance RT: 2.02 RT: 5.84 RT: 10.50 0 RT: 2.21 NL: 3.94E4 100 m/z= 353.0868-353.0888 80 F: FTMS - p ESI Full ms [120.00-1000.00] MS 60 ICIS 140526_m_nonspiked 40

20 RT: 3.93

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

353.0869 M. australis NL: 100 2.62E4 80 140526_m_nonspiked#1 07-127 RT: 2.02-2.40 60 AV: 11 F: FTMS - p ESI Full ms 40 [120.00-1000.00] 287.0553 20 537.1020 575.0789

Relative Relative Abundance 312.0941 373.0554 467.0792 519.0914 636.1281 0 353.0878 NL: 100 Simulated data 8.22E5 80 C16 H18 O9 +H: C16 H17 O9 60 pa Chrg -1

0 250 300 350 400 450 500 550 600 650 m/z

Appendix 5. Comparison of molecular ion [M-1]- (m/z 353) of chlorogenic acid in M. australis and the simulated data generated from Xcalibur by LC-HRMS.

219

Breakdown Curve of Ion 359.0 m/z Breakdown Curve of Ion 271.0 m/z Intensity: 2.88e+07 Product Ions Coll.Energy Intensity: 8.97e+05 Product Ions Coll.Energy Pressure: 1.5 mTorr 161.2 m/z 18 v Pressure: 1.5 mTorr 119.4 m/z 30 v 100 133.2 m/z 43 v 100 151.2 m/z 19 v 135.3 m/z 40 v 177.2 m/z 20 v 179.1 m/z 26 v 80 80

Rosmarinic acid Naringenin 60 60

40 40

Relative Intensity Relative Relative Intensity Relative

20 20

0 0 5 20 35 50 65 80 5 20 35 50 65 80 Collision Energy (V) Collision Energy (V)

Breakdown Curve of Ion 301.0 m/z Intensity: 5.56e+05 Product Ions Coll.Energy Breakdown Curve of Ion 269.0 m/z Pressure: 1.5 mTorr 164.1 m/z 24 v Intensity: 4.99e+05 Product Ions Coll.Energy 100 242.1 m/z 20 v Pressure: 1.5 mTorr 117.4 m/z 41 v 200.9 m/z 25 v 100 151.2 m/z 26 v 199.1 m/z 22 v 149.2 m/z 25 v 80 80

60 Hesperetin Apigenin 60

Relative Intensity Relative Relative Intensity Relative

20 20

0 0 5 20 35 50 65 80 5 20 35 50 65 80 Collision Energy (V) Collision Energy (V)

Breakdown Curve of Ion 353.0 m/z Breakdown Curve of Ion 609.0 m/z Intensity: 1.08e+06 Product Ions Coll.Energy Intensity: 1.29e+07 Product Ions Coll.Energy Pressure: 1.5 mTorr 191.1 m/z 10 v Pressure: 1.5 mTorr 301.0 m/z 27 v 100 100 266.2 m/z 26 v 242.0 m/z 43 v 338.2 m/z 17 v 324.8 m/z 33 v 320.9 m/z 18 v 286.1 m/z 46 v 80 80

Chlorogenic acid Hesperidin 60 60

40 40

Relative Intensity Relative Relative Intensity Relative

20 20

0 0 -15 0 15 30 45 60 5 20 35 50 65 80 Collision Energy (V) Collision Energy (V)

Appendix 6. Product ions of reference compounds for Tandem Mass Spectrometry

220

Collision Energy Curves of SRM Transitions SRM Transition Coll.Energy Pressure: 1.5 mTorr 271.0 m/z => 119.3 m/z 26 v 100 285.0 m/z => 133.2 m/z 38 v 301.0 m/z => 300.4 m/z 10 v 317.0 m/z => 316.4 m/z 11 v 80 359.0 m/z => 161.2 m/z 31 v

40 Relative Intensity Relative

0 5 20 35 50 65 80 Collision Energy (V)

Optimizing Tube Lens for Q1MS Previous Setting Optimum Setting 100

40 Relative Intensity Relative

Mass 359.00m/z : 16 % Improvement 20 Mass 317.00m/z : 7 % Improvement Mass 301.00m/z : 6 % Improvement Mass 285.00m/z : 2 % Improvement Mass 271.00m/z : 13 % Improvement 0 -240 -190 -140 -90 -40 10 Tube Lens Voltage (V) Apr 3, 2014 Appendix 7. Collision energy obtained from phenolic standards.

221

Appendix 8. 1H NMR spectrum of purified fraction of peak 6 extracted from M. australis.

2 16

222

Appendix 9. 13C NMR spectrum of purified fraction of peak 6 extracted from M. australis.

223

217 Appendix 10. HSQC spectrum of purified fraction peak 6 from M. australis (CD3OD).

224

218 Appendix 11. HMBC spectrum of purified fraction peak 6 from M. australis (CD3OD).

225

219 Appendix 12. COSY spectrum of purified mixture from M. australis (CD3OD).

226

RT: 0.00 - 20.01 SM: 7G 100 NL: 1.71E7 TIC F: FTMS - p APCI corona Full 90 ms [120.00-700.00] 80 MS 140521_P_1ug_a_ neg 70

40 RelativeAbundance

0 0 2 4 6 8 10 12 14 16 18 20 Time (min) Appendix 13. Total ion current (TIC) chromatogram from LC-HRMS of anions in P. rotundifolia using negative atmospheric pressure chemical ionisation.

227

221

Appendix 14. HSQC spectrum of purified fraction peaks 4 (verbascoside) from P. rotundifolia (CD3OD).

228

Appendix 15. HMBC spectrum of purified fraction peaks 4 (above) and 6 (below) from

P. rotundifolia (CD3OD).

229