Tyrosinase Inhibitors from Traditional Chinese Medicine As Potential Anti-Hyperpigmentation Agents

Tyrosinase Inhibitors from Traditional Chinese Medicine as Potential Anti-hyperpigmentation agents

Author Shi, Gengxuan

Published 2019

Thesis Type Thesis (Masters)

School School of Environment and Sc

DOI https://doi.org/10.25904/1912/1780

Downloaded from http://hdl.handle.net/10072/387690

Griffith Research Online https://research-repository.griffith.edu.au

Tyrosinase Inhibitors from Traditional Chinese Medicine as

Potential Anti-hyperpigmentation agents

Gengxuan SHI

School of Environment and Science

Griffith Science

Griffith University

Submitted in fulfilment of the requirements of the degree of

Master of Science

July 2019

Abstract

Several treatments for skin pigmentation are available today, however, many have unwanted side effects. Classic drugs like hydroquinone, arbutin, mequinol and kojic acid have been considered strongly carcinogenic, with related adverse effects. Traditional Chinese medicines (TCMs) have long been documented for their skin- lightening properties with little known negative effects. However, little is known about precisely how these herbal medicines work and what are the chemical basis for their activity.

In this project, a 96-well plate based tyrosinase assay was established and used to test 44 TCM with known skin-lightening properties. Out of 44, 17 TCM extracts showed over 60% inhibition against tyrosinase at the concentration of 0.5 mg/mL. One of the TCM extracts, Xanthium strumarium L. extract, possessed 81.7% inhibition at 0.5 mg/mL. Further bioassay-guided isolation of the crude extract resulted in 10 compounds. Three of the compounds showed moderate activity against tyrosinase with IC50 values of 0.18 mM (cytidine), 0.29 mM (1,4-dicaffeoylquinic acid) and 2.47 mM (xanthiside).

Chapter 1 contained the introduction of this project as well as the literature review. Chapter 2 was mainly about the screening of TCM extracts against mushroom tyrosinase. Chapter 3 contained the bioassay-guided extraction and isolation of

Xanthium strumarium L. Additionally, the IC50 values of the active compounds were included in chapter 3 as well.

Statement of Originality

This work has not previously been submitted for a degree or diploma in any university. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made in the thesis itself.

(Signed)_

iii

Table of contents

Chapter 1. Introduction and Literature Review ...... 1 1.1 Hyperpigmentation ...... 1 1.2 Melanin Biosynthesis ...... 4 1.3 Tyrosinase ...... 7 1.4 Tyrosinase inhibitors from natural products ...... 8 1.4.1 Flavonoids ...... 9 1.4.2 Organic acids ...... 10 1.4.3 Glycoside ...... 11 1.4.4 Terpenes ...... 12 1.4.5 Esters ...... 14 1.4.6 Other compounds ...... 15 1.5 Traditional Chinese Medicine (TCM) ...... 16 1.6 Project aims ...... 18 Chapter 2. Tyrosinase assay and the screening of TCM extracts against tyrosinase .. 23 2,1 Mushroom tyrosinase assay ...... 23 2.2 Positive controls ...... 26 2.3 selection of TCM against hyperpigmentation ...... 27 2.4 Anti-tyrosinase activity ...... 30 2.5 Experimental ...... 32 Protein sequence alignments...... 32 Chemicals ...... 32 Tyrosinase assay ...... 32 Crude extracts preparations ...... 33 Chapter 3 Chemical investigation of plant Xanthium Strumarium L...... 35 3.1 Xanthium Strumarium L...... 35 3.2 Bioassay-guided extraction and isolation ...... 40 3.2.1 Further purification for fraction B ...... 42 3.2.2 Further purification for fraction C ...... 49 3.2.3 Further purification for fraction D ...... 51 3.2.4 Further purification for fraction E ...... 57 3.3 Inhibitory activity of the pure compounds against tyrosinase ...... 60 3.4 Conclusions and discussions ...... 61 3.5 Experimental ...... 64 General instrument ...... 64 Bioassay-guided extraction and isolation ...... 64 Appendix ...... 72

List of figures Figure 1.1 Atomic force microscopy performed on epidermal melanocytes...... 2 Figure 1.2 Hypopigmentation (left) verse hyperpigmentation (right) ...... 2 Figure 1.3 Symptom of three kinds of hyperpigmentation...... 3 Figure 1.4 Biosynthesis of melanin ...... 5 Figure 1.5 Signaling pathways of series synthesis of eumelanin and pheomelanin.[34] ...... 7 Figure 1.6 Representitive tyrosinase inhibitory active flavonoids ...... 10 Figure 1.7 Representative reported natural organic acids with inhibitory activity against tyrosinase ...... 11 Figure 1.8 Representative natural glycoside with inhibitory activity against tyrosinase ...... 12 Figure 1.9 representative terpenes reported as tyrosinase inhibitors...... 14 Figure 1.10 Representative esters as natural tyrosinase inhibitors ...... 15 Figure 1.11 Other compounds reported as natural tyrosinase inhibitors ...... 16 Figure 1.12 Representative TCM traditionally used for skin-lightening purpose ...... 17

Figure 2.1 Sequence alignment of human tyrosinase ...... 24 Figure 2.2 Sequence alignment of mushroom tyrosinase (C7FF04) and TYRP1(P17643)...... 25 Figure 2.3 The metal binding sites of TYRP1 (A) and mushroom tyrosinase (B) ...... 26 Figure 2.4 The inhibition of arbutin against tyrosinase...... 26 Figure 2.5 The inhibition of kojic acid against tyrosinase...... 27 Figure 2.7 HTS tyrosinase assay results ...... 31 Figure 3.1A Photographs of Xanthium Strumarium L...... 35 Figure 3.1B Isolated compounds with anti-cancer effects from Xanthium ...... 36 Figure 3.1C Representative isolated compounds from Xanthium ...... 37 Scheme 3.2A Bioassay-guided isolation of 10 compounds from Xanthium...... 38 Figure 3.2B All 10 compounds isolated from Xanthium Strumarium L...... 39 Figure 3.3 purchased TCM Xanthium...... 40 Figure 3.2 Inhibitory activity of hexane, DCM, methanol extract against tyrosinase ...... 40 Figure 3.4A Semi-preparative HPLC UV trace of methanol extract of Xanthium with gradient

MeOH/H2O (0.1% TFA)...... 41 Figure 3.4B Inhibitory activity of seven fractions from semi-preparative HPLC chromatography ...... 41

1 Figure 3.5 H NMR (DMSO-d6, 800MHz) spectrum of Fraction A-E ...... 42

Figure 3.6 Semi-preparative HPLC UV trace of Fraction B with gradient MeOH/H2O (0.1%TFA)...... 43 Figure 3.7 Inhibitory activity of combined fractions after HPLC chromatography of Fraction B 43

Figure 3.8A Semi-preparative HPLC UV trace of Fraction B2 with gradient MeOH/H2O (0.1% TFA). Timed fractions (1 min x 30) were collected. (Black at 210 nm, green at 280 nm, pink at 254 nm, blue at 320 nm) ...... 44

Figure 3.8B Semi-preparative HPLC UV trace of Fraction B4 with gradient MeOH/H2O (0.1% TFA) ...... 44

1 Figure 3.9 H NMR (DMSO-d6, 800 MHz) spectra of Uridine (3.43) ...... 45

1 Figure 3.10 H NMR (DMSO-d6, 800 MHz) spectra of Cytidine (3.44) ...... 45

1 Figure 3.11 H NMR (CD3OD, 800 MHz) spectrum of Caffeoylcholine (3.41) ...... 46

1 Figure 3.12 H NMR (CD3OD, 800 MHz) spectrum of 4-caffeoylquinic acid (3.36) ...... 48

Figure 3.13A Semi-preparative HPLC UV trace of Fraction C with gradient MeOH/H2O (0.1% TFA) ...... 49 Figure 3.13B Inhibitory activity of combined fractions after HPLC chromatography of Fraction C ...... 49

1 Figure 3.14 H NMR (CD3OD, 800 MHz) spectrum of 3-caffeoylquinic acid (3.35) ...... 50

Figure 3.15A Semi-preparative HPLC UV trace of Fraction D with gradient MeOH/H2O (0.1% TFA) ...... 51 Figure 3.15B Inhibitory activity of combined fractions after HPLC chromatography of Fraction D ...... 52

Figure 3.16A Semi-preparative HPLC UV trace of Fraction D5 with gradient MeOH/H2O (0.1% TFA) ...... 52 Figure 3.16B Inhibitory activity of combined fractions after HPLC chromatography of Fraction D5 ...... 53

1 Figure 3.17 H NMR (CD3OD, 800 MHz) spectrum of Xanthiside (3.42) ...... 53

1 Figure 3.18 H NMR (CD3OD, 800 MHz) spectrum of 1,3-dicaffeoylquinic acid (3.37) ...... 54

1 Figure 3.19 H NMR (CD3OD, 800 MHz) spectrum of 3,5-dicaffeoylquinic acid (3.38) ...... 55

Figure 3.20A Semi-preparative HPLC UV trace of Fraction E with gradient MeOH/H2O (0.1% TFA) ...... 57 Figure 3.20B Inhibitory activity of combined fractions after HPLC chromatography of Fraction E ...... 57

1 Figure 3.21A H NMR (CD3OD, 800 MHz) spectrum of 1,5-dicaffeoylquinic acid (3.39) ...... 58

1 Figure 3.22 H NMR (CD3OD, 800 MHz) spectrum of 4,5-dicaffeoylquinic acid (3.40) ...... 59 Figure 3.23 Inhibitory activity of isolated compounds against tyrosinase under concentration of 0.5mg/mL ...... 60 Figure 3.24 Dose-response curve of the inhibition of cytidine (3.44), 1,3-dicaffeoylquinic acid (3.37), xanthiside (3.42) against tyrosinase...... 61

Figure 3.25 Comparison of IC50 values of cytidine (3.44), 1,3-dicaffeoylquinic acid (3.37) and xanthiside (3.42) to kojic acid and arbutin...... 62

List of tables

Table 2.6 76 TCMs from literature with reported skin-lightening properties ...... 27 Table 3.9 NMR data for uridine (3.43) ...... 45 Table 3.10 NMR data for cytidine (3.44) ...... 46 Table 3.11 NMR data for caffeoylcholine (3.41) ...... 47 Table 3.12 NMR data for 4-caffeoylquinic acid (3.36) ...... 48 Table 3.14 NMR data for 3-caffeoylquinic acid (3.35) ...... 50 Table 3.17 NMR data for xanthiside (3.42) ...... 53 Table 3.18 NMR data for 1,3-dicaffeoylquinic acid (3.37)...... 55 Table 3.19 NMR data for 3,5-dicaffeoylquinic acid (3.38)...... 56 Table 3.21 NMR data for 1,5-dicaffeoylquinic acid (3.39)...... 58 Table 3.22 NMR data for 4,5-dicaffeoylquinic acid (3.40)...... 60

vii

List of abbreviation

℃ degree Celsius μm micrometre 1D one dimensional 2D two dimensional C18 Octadecyl bonded silica COSY correlation spectroscopy CD3OD deuterated methabol d doublet DCM dichloromethane t triblet DMSO dimethylsulfoxide DMSO-d6 deuterated DMSO g garm h hours

H2O water HMBC heteronuclear multiple-bond correlation spectroscopy HPLC high-performance liquid chromatography HSQC heteronuclear single-quantum coherence spectroscopy HTS high-throughput screening Hz Hertz

IC50 half maximal inhibitory concentration J Coupling constant LC-MS Liquid chromatography/mass spectrometry MeOH methanol MS mass spectrometry mM milimoler μM micromoler ppm parts per million TCM Traditional Chinese Medicine TFA trifluroacetic acid or trifluoroacetate Tyr tyrosinase TYRP1 tyrosinase related protein 1 UV ultraviolet

viii

Acknowledgement

I am truly appreciated to my principle supervisor Associate Professor Yun Jiang Feng and my co-supervisor Associate Professor Chris L. Brown. They provided me huge support as well as valuable guidance, which led to the completion of this thesis. Their kind encouragements make me feel confident to finish my project. Many thanks to my respectful supervisors for sharing their academic experience and wisdom with me, which will definitely benefit my future.

I am grateful to my group members, William Miaogang, Jianying Han and Duy Thanh Nguyen for their advice on lab work, Jiayu Zhang and Daozhi Ge for their suggestions on coursework, Eden Little for her generous advice on English speaking. I also want to thank all group members in Professor Ronald James Quinn’s Group for scientific knowledge in cross group meeting. I am particularly thankful my collaborator, Yaoying Lu and Associate Professor Kathryn Tonissen for their guidance on biological assay. I would like to thank Mr Alan White for the purchase of chemicals and advice on chemical investigations, Dr Wendy Loa-Kum-Cheung for instrumental technique support.

I am so appreciated to my family, especially my parents. They not only provided me financial support, but also mental support for me to chase my dream. They gave me love, trust and encouragements that helped me solve difficulties.

Last but not least, I would like to thank to my roommate Zhenya Chang for his cares on living. Many thanks to my friends with me in Australia, Hui Wang, Xing Wang, Sicong Wang, Yuan Zhang, Mingxuan Wan, Jiaye Lu, Dehui Kong, Weizhi Xu, Yucheng Tang and Min Jin. They all shared joys and sorrows with me in Australia. I am also appreciated to my friends in China, Yuting Zhang, Yan Li, Yuanda Zhang, Qianqian Xu, Yonglun Yuan for their accompanies and encouragements across distance, which made me feel not alone. Your love and friendship are significant to me, Thank you very much!

Chapter 1. Introduction and Literature Review

1.1 Hyperpigmentation Hyperpigmentation is a common skin disorder, which occurs through the excessive production of melanin in the skin, especially with sun exposure [1]. Under most circumstances, the cause of hyperpigmentation can be traced to the activity and presence of melanocytes in which melanin is produced [2]. As the human body ages, the distribution of melanocytes gradually becomes less diffuse and the regulation of its activity less controlled by the body. As ultraviolet (UV) radiation stimulates the activity of melanocytes, the concentration of these cells increases which results in hyperpigmentation [3]. Hyperpigmentation is associated with a number of skin diseases such as Addison’s disease and Cushing’s disease or some skin conditions like melasma and acne scarring [4].

Melanocytes are dendritic cells (Figure 1.1) where melanin is synthesized and secreted.[5] As shown in the figure, epidermal melanocyte is highly dendritic cell similar to neural cell. Melanocytes exhibited approximately 1.5 microns in diameter and the shapes of them are rounded, oval or even triangular. They locate in basal cell layer of epidermis and sometimes in mucous membrane as well [6]. Melanocytes function as a factory of melanin production and medium to transport melanin to skin cells. Normally, every four to ten basal keratinocytes share one melanocyte depends on different regions of skin [7]. In spite of melanin production, melanocytes can survive considerable genotoxic stress and even weaken damage induced by UV radiation. The normal production of melanin can be considered as photoprotection and thermoregulation provided by melanocytes [8]. These kinds of protection prevent further damage to histiocytic cells from UV radiation. Due to different degree of exposure to UV radiation, the distribution of melanocytes appears differently. Neck skin contains the most melanocytes, followed by arms, back, legs and chest, which illustrates that the distribution of melanocytes is consistent with the extent of UV radiation [9].

Figure 1.1 Atomic force microscopy performed on epidermal melanocytes. (A) Two- dimensional image of an epidermal melanocyte; (B) Three-dimensional image of a melanocyte. [5]

Figure 1.2 Hypopigmentation (left) verse hyperpigmentation (right) (resource from https://www.healthyskinsolutions.com/hyperpigmentation-vs-hypopigmentation/)

Depends on the amount of melanin production, there are two kinds of skin disorders (Figure 1.2). Hypopigmentation refers to condition that the production of melanin is

2 lower than normal while hyperpigmentation refers to higher melanin production than normal [1]. Normally, hormone changes in body can cause melanin deposition on skin which leads to hyperpigmentation [3]. Skin hyperpigmentation can also be caused by the deposition of iron and silver in skin tissues, such as hemochromatosis and argyrophilia [10].

A B

Figure 1.3 Symptom of three kinds of hyperpigmentation. (A. age spot; B. melasma; C. post-inflammatory hyperpigmentation.) (resource from https://edrugsearch.com/age- spot-removal/ https://russakplus.com/condition/treating-melasma/ https://www.dermnetnz.org/topics/postinflammatory-hyperpigmentation/)

There are three types of hyperpigmentation. Age spots (Figure 1.3 A) are a predominant type of skin threat among ladies over 40 years old [11]. They are flat, brown, grey or black spots on skin resulting from excess production of melanin. Age spots occasionally obscure the detection of skin cancer [12] while majority of cases believed exhibit no threat [13]. As far as cosmeticians concerned, age spots are unsightly and some people especially women choose to have them removed [14].

Melasma (Figure 1.3 B) is another type of hyperpigmentation caused by excessive sun exposure [15]. Melasma is caused by the stimulation of melanocytes, it can be treated by physical treatment like laser treatment [16]. However, it will return after small amount of sun exposure after healed [17]. Melasma is also known as mask of pregnancy due to frequent appearance on pregnant women [18]. It can be triggered by birth control pills and hormone replacement medicine [19]. While after giving birth or

3 stopping the intake of medicine, melasma can be faded. In spite of skin conditions, melasma can also be treated as primary condition of thyroid disease because excess production of melanocyte-stimulating hormone by thyroid results in melasma [20].

The third type of hyperpigmentation is post-inflammatory hyperpigmentation (Figure 1.3 C). Post-inflammatory hyperpigmentation refers to skin pigmentation after acute or chronic inflammation [21]. The depth and duration of post-inflammatory hyperpigmentation vary from skin colours [22]. Although people with darker skin tone may suffer from longer hyperpigmentation, the severity is not significantly related to the severity of inflammation [23].

1.2 Melanin Biosynthesis Under normal circumstances, melanin synthesis (Figure 1.4) starts from an essential amino acid, tyrosine. Tyrosine is firstly hydroxylized into dihydroxyphenylalanine (DOPA), a slow reaction. DOPA is then oxidised into dopaquinone fast. The first two steps of oxidations are both catalysed by an enzyme known as tyrosinase. The second oxidation is a rate-limiting step in melanin production. Dopaquinone further reacts with cysteine to form cysteinyl-DOPA which leads to the assembling of pheomelanin, red to yellow coloured pigment [24]. Without the presence of cysteine, two carbonyls in dopaquinone are reduced to hydroxy groups, meanwhile, the amine group is fused to the aromatic ring to form leucodopachrome. With the catalysation of dopachrome tautomerase, leucodopachrome is oxidised to dopachrome. Dopachrome is then catalysed by either dihydroxyindolecarboxylic acid (DHICA) oxidase or dihydroxyindole (DHI) oxidase to produce indole-5,6-quinone carboxylic acid and indole-5,6-quinone, respectively. They eventually result in the synthesis of brown eumelanin and black eumelanin [25].

Figure 1.4 Biosynthesis of melanin (Modified by author based on https://biology.stackexchange.com/questions/54294/how-does-low-cysteine- conditions-affect-pheomelanin-production) 5

Melanosomes are subcellular lysosome-like organelles and responsible for melanin production and accumulation.[26] The synthesis of eumelanin and pheomelanin take place within melanosomes by specific melanogenic enzymes through a number of signalling pathways (Figure 1.5). Among all the critical enzymes, tyrosinase and tyrosinase related protein 1 (TYRP1) are responsible for the quality and quantity of melanin synthesis whilst Promelanosome protein (Pmel17) and Melanoma Antigen Recognized by T-cells 1 (MART1) are necessary proteins for the structural maturation of melanosomes to provide places for melanin production.[27] Productions of these significant enzymes and structural proteins are induced by the Microphthalmia- associated transcription factor (MITF).[28] Furthermore, the activity of MITF is regulated by series of signalling pathways contains Cyclic adenosine monophosphate (cAMP), mitogen-activated protein kinase & extracellular signal-regulated kinase (MEK) and Wingless/Integrated signalling pathway (Wnt).[29] Specific receptors such as KIT, MCIR activate these signalling pathway upstreamly.[30] The expression of genes such as SOX10 and PAX3 are driven by MITF, which leads to the synthesis of TYRP1 and tyrosinase to catalyse melanin production.[31]

Melanin can be metabolised in two different ways. After the absorption in skin, melanin penetrates the basic membrane to enter into dermis cells. The melanophilic dermis cell can devour the metabolites and then transport them to lymphonodus. Finally, melanin metabolites can be excreted by kidney through blood circulation.[32] Melanin can also be metabolised through keratinocytes. Then the epidermal cells ascend to the stratum corneum, which excrete melanin within the aging keratinocytes [33].

Figure 1.5 Signalling pathways of series synthesis of eumelanin and pheomelanin.[34] (Modified by author based on D’Mello, P. et al.)

1.3 Tyrosinase Tyrosinase is a copper-containing enzyme with two copper ions in the active site. They are responsible for catalysation of monophenols to o-diphenols, followed by the oxidation of o-diphenols to o-quinone derivatives.[35] However, the precise catalytic mechanism on how the two copper ions interact with substrate and how the

7 monophenols are activated is still not clear. Generally, it is believed that the hydroxyl group of monophenol is deprotonated after it is bound as phenolate ions. During the deprotonation, several residues like the glutamate residue near CuA (Copper ion A) and nearby conserved asparagine residue have been proven to assist the deprotonation.[36] Tyrosinase plays the key role for the pigmentation in skin, hair and eyes.[37] Similarly, in plant, tyrosinase is significant in primary immune response and healing of wound.[38] In fungi, the enzyme provides protection during spore forming and after lesion or injury. For medical purposes, tyrosinase can be used for melanin synthesis in vitro, L-DOPA production, antibiotic lincomycin production. Additionally, tyrosinase has been utilized as a target in finding new drugs against Parkinson’s disease and various neurological diseases.[39]

Much efforts have been devoted to developing treatments against hyperpigmentation. Therapeutic options include laser surgery, cryotherapy with liquid nitrogen and dermabrasion. However, due to the epidermal trauma and even worsening in pigmentation induced by these physical treatments, more attention have been paid to safer methods with fewer side effects.[40, 41] Tyrosinase plays a significant role in melanogenesis, as it is a rate-limiting enzyme. Therefore, the inhibition of tyrosinase activity can be considered as a potential cure for hyperpigmentation.[42] According to previous research, a number of tyrosinase inhibitors were investigated naturally, semisynthetically and synthetically.[43]

1.4 Tyrosinase inhibitors from natural products Among all the inhibitors against tyrosinase, those from natural materials are considered to be more desirable as plenty of prescribed drugs currently used have been inspired or derived from natural products and they catch more attention against chemically synthetic compounds due to cosmetic demand. In addition, systemic protective strategies avoid from UV radiation have been developed by natural organisms.[44] Therefore, a variety of tyrosinase inhibitors are from nature. Huge amount of research has been devoted to searching for tyrosinase inhibitors from plant, fungal metabolites and marine algae and solving their structure-activity relationship (SAR).

1.4.1 Flavonoids Flavonoids are widely distributed in leaves, seeds, skins and flowers of plants. Almost 4000 kinds of flavonoids have been identified so far, including flavanols, flavones, flavanols, flavanones, isoflavones and anthocyanidins. As far as some plants are concerned, flavonoids and their derivatives provide protection against UV radiation, pathogens and some herbivores. The research[45] on acetone extract of Glycyrrhiza glabra showed that glabrene (1.1) and isoliquiritigenin (1.2) can inhibit the activity of tyrosinase in a dose-dependent manner, firstly indicating that isoflavenes and chalcones might be candidates for anti-pigmentation agents. Recently, calycosin (1.3)[46], an isoflavene, isolated from the root of Pueraria lobata, has exhibited potent inhibitory activity against tyrosinase with IC50 of 1.45 M. On the other hand, chalcones are a group of natural products that have been considered as precursors of flavonoids and isoflavonoids. Chalcones such as licochalcone A (1.6)[47], kuraridin (1.4)[48], kuraridinol (1.5)[48] and 2,4,2’,4’-tetrahydroxy-3-(3-methyl-2-butenyl) chalcone (TMBC) (1.7)[49] have been identified as possessing inhibitory activity against tyrosinase. Some flavonoids with the structure of 3-hydroxy-4-ketone, such as kaempferol (1.8)[50] and quercetin (1.9)[51], are able to chelate the copper in tyrosinase active site to inhibit the enzymatic activity, which leads to irreversible inhibition against tyrosinase.

Figure 1.6 Representative tyrosinase inhibitory active flavonoids

1.4.2 Organic acids According to the literatures, several compounds in tyrosinase assay such as kojic acid (1.10), gallic acid (1.11)[52] are organic acids. The secretion of organic acids from plants prevents damage from microbials and insects. For example, in order to prevent the invasion of aphids, the leaves of Rhus javanica secrete tannic acid rapidly through the veins. Tannic acid (1.17) can inhibit the oxidation of L-DOPA catalysed by tyrosinase

[53] with IC50 of 22 M. As another example, the leaves of Korean ginseng can also inhibit tyrosinase activity with the presence of p-coumaric acid (1.12), which is a ubiquitous plant metabolite.[54] Recent studies of comparisons of 17 natural tyrosinase inhibitors showed that, compared to kojic acid, p-coumaric acid as well as caffeic acid (1.14) exhibited highly effective inhibitory activity against tyrosinase tested by mushroom tyrosinase assay with 10-fold and 3-fold inhibition.[55] Geranic acids (1.15, 1.16) isolated from Cymbopogon citratus are cis-trans isomers and both isomers showed efficient inhibitory activity against tyrosinase while the related aldehyde analogues, geraniol, exhibited low inhibition.[56] Apparently, these results indicate that the carboxylic acid parts of the compounds possess significant effect on tyrosinase inhibitory activity. Natural occurring cinnamic acid (1.13) and its analogs like methoxycinnamic acid are well known for their tyrosinase inhibitory activity while their bindings to the active site of tyrosinase are reversible. Therefore, these noncompetitive inhibitors showed weak-to-moderate inhibitory activity whereas none of them so far has been stronger than kojic acid in terms of inhibitory strength against tyrosinase.

Figure 1.7 Representative reported natural organic acids with inhibitory activity against tyrosinase

1.4.3 Glycosides A range of cycloartane and cucurbitane glycosides from different leguminous and cucurbitaceous plants have been elucidated for their SAR activity against tyrosinase. The results indicated that the only differences in their skeletons are the sugar units attached and the groups attached on C-3, C-6 and C-25. As far as the inhibitory activities are concerned, the IC50 range of these compounds is from 13.95 M to 102.39 M. Askendoside B (1.18) (13.95 M) showed the most potent inhibition among these glycosides.[57] Hesperidin (1.21) isolated from citrus peel extract has been identified as non-competitive inhibitor against tyrosinase and it only brought low cytotoxicity to mouse melanoma cells.[58, 59] Asiaticoside (1.25) isolated from Centella asiatica showed dose-dependent inhibitory activity against tyrosinase. Additionally, it is reported that asiaticoside can inhibit tyrosinase in Cloudman S19 melanoma cells and B16F10 mouse melanoma cells.[60, 61] Arbutin (1.19) was used as a skin-lightening agent in the 1990s, and it has been used as the standard in tyrosinase inhibitory assay as well as kojic acid.[62] Isotachioside (1.20), as well as its glycoside derivatives (1.22 glucoside, 1.23 xyloside, 1.24 cellobioside and 1.25 maltoside) isolated from Isotachis japonica and Protea neriifolia are categorised as arbutin analogs.[63] Although arbutin

11 and isotachioside are not potent inhibitors against tyrosinase, their analog (1.26) showed better inhibitory activity. Furthermore, glucoside derivatives showed the most potent inhibition, which suggested the combination of resorcinol and glucose significantly induced the inhibitory activity.[64]

Figure 1.8 Representative natural glycoside with inhibitory activity against tyrosinase

1.4.4 Terpenes

Khan et al. have isolated 8 cycloartane triterpenoids from Amberboa ramosa and compared them in terms of the tyrosinase inhibitory activity. Out of these eight compounds, 3,21,22,23-tetrahydroxy-cycloart-24 (31), 25 (26)-diene (1.27) was found to be the most efficient inhibitor against tyrosinase, even more potent than kojic acid (16.67 M) and L-mimosine (3.68 M). [65] Another study on isolation of Atractylodis rhizoma alba showed that the compound, Selina-4 (14), 7 (11)-dien-8-one (1.28) could not inhibit free extracellular tyrosinase. Surprisingly, it inhibits tyrosinase in melanocytes, since such compounds can reduce the expression of melanin-synthesis- related enzyme like TRP-1 and TRP-2.[66] Arjunilic acid (1.32) isolated from Rhododendron collettianum has been identified as the most potent inhibitor among 9 pentacyclic triterpenoids isolated from the same plant. The result of SAR showed that the tyrosinase inhibitory activity of these kinds of triterpenoids decreases as the configuration of the hydroxy group on C-2 changes from −configuration to −configuration.[67]

Figure 1.9 Representative terpenes reported as tyrosinase inhibitors. (Compound 1.29: maslinic acid; Compound 1.30: 2 , 23-trihydroxyolean-12-en-28-oic acid; Compound 1.31: bayogenin; Compound 1.32: arjunilic acid; Compound 1.33: methyl arjunolate; Compound 1.34: arjungenin; Compound 1.35: 3, 23, 24-trihyfroxyolean- 12-en-28-oic acid.)

1.4.5 Esters Khatib et al. isolated 3-(2,4-dihydroxyphenyl) propionic acid (1.36) (DPPA) from the plant Ficus carica and transformed it into esters. SAR on tyrosinase inhibition showed that one of the esters, DPPA isopropyl ester, possesses effective inhibitory activity against tyrosinase.[68] Two compounds, methyl 2(2S)-hydroxyl-7(E)-tritriacontenoate (1.39) and methyl 2(2S)-O--D-galactopyranosyl-7(E)-tetratriacontenoate (1.40), 14 were isolated from Amberboa ramose, and both showed strong inhibition against

[69] tyrosinase with IC50 of 1.36 M and 11.68 M respectively. Other research on tyrosinase inhibitory activity of 3,5-dihydroxyphenyl decanoate (1.37) suggested that with the increasing concentration of 3,5-dihydroxyphenyl decanoate at low range of 5- 60 M, the enzymatic activity decreases slowly. At high concentration, the compound strongly inhibits tyrosinase activity irreversibly, which indicated that 3,5- dihydroxyphenyl decanoate can inactivate tyrosinase. On the contrary, its analog, nonyl-3,5-dihydroxyphenyl decanoate (1.38) did not show any inhibition against tyrosinase.[70] Methyl p-hydroxybenzoate (1.42)and p-hydroxybenzoic acid (1.41) isolated from the branches of Ficus erecta var. sieboldii King have been identified as

[71] comparable tyrosinase inhibitors compared to arbutin (IC50= 0.32 mM).

Figure 1.10 Representative esters as natural tyrosinase inhibitors

1.4.6 Other compounds Trifolirhizin (1.43) isolated from Sophora flavescens has been identified as a noncompetitive inhibitor against tyrosinase. It inhibited the production of melanin 15 obviously in B16 melanoma cells.[48] Inhibitory activity of boldine (1.44) against tyrosinase is reversible and dose-dependent. According to the results of molecular docking and dynamics simulation, boldine can interact with several residues near the binding site of tyrosinase, which led to some possibility of alkaloids inhibiting tyrosinase activity.[72] Paeonol (1.45) as well as resveratrol (1.46) was a mixed type tyrosinase inhibitor as it can bind to the copper ion at the active site.[73, 74]

Figure 1.11 Other compounds reported as natural tyrosinase inhibitors

In conclusion, only a few natural tyrosinase inhibitors are commercially available on market as depigmenting agents. Additionally, a number of natural tyrosinase inhibitors such as flavonoid and terpenes have been identified that affect many other targets,[75] which are not specific enough to treat hyperpigmentation. Apart from that, some of the inhibitors like kojic acid have been proved to have high toxicity towards cells.[76] The development of depigmenting agents continues to be the subjects of extensive research due to the vast clinical results.

1.5 Traditional Chinese Medicine (TCM) TCM has provided a large number of natural products, which have been applied for over 5000 years. The long-term practical use, unique theory and technology of TCM attracts a number of researchers.[77] Although there are few records of TCM being used for hyperpigmentation, the application of TCM as cosmetics to lighten skin colour has

16 been developed for centuries. According to some ancient works, ancient Chinese practitioners mastered the technologies to manufacture simple skin-lightening agents. The medical records from the Warring State Period to the Late Qing Dynasty contained 269 traditional Chinese herbal medicines and 1233 cosmetic prescriptions, most of which mainly were for the nobility of the imperial court.[78] It was recorded on shells that TCM practitioners had applied herbs in treating pigment-related diseases like melanin spots from Ancient Times to Pre-Qin Period, which indicates that TCM has had more than 2000 years of application for skin-lightening purposes. “ Shen Nong’s Herbal Classic”, the earliest book on materia medica in the world, contained more than 100 kinds of cosmetical medica that possessed cosmetic applications including skin- lightening properties.[79]

A B

C D

Figure 1.12 Representative TCM traditionally used for skin-lightening purpose. (A. Atractylodes macrocephala; B. Angelica dahurica; C. Bletilla striata; D. Ampelopsis japonica. Resource from http://www.yobo360.com/shop530/product-2893.aspx)

Based on ancient TCM records, four herbal TCMs were frequently used for skin- lightening and freckle elimination purposes. The rhizome of white Atractylodes Macrocephala was recoded as part of a prescription to lighten skin colour. From the book “Drug Properties”, melanin spots could be cured by applying Atractylodes Macrocephala in the form of a highly concentrated liquor on affected parts. Secondly, angelica dahurica was another TCM commonly applied as cosmetics in ancient China. It possessed a pleasant smell which was considered to be one of the best cosmetic products. According to “Qian Jin Mian Zhi Fang” and “Yu Rong San”, it was the basis to 17 make ancient anti-aging facial creams. The third TCM was Bletilla Striata which mainly contains starch, volatile oil and mucoid substance. Due to its sticky property, the addition of its extract into cosmetics can lubricate skin. It was written in “Drug Properties” that Bletilla Striata had been used for treatment of facial sores. Furthermore, from “Compendium of Materia Medica”, Bletilla Striata was used for lightening skin colour and even treatment for chapped hands and feet. Finally, Ampelopsis Japonica was recorded as being able to treat hyperpigmentation when ground with albumen.

1.6 Project aims There are high demands for skin-lightening and anti-pigmentation agents from natural materials. TCM is a rich resource of natural products, which has little side effects reported over thousands of years. This project aims at discovering tyrosinase inhibitors from TCM. More specifically, I will: 1) Undertake selection for TCM with known skin-lightening properties. 2) Establish in vitro tyrosinase inhibitory assay; 3) Make extracts of TCM materials; 4) Screen TCM extracts for tyrosinase inhibitory activity; 5) Isolate tyrosinase inhibitors from active TCM extracts

References 1. Nicolaidou, E. and A.D. Katsambas, Pigmentation disorders: hyperpigmentation and hypopigmentation. 2014. 32(1): p. 66-72. 2. Quevedo Jr, T.B. Fitzpatrick, and K. Jimbow, Human skin color: origin, variation and significance. 1985. 14(1): p. 43-56. 3. Ortonne, J.-P. and D.L. Bissett. Latest insights into skin hyperpigmentation. in Journal of Investigative Dermatology Symposium Proceedings. 2008. Elsevier. 4. Yadav, M., et al., Familial progressive hyperpigmentation: a case report. 2012. 5. Zhang, R.z., et al., Morphology of cultured human epidermal melanocytes observed by atomic force microscopy. 2004. 17(1): p. 62-65. 6. Gauthier, Y., et al., Melanocyte detachment after skin friction in non lesional skin of patients with generalized vitiligo. 2003. 148(1): p. 95-101. 7. Mulekar, S., Melanocyte–keratinocyte cell transplantation for stable vitiligo. 2003. 42(2): p. 132-136. 8. Sarna, T., et al., Loss of melanin from human RPE with aging: possible role of melanin photooxidation. 2003. 76(1): p. 89-98. 9. Robins, A.H., Biological perspectives on human pigmentation. Vol. 7. 2005: Cambridge University Press. 18

10. Stains, C., Special stains in interpretation of liver biopsies. 2010: p. 92. 11. Choi, W., et al., Molecular and histological characterization of age spots. 2017. 26(3): p. 242-248. 12. Queen, L., Skin Cancer: Causes, Prevention, and Treatment. 2017. 13. Tobin, D.J.J.J.o.t.v., Introduction to skin aging. 2017. 26(1): p. 37-46. 14. Shah, F., et al., A Prospective Study of Knowledge and Perception towards the Efficacy of Anti-Aging Cosmetics among Female Population of Ras Al Khaimah, UAE. 2017. 7(03): p. 275. 15. Park, K.C. and I. Kim, Pathogenesis of Melasma, in Melasma and Vitiligo in Brown Skin. 2017, Springer. p. 21-31. 16. Zhang, Q., et al., Clinical staging of melasma. 2018. 24(4): p. 274-278. 17. Zhu, L., et al., Combination therapy for melasma. 2016. 49(2): p. 147-150. 18. Moin, A., Z. Jabery, and N. Fallah, Prevalence and awareness of melasma during pregnancy. 2006. 45(3): p. 285-288. 19. Sonthalia, S.J.M.A.M.N.D.J., Etiopathogenesis of melasma. 2015: p. 6-14. 20. Yazdanfar, A. and B.J.I.J.o.D. Hashemi, Association of melasma with thyroid autoimmunity: a case-control study. 2010. 13(2): p. 51-53. 21. Callender, V.D., et al., Postinflammatory hyperpigmentation. 2011. 12(2): p. 87-99. 22. Davis, E.C., V.D.J.T.J.o.c. Callender, and a. dermatology, Postinflammatory hyperpigmentation: a review of the epidemiology, clinical features, and treatment options in skin of color. 2010. 3(7): p. 20. 23. Chan, H.H., et al., The prevalence and risk factors of post‐inflammatory hyperpigmentation after fractional resurfacing in Asians. 2007. 39(5): p. 381-385. 24. Le Pape, E., et al., Regulation of eumelanin/pheomelanin synthesis and visible pigmentation in melanocytes by ligands of the melanocortin 1 receptor. 2008. 21(4): p. 477-486. 25. Hearing, V.J.J.J.I.D., Determination of melanin synthetic pathways. 2011. 131(E1): p. E8-E11. 26. Marks, M.S. and M.C.J.N.r.M.c.b. Seabra, The melanosome: membrane dynamics in black and white. 2001. 2(10): p. 738. 27. Yamaguchi, Y., M. Brenner, and V.J.J.J.o.b.c. Hearing, The regulation of skin pigmentation. 2007. 282(38): p. 27557-27561. 28. Bae, I.H., et al., Mannosylerythritol lipids inhibit melanogenesis via suppressing ERK‐CREB‐MiTF‐tyrosinase signalling in normal human melanocytes and a three‐dimensional human skin equivalent. 2018. 29. Su, T.-R., et al., Inhibition of melanogenesis by gallic acid: Possible involvement of the PI3K/Akt, MEK/ERK and Wnt/β-catenin signaling pathways in B16F10 cells. 2013. 14(10): p. 20443-20458. 30. Rosenblum, E.B., H.E. Hoekstra, and M.W.J.E. Nachman, Adaptive reptile color variation and the evolution of the MCIR gene. 2004. 58(8): p. 1794- 1808. 31. Wan, P., et al., Regulation of melanocyte pivotal transcription factor MITF by some other transcription factors. 2011. 354(1-2): p. 241-246. 32. Wu, Q., et al., Evidence for chronic kidney disease-mineral and bone disorder associated with metabolic pathway changes. 2015. 94(32). 33. Slominski, A., et al., Melanin pigmentation in mammalian skin and its hormonal regulation. 2004. 84(4): p. 1155-1228. 34. D’Mello, S., et al., Signaling pathways in melanogenesis. 2016. 17(7): p. 1144.

35. Ramsden, C.A., P.A.J.B. Riley, and m. chemistry, Tyrosinase: The four oxidation states of the active site and their relevance to enzymatic activation, oxidation and inactivation. 2014. 22(8): p. 2388-2395. 36. Kanteev, M., M. Goldfeder, and A.J.P.S. Fishman, Structure–function correlations in tyrosinases. 2015. 24(9): p. 1360-1369. 37. Ando, H., et al., Approaches to identify inhibitors of melanin biosynthesis via the quality control of tyrosinase. 2007. 127(4): p. 751-761. 38. Bonesi, M., et al., Advances in the tyrosinase inhibitors from plant source. 2018. 39. Valipour, E., B. Arikan, Biotechnology, and F. Sciences, Characteriszation of Tyrosinase Enzyme From Native Bacillius Megaterium SP. Strain M36. 2016. 5(5): p. 465. 40. Passeron, T., et al., Laser treatment of hyperpigmented lesions: position statement of the European Society of Laser in Dermatology. 2019. 41. Hirobe, T. and H.J.D.s. Enami, Melanoblasts but not melanocytes decrease in number in human epidermis of idiopathic guttate hypomelanosis. 2018. 36(3): p. 131-135. 42. Kim, Y.-J., H.J.C. Uyama, and m.l.s. CMLS, Tyrosinase inhibitors from natural and synthetic sources: structure, inhibition mechanism and perspective for the future. 2005. 62(15): p. 1707-1723. 43. Ullah, S., et al., Tyrosinase inhibitors: a patent review (2011-2015). 2016. 26(3): p. 347-362. 44. Saewan, N. and A.J.J.o.c.d. Jimtaisong, Natural products as photoprotection. 2015. 14(1): p. 47-63. 45. Nerya, O., et al., Glabrene and isoliquiritigenin as tyrosinase inhibitors from licorice roots. 2003. 51(5): p. 1201-1207. 46. Wagle, A., et al., Identifying an isoflavone from the root of Pueraria lobata as a potent tyrosinase inhibitor. 2019. 276: p. 383-389. 47. Fu, B., et al., Isolation and identification of flavonoids in licorice and a study of their inhibitory effects on tyrosinase. 2005. 53(19): p. 7408-7414. 48. Hyun, S.K., et al., Inhibitory effects of kurarinol, kuraridinol, and trifolirhizin from Sophora flavescens on tyrosinase and melanin synthesis. 2008. 31(1): p. 154-158. 49. Zhang, X., et al., Inhibitory effect of 2, 4, 2′, 4′-tetrahydroxy-3-(3-methyl-2- butenyl)-chalcone on tyrosinase activity and melanin biosynthesis. 2009. 32(1): p. 86-90. 50. Rho, H.S., et al., Kaempferol and kaempferol rhamnosides with depigmenting and anti-inflammatory properties. 2011. 16(4): p. 3338-3344. 51. Fan, M., et al., Quercetin as a tyrosinase inhibitor: inhibitory activity, conformational change and mechanism. 2017. 100: p. 226-233. 52. Fatiha, B., et al., Phenolic composition, in vitro antioxidant effects and tyrosinase inhibitory activity of three Algerian Mentha species: M. spicata (L.), M. pulegium (L.) and M. rotundifolia (L.) Huds (Lamiaceae). 2015. 74: p. 722- 730. 53. Kubo, I., et al., Tyrosinase inhibitors from galls of Rhus javanica leaves and their effects on insects. 2003. 58(9-10): p. 719-725. 54. An, S.M., J.S. Koh, and Y.C.J.P.R. Boo, p‐coumaric acid not only inhibits human tyrosinase activity in vitro but also melanogenesis in cells exposed to UVB. 2010. 24(8): p. 1175-1180. 55. Okombi, S., et al., Analogues of N-hydroxycinnamoylphenalkylamides as inhibitors of human melanocyte-tyrosinase. 2006. 16(8): p. 2252-2255. 20

56. Masuda, T., et al., Identification of geranic acid, a tyrosinase inhibitor in lemongrass (Cymbopogon citratus). 2007. 56(2): p. 597-601. 57. Khan, M.T.H., et al., Tyrosinase inhibition studies of cycloartane and cucurbitane glycosides and their structure–activity relationships. 2006. 14(17): p. 6085-6088. 58. Man, M.-Q., et al., Benefits of Hesperidin for Cutaneous Functions. 2019. 2019. 59. Zhang, C., et al., Tyrosinase inhibitory effects and inhibition mechanisms of nobiletin and hesperidin from citrus peel crude extracts. 2007. 22(1): p. 91- 98. 60. Kwon, K.J., et al., Asiaticoside, a component of Centella asiatica, inhibits melanogenesis in B16F10 mouse melanoma. 2014. 10(1): p. 503-507. 61. LU, Y., D. WU, and D.-g.J.C.J.o.A.M. WANG, Therapeutic effects of new Kligman formula based on the depigmention agent asiaticoside on melasma [J]. 2008. 1. 62. Hori, I., et al., Structural criteria for depigmenting mechanism of arbutin. 2004. 18(6): p. 475-479. 63. Matsumoto, T., et al., Chemical synthesis and tyrosinase-inhibitory activity of isotachioside and its related glycosides. 2018. 465: p. 22-28. 64. Ishioka, W., et al., Resorcinol alkyl glucosides as potent tyrosinase inhibitors. 2019. 29(2): p. 313-316. 65. Khan, M.T.H., et al., Tyrosinase inhibitory cycloartane type triterpenoids from the methanol extract of the whole plant of Amberboa ramosa Jafri and their structure–activity relationship. 2006. 14(4): p. 938-943. 66. Chang, Y.-H., et al., Inhibition of melanogenesis by selina-4 (14), 7 (11)-dien-8- one isolated from Atractylodis Rhizoma Alba. 2007. 30(4): p. 719-723. 67. Ullah, F., et al., Tyrosinase inhibitory pentacyclic triterpenes and analgesic and spasmolytic activities of methanol extracts of Rhododendron collettianum. 2007. 21(11): p. 1076-1081. 68. Khatib, S., et al., Enhanced substituted resorcinol hydrophobicity augments tyrosinase inhibition potency. 2007. 50(11): p. 2676-2681. 69. Khan, S.B., et al., Tyrosinase-inhibitory long-chain esters from Amberboa ramosa. 2005. 53(1): p. 86-89. 70. Qiu, L., et al., Irreversibly inhibitory kinetics of 3, 5-dihydroxyphenyl decanoate on mushroom (Agaricus bisporus) tyrosinase. 2005. 13(22): p. 6206-6211. 71. Park, S.H., et al., Constituents with tyrosinase inhibitory activities from branches of Ficus erecta var. sieboldii King. 2012. 27(3): p. 390-394. 72. Si, Y.-X., et al., Effects of boldine on tyrosinase: Inhibition kinetics and computational simulation. 2013. 48(1): p. 152-161. 73. Peng, L.-H., et al., Inhibitory effects of salidroside and paeonol on tyrosinase activity and melanin synthesis in mouse B16F10 melanoma cells and ultraviolet B-induced pigmentation in guinea pig skin. 2013. 20(12): p. 1082- 1087. 74. Uesugi, D., et al., Synthesis, oxygen radical absorbance capacity, and tyrosinase inhibitory activity of glycosides of resveratrol, pterostilbene, and pinostilbene. 2017. 81(2): p. 226-230. 75. Panche, A., A. Diwan, and S.J.J.o.n.s. Chandra, Flavonoids: an overview. 2016. 5. 76. Aytemir, M., et al., Kojic acid-derived mannich bases with biological effect. 2018, Google Patents.

77. Tang, W. and G. Eisenbrand, Chinese drugs of plant origin: chemistry, pharmacology, and use in traditional and modern medicine. 2013: Springer Science & Business Media. 78. Wang, K.-H., et al., Cosmetic applications of selected traditional Chinese herbal medicines. 2006. 106(3): p. 353-359. 79. Zhang, M. and S. Chen, Review of current situations and prospects of research and development on traditional Chinese medicine cosmetics. 2007. 32(23): p. 2457-2460.

Chapter 2. Tyrosinase assay and the screening of TCM extracts

against tyrosinase

This chapter describes the initial screening of TCM extracts for tyrosinase inhibitory activity using a mushroom tyrosinase assay[1]. In TCM terminology, there are no specific descriptions for skin hyperpigmentation disorders, skin-pigmentation-related terms such as ‘black spots’, ‘age spots’ and ‘skin care’ are generally used for TCM selection in the Chinese pharmacy. Of course, these terms are related to skin-hyperpigmentation issues, that leads to the assumption which materials have activity against these conditions have therapeutic effects on skin-hyperpigmentation.

2,1 Mushroom tyrosinase Tyrosinase is a rate-limiting enzyme for melanin production[2]. Therefore, the assay was established by using tyrosine as substrate and potassium phosphate solution (pH 6.8) as buffer as reported previously. Kojic acid (1.19) and arbutin (1.10) have been reported to be effective tyrosinase inhibitors and were used as positive controls in the experiments.

Mushroom (Agaricus bisporus) tyrosinase was used in the initial assay given it is low cost and is readily available from commercial suppliers. Human tyrosinase, by contrast, is prohibitively expensive (A$139/1000unit/mg). To date no common tyrosinase protein structure occurring across all species has been found[3]. To justify the use of mushroom tyrosinase in the initial TCM screening we undertook some straightforward bioinformatic analysis in order to determine the degree of structural similarity between the two proteins, initially using sequence alignment and binding studies. There are 529 amino acids in the primary sequence of the 60 kDa human tyrosinase[4], whilst there are 576 amino acids in the 66 kDa mushroom tyrosinase[5]. To date, the solid structure of human tyrosinase has not been reported, which has limited the understanding of the

23 specific modes of interaction with substrates. In an attempt to gain some insight in the structure of the human protein, particularly around the active site of the enzyme, the crystal structure of human tyrosinase related protein 1 (TYRP1) was utilized to study the similarity between human and mushroom tyrosinase with the assistant of SwissModel online resource.[6] TYRP1 is one of three tyrosinase -like glycoenzymes in human melanocytes with 537 amino acids for 60 kDa. According to the sequence alignment (Figure 2.1) between human tyrosinase (Code: P14679) and TYRP1 (Code: P17643), the sequence identity was determined to be 0.44 (44% sequence similarity, over 30%) by adding protein sequences using the default method, which indicated that TYRP1 was appropriate as the homologous model of human tyrosinase.[7] In addition, the crystal structure of TYRP1 is available online from http://www.rcsb.org/ with code 5M8Q. In the construct, the active site copper ions were replaced with zinc ions rendering the TYRP1 protein inactive in melanin production.[7]

Figure 2.1 Sequence alignment of human tyrosinase (Model_01 P14679)and TYRP1 (5m8q.2.A P17643) The sequences highlighted indicate the similarity.

The sequence identity was only 0.11 (11% similarity) by adding protein sequences using the default settings between mushroom tyrosinase (Code: C7FF04) and TYRP1 (Code: P17643). However, analysis of the sequence alignment suggests the copper binding sites of the two proteins were highly similar (32% similar, using partial sequences alignment on SwissModel with default settings), which was selected in Figure 2.2. the copper ions of both proteins are localised by several proximal histidine residues. The preserved

24 binding of copper ions across the two proteins indicates the similarity proximal coordinating ligands between mushroom and human tyrosinase and a similar active site.

Figure 2.2 Sequence alignment of mushroom tyrosinase (C7FF04) and TYRP1(P17643). * represent the same amino acids, ﹒represent the similar amino acid, metal binding sites were highlight in purple.

The 3-D models of the copper binding sites from mushroom tyrosinase (PDB code: 5M6B)[8] and TYRP1[7] (PDB code: 5M8Q) were derived from X-ray analysis collected from online protein databank. Figure 2.3 A showed the copper binding site of TYRP1, Figure 2.3 B gave a 3-D view of copper binding site of mushroom tyrosinase when Kojic acid (a known tyrosinase inhibitor) inhibited the enzymatic activity. The sequences highlighted in red are histidine residues, which coordinate the copper ions. The 3-D models showed similar attachments of histidine residues between mushroom tyrosinase and TYRP1, which suggests that the using of mushroom tyrosinase in initial screening was feasible.

A B Figure 2.3 The metal binding sites of TYRP1 (A) and mushroom tyrosinase (B), the red sequences in both figures represent the histidine residues.

2.2 Positive controls To validate the assay, two positive controls were used. Dose-response curves were constructed using 11 different concentrations. For arbutin, 2.0, 1.0, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2 and 0.1 mM respectively were used. As shown in Figure 2.4, arbutin showed dose dependent inhibition against tyrosinase with an IC50 value of 0.79 mM.

Figure 2.4 The inhibition of arbutin against tyrosinase. Data was presented as mean ± SEM

For kojic acid, 11 different concentrations ranging from 3.5 to70 μM were used (3.5, 7.0, 10.6, 14.1, 21.1, 28.1, 35.2, 42.2, 49.3, 56.3, 56.3, 63.3 and 70.4 μM respectively). As shown in Figure 2.5, kojic acid also showed dose dependent inhibition against tyrosinase with an IC50 value of 35.55μM. Both IC50 values were comparable with those reported in

[9, 10] the literature (IC50(kojic acid)= 30.61 μM, IC50(arbutin)= 2.10 mM), which indicated the tyrosinase assay worked well.

Figure 2.5 The inhibition of kojic acid against tyrosinase. Data was presented as mean ± SEM

2.3 Selection of TCM A comprehensive literature review on Chinese Pharmacopoeia and other Chinese references [10-16] identified 76 TCMs which have been traditionally used for skin care products such as skin-lightening agents, anti-aging agents, whitening creams and moistures. They were selected for the project. Out of 76 TCMs, only 42 samples were available from local Chinese pharmacy, which were the subject of the current study. Table 2.6 summarised the 76 TCM samples, including English name, Chinese name, reported properties and applications. The table also indicated whether these herbs were available from the local Chinese pharmacy. The TCM materials that were selected for the following screening were indicated by availability.

Table 2.6 76 TCMs from literature with reported skin-lightening properties

Chinese Reported properties Medicinal ID English/Latin Availability name and applications part Anti-aging cosmetics[11], Root 1 Angelica dahurica 白芷 whitening cream[12] In vitro tyrosinase Fruit 2 Tribulus terrestris 白蒺藜 V inhibitory activity[13] 3 Paeonia lactiflora Pall. 白芍 Anti-aging products[14] V Root In vitro tyrosinase Fungus 4 Poria cocos 茯苓 V inhibitory activity[13] Stem and 5 Bletilla striata 白芨 Anti-aging cosmetics[11] leaf In vitro tyrosinase Root 6 Ampelopsis Radix 白蔹 inhibitory activity[15], V whitening cream[12] 27

7 Cortex dictamni 白鲜皮 Whitening cream[12] V Root bark Anti-aging cosmetics[11], in vitro Stem 8 Typhonium giganteum 白附子 tyrosinase inhibitory activity[15] 9 Lithargyrum 密陀僧 Anti-aging products[14] Lead 10 Paeonia lactiflora Pall. 芍药 Anti-aging products[14] Root 11 Amygdalus Communis Vas 杏仁 Anti-aging cosmetics[11] V Seed Root and 12 Saposhnikovia divaricata 防风 Anti-aging products[14] V stem In vitro tyrosinase Root 13 Asarum sieboldii Miq. 细辛 inhibitory activity[13] Skin care, analgesic Stem 14 Ligusticum chuanxiong hort 川芎 effect[16], anti-aging V cosmetics[11] 15 Ligusticum sinense Oliv. 藁本 Anti-aging products[14] V Stem 16 Bombyx mori Linnaeus 僵蚕 Anti-aging products[14] V Larva Whole 17 Lysimachia foenum-graecum 零陵香 Anti-aging products[14] plant Stem and 18 Atractylodes macrocephala 白术 Anti-aging cosmetics[11] root Animal 19 Moschus berezovskii Flerov. 麝香 Anti-aging products[14] secretion Moisture and Stem 20 Angelica sinensis 当归 V whitening[13] 21 Phytolacca acinosa Roxb 商陆 Anti-aging cosmetics[11] V Root 22 Trichosanthes cucumeroides 土瓜根 Anti-aging products[14] Root 23 Prunus persica 桃仁 Anti-aging cosmetics[11] Seed 24 Benincasa hispida 冬瓜仁 Anti-aging cosmetics[11] Seed 25 lily magnolia;magnolia wood 辛夷 Anti-aging cosmetics[11] V Flower 26 Magnolia 木兰皮 Whitening products[14] Root skin 27 Colophony 松香 Whitening products[14] Resin Stem and 28 Agastache rugosa 藿香 Whitening products[14] leaf 29 Aristolochia debilis 青木香 Whitening products[14] Root Root and 30 Nardostachys jatamansi DC 甘松 Whitening products[14] stem 31 Syzygium aromaticum 丁香 Whitening products[14] Flower Animal 32 Eagle white 鹰屎白 Whitening products[14] secretion 33 Cuscuta chinensis Lam. 菟丝子 Anti-aging cosmetics[11] V Leaf In vivo tyrosinase Root 34 Scutellaria baicalensis Georgi 黄芩 V inhibitory activity[17] 35 Astragalus membranaceus 黄芪 Whitening products[14] V Root Anti-aging cosmetics[11], Root 36 Salvia miltiorrhiza Bge. 丹参 treatment of facial skin V diseases[13] 37 Lithospermum erythrorhizon 紫草 Whitening products[14] V Leaf 38 Carthamus tinctorius L. 红花 Whitening products[14] V Flower Root and 39 Panax notoginseng 三七 Whitening products[14] V stem 28

Whitening and freckle Leaf juice 40 Aloe vera 芦荟 V removing cosmetics[16] Whitening and freckle Leaf 41 Glycyrrhiza uralensis Fisch 甘草 V removing cosmetics[16] 42 Morus alba L. 桑白皮 Whitening products[13] Root 43 Rhus chinensis Mill. 五倍子 Skin care products[14] V Gall 44 Cicada Slough 蝉衣 Whitening products[14] V Pupa 45 Paeonia 牡丹皮 Whitening products[14] V Root skin 46 Sophora 苦参 Whitening products[14] V Root 47 Xanthium strumarium L. 苍耳子 Skin care products[14] V Fruit Root and 48 Rheum palmatum L．大黄 Anti-aging cosmetics[11] V stem Whole 49 Serissa japonica 六月雪 Skin care products[14] plant 50 Lycium chinense Mill 地骨皮 Skin care products[14] V Root skin 51 Dipsacus asperoides 炒川断 Skin care products[14] V Root Root and 52 Gastrodia elata Bl. 天麻 Skin care products[14] V stem 53 Forsythia suspensa 连翘 Skin care products[14] V Fruit 54 Acer ginnala Maxim. 茶条槭 Skin care products[14] Leaf 55 Thlaspi arvense Linn 败酱草 Skin care products[14] V Leaf Root and 56 Uxus sinica 黄杨 Skin care products[14] leaf 57 Lonicera japonica 金银花 Skin care products[14] Flower 58 Platycodon grandiflorus 桔梗 Skin care products[14] V Root Whole 59 Tropaeolum majus. 旱金莲 Whitening products[14] plant Alisma plantago- Stem 60 泽泻 Whitening products[14] aquatica Linn. Root and 61 Polygonatum odoratum 玉竹 Whitening products[14] V stem 62 Bambusa tuldoides Munro 竹茹 Whitening products[14] Stem In vitro tyrosinase Root and 63 Rhodiola rosea L. 红景天 inhibitory activity[16] stem Whitening and freckle Fruit 64 Ginkgo biloba L. 银杏 removing cosmetics[16] 65 Chaenomeles sinensis 木瓜 Whitening products[13] V Fruit 66 Acanthopanar gracilistμlus 五加皮 Whitening products[13] V Root in vitro tyrosinase Root 67 Rehmannia glutinosa 熟地 V inhibitory activity[13] In vitro tyrosinase Root and 68 Atractylodes Lancea 苍术 V inhibitory activity[13] stem In vitro tyrosinase Root 69 Salutem Phytolaccae 生商陆 inhibitory activity[13] In vitro tyrosinase Stem and 70 Cinnamomum cassia Presl 桂枝 V inhibitory activity[13] leaf 71 Glehnia littoralis 北沙参 Whitening products[14] V Root 72 Fructus piperis longi 荜 Whitening products[14] Fruit

In vivo tyrosinase Root and 73 Polygonum cuspidatum 虎杖 V inhibitory activity[17] stem In vivo tyrosinase Root 74 Sanguisorba officinalis L. 地榆 inhibitory activity[17] Whitening and freckle Root 75 Radix Cynanchi Auriculati 白首乌 removing cosmetics[16] In vitro tyrosinase Resin 76 Schinus terebinthifolius Raddi 肖乳香 inhibitory activity[13]

2.4 Anti-tyrosinase activity 1 g of each TCM was extracted by ethanol to provide the crude extracts for testing. Ethanol was chosen in consideration of traditional TCM preparation in which is soaking herbal material in alcohol or boiling water. The crude extracts were tested against tyrosinase at the concentration of 2 mg/mL. The activity against tyrosinase were evaluated and the results were shown in Figure 2.7.

73 71 70 68 67 66 65 61 58 55 53 52 51 50 48 47 46 45 44 43 41 40 39

38 Sample ID Sample 37 36 35 34 33 25 21 20 16 15 14 12 11 7 6 4 3 2 - + 0 20 40 60 80 100 120 inhibition/% Figure 2.7 HTS tyrosinase assay results (all extracts including controls are under concentration of 1 mg/mL, values present mean+SEM)

At the concentration of 1 mg/mL, the positive control, kojic acid, showed almost 100% inhibition against tyrosinase. The negative control, phosphate buffer with 1% DMSO, showed approximately 0.15% inhibition against tyrosinase. At the concentration of 1 mg/mL, 7 out of 42 TCM extracts (17%) showed little inhibition against tyrosinase. 17 showed over 60% of inhibition against tyrosinase accounting for 40%, 5 had over 80% 31 inhibition against tyrosinase, which accounted for 11% of all TCM. One of the TCM extracts, Xanthium strumarium L., showed approximately 81% inhibition at 1 mg/mL. The plant material was readily available and was chosen for further isolation of pure compounds. The structures of the compounds were determined by NMR and MS spectroscopic data. Their IC50 values were determined using tyrosinase assay.

2.5 Experimental Protein sequence alignments The full length protein coding sequence of mushroom tyrosinase (C7FF04), human tyrosinase (P14679) and TYRP1(P17643) were obtained from the online database www.uniprot.org and then import into the online structure prediction application on the server, SwissModel.

Chemicals The mushroom tyrosinase inhibitor kit (kojic acid, L-tyrosine, mushroom tyrosinase, potassium phosphate, potassium hydroxide) was purchased from Sigma-Aldrich, Castle Hill, New South Wales, Australia.

Tyrosinase assay The spectrophotometric tyrosinase assay was setup based on a previous reported method with slight modifications.[18] Stock solution of 0.1 M potassium phosphate buffer under pH 6.8 was readily made for the assay. 0.625 mM L-tyrosine in phosphate buffer and 217 U/mL mushroom tyrosinase in phosphate buffer were freshly prepared 30min before using. For testing two positive controls, different concentrations of arbutin (2.0, 1.0, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2 and 0.1 mM respectively) and kojic acid (3.5, 7.0, 10.6, 14.1, 21.1, 28.1, 35.2, 42.2, 49.3, 56.3, 56.3, 63.3 and 70.4 μM respectively) were freshly prepared using serial dilution of 4mM arbutin and 140.8 μM kojic acid stock solutions. Briefly, 80 μL L-tyrosine solution, 100 μL arbutin solution were mixed in a 96- well plate. Then, 20 μL mushroom tyrosinase solution was added to each mixture, which was then incubated for 35 min under 37 °C. After the incubation, the absorbance of the mixture was determined under 475 nm UV wavelength using microplate reader (Multiskan MK3, Thermo, USA) . The mixture of 80μL L-tyrosine solution and 120 μL

32 phosphate buffer was used as negative control for both testings. The inhibition percentage was calculated using the following equation.

퐴1−퐴2 Inhibition%=(1 − ) × 100 퐴3−퐴4 In this equation, A1 is the absorbance of the sample with tyrosinase, A2 is the absorbance of the sample without tyrosinase, A3 is the absorbance of tyrosinase without sample and A4 is the absorbance of negative control. The 50% inhibitory concentration (IC50) values were calculated by building the nonlinear regression curve with GraphPad Prism 7.0b (GraphPad software, La Jolla, California, USA)

For testing TCM extracts, 2 mg/mL solution of each TCM extract was prepared with 1% of DMSO buffer solution. 1% of DMSO was also the up-limit tolerance of microplate reader against DMSO. Briefly, 100 μL TCM solution and 80μL L-tyrosine solution were added into a 96-well plate and then 20 μL tyrosinase was added finally. The plate was then incubated for 35 min under 37 °C. After the incubation, the absorbance of the mixture was determined under 475 nm UV wavelength using microplate reader as well. 1 mg/mL kojic acid was used as the positive control in the screening, 1% DMSO phosphate buffer was used as negative control. The percentage inhibition was calculated as described previously.

Crude extracts preparations 42 TCM samples were purchased from Hong Ren Tang Herbal, Sunnybank Plaza, Queensland, Australia. 1 g of dried TCM samples were ground and extracted with ethanol (200 mL) at room temperature for 3 h. Then the solution was filtered, and the TCM residue was extracted 2 more times using the same methods. The 3 extracts were combined using rotary evaporator to give the crude extracts.

References 1. Bonesi, M., et al., Advances in the tyrosinase inhibitors from plant source. 2018. 2. Hearing, V.J.J.J.I.D., Determination of melanin synthetic pathways. 2011. 131(E1): p. E8-E11. 3. Jaenicke, E. and H.J.B.J. Decker, Tyrosinases from crustaceans form hexamers. 2003. 371(2): p. 515-523. 4. Teramae, A., et al., The Molecular Basis of Chemical Chaperone Therapy for Oculocutaneous Albinism Type 1A. 2019. 139(5): p. 1143-1149. 33

5. Li, N.-Y., et al., Molecular cloning and expression of polyphenoloxidase genes from the mushroom, Agaricus bisporus. 2011. 10(2): p. 185-194. 6. Waterhouse, A., et al., SWISS-MODEL: homology modelling of protein structures and complexes. 2018. 46(W1): p. W296-W303. 7. Lai, X., et al., Structure of human tyrosinase related protein 1 reveals a binuclear zinc active site important for melanogenesis. 2017. 56(33): p. 9812-9815. 8. Pretzler, M., A. Bijelic, and A.J.S.r. Rompel, Heterologous expression and characterization of functional mushroom tyrosinase (Ab PPO4). 2017. 7(1): p. 1810. 9. Lee, Y.S., et al., Synthesis of tyrosinase inhibitory kojic acid derivative. 2006. 339(3): p. 111-114. 10. Sugimoto, K., et al., Syntheses of arbutin-α-glycosides and a comparison of their inhibitory effects with those of α-arbutin and arbutin on human tyrosinase. 2003. 51(7): p. 798-801. 11. Yangjun Ou, Gudai dui hufumeirongzhongyao de yingyong 1990. 12. Shaodan Huang, et al., Zhongyaomeibaichengfendeshaixuan jiMeibaishuangdezhibei. 2014. 13. Shumiao Shu, Xiao Feng, and Pan.J. Modern pharmacy and clinics, 297Meibaizhongyao yu zhiwuyao de yanjiujinzhan. 2007. 22(6): p. 252-255. 14. Zhongguo chengxiangqiye Weisheng, Lu, J., Waiyongmeibaizhongyao de Yanjiujinkuang (Review) . 2014(1): p. 165-167. 15. Haiyi Tang, Guanbang He, and Xilin Zhou. Meibaizhongyaozhishui ji Yichuntiquwu dui Laoansuanmeiyizhi gongxiaozhibijiao. 2005. 16. Beibei Zhang, et al., Chuantong zhongyao Zaimeibaishangdeyingyong. 2015(2015.10): p. 73-75. 17. Mu Zhang, et al., Jizhongmeibaizhongyao Zuoyongdeyanjiu. 2009. 2009(1): p. 33-36. 18. Chen, M.-J., et al., Novel synthetic kojic acid-methimazole derivatives inhibit mushroom tyrosinase and melanogenesis. 2016. 122(6): p. 666-672.

Chapter 3 Chemical investigation of plant

Xanthium strumarium L.

3.1 Xanthium Strumarium L. Xanthium strumarium L. is a well-known TCM (Chinese name: Cang-Er-Zi) belonged to daisy family.[1] It is native to the Americas and eastern China. The fruits of Xanthium have been listed in the Pharmacopoeia of the People’s Republic of China (CH. P) since 1963. It has been commonly applied for the treatment of rhinitis and headache for thousands of years. The whole plant, leaves, flowers and fruits are shown in Figure 3.1A. Currently, over 60 kinds of prescriptions contained Xanthium strumarium L. for treating various diseases, including anti-melanogenesis activity. Crude extract has been reported to inhibit melanin synthesis in a concentration-dependent manner at the concentration of 0-50 µg/mL using Mel-Ab cells[2]. Two other major biological activities associated with Xanthium were anti-allergic-rhinitis (anti-AR)[3] and anti-cancer effects[4].

A B C Figure 3.1A Photographs of Xanthium strumarium L. Figure A adapted from https://www.flickr.com/photos/99758165@N06/ Figure B adapted fromhttps://www.jadeinstitute.com/jade/ Figure C adapted from https://commons.wikimedia.org/wiki/

According to modern pharmacological studies, Xanthium has been frequently developed for the treatment of nasal diseases, especially allergic rhinitis. Water extract of Xanthium was reported to inhibit systemic anaphylaxis in mice (0.01 to 1 g/kg, p.o.),[5] indicating anti-AR effect on animal model. Additionally, water extract of Xanthium (0.25-1 mg/mL)[6] can regulate human mast cell-mediated and peripheral 35 blood mononuclear cell-mediated inflammatory and immunological reactions.[7]

Figure 3.1B Isolated compounds with anti-cancer effects from Xanthium strumarium L.

Xanthatin (3.10)[8] has significant inhibitory activity against lung cancer cells with an

[9] IC50 value of 13.9 µg/mL (Cell lines of A549, H1299, H1650 and HCC827) . In addition,

[10] its analogs, xanthinosin (3.11) (IC50 values of 4.8 µg/mL) , 8-epi-xanthatin (3.12) (IC50 values of 0.1 µg/mL)[11] have also been identified that possessed inhibitory activity against breast cancer MDA-MB-231 cells and colon cancer HCT-15 cells. Furthermore, eremophil-1(10),11(13)-dien-12,8β-olide (3.13),8-epi-xanthatin-1β,5β-epoxide (3.14) and tomentosin (3.15) (IC50 values of 13.22, 2.43, 4.53 µM) isolated from the leaves of Xanthium were determined to have anti-cancer effect against BGC-823 cells and KE-97 cells.[12]

Other biological activities have also been reported includes antioxidant effects[13], anti- inflammatory effects[14], antidiabetic effects[15], antilipidemic effects[16], antibacterial[17] and antifungal effects[18], insecticide and antiparasitic effects[19]. Large amount of effort has been devoted to the pharmacological and phytochemical research on Xanthium, which resulted in the isolation of over 170 compounds from this plant.[20]

Figure 3.1C Representative compounds isolated from Xanthium strumarium L.

Apart from some sesquiterpenoids such as xanthatin, many other types of compounds have been isolated from Xanthium. For examples, phenylpropenoids like xanthiumnolic A (3.16), xanthiumnolic C (3.17)[21] and coniferine (3.18)[22], courmarins like xanthiumnolic B (3.19)[23], leptolepisol D (3.20) and diospyrosin (3.21)[24], steroids like daucosterol (3.22), stigmasterol (3.23) and taraxasteryl acetate (3.24)[24], glycosides like (6Z)-3-hydroxymethyl-7-methylocta-1,6-dien-3-ol 8-O-β-D- glucopyranoside (3.25), (6E)-3hydroxymethyl-7-methylocta-1,6-dien-3-ol 8-O-β-D- glucopyranoside (3.26) and everlastoside C (3.27)[25], flavonoids like ononin (3.28), quercetin (3.29) and allopatuletin (3.30) and other compounds like uracil (3.31), sibiricumthionol (3.32) and indole-3-carbaldehyde (3.33)[26].

Dry powder of Xanthium

Hexane Methanol DCM extract extract extract C18 Semi-preparative HPLC chromatography

Fraction A Fraction B Fraction C Fraction D Fraction E Fraction F Fraction G * * * * Sugar (not Caffeoyl Not active ( not 3-CQA 1,3-DCQA 1,5-DCQA investigated) choline investigated)

Uridine 3,5-DCQA 4,5-DCQA

Cytidine Xanthiside

4-CQA

Scheme 3.2A Bioassay-guided isolation of 10 compounds from Xanthium. (* Repeated Luna C18 HPLC chromatography with MeOH/H2O gradient elution. CQA = caffeoylquinic acid, DCQA = dicaffeoylquinic acid.)

Figure 3.2B All 10 compounds isolated from Xanthium Strumarium L.

Little data was reported associated with Xanthium on anti-hyperpigmentation activity. To investigate the pure compounds in Xanthium with inhibitory activity of tyrosinase, bioassay-guided extraction and isolation was applied in this project. As shown in Scheme 3.2A, further purifications were carried out by repeated C18 and Luna C18 HPLC chromatography to afford 10 compounds, including 6 caffeoylquinic acids, two alkaloids, one choline and one glycoside. They are 3-caffeoylquinic acid (3.35)[27], 4- caffeoylquinic acid (3.36)[28], 1,3-dicaffeoylquinic acid (3.37)[29], 3,5-dicaffeoylquinic acid (3.38)[30], 1,5-dicaffeoylquinic acid (3.39)[31], 4,5-dicaffeoylquinic acid (3.40), caffeoylcholine (3.41)[32], xanthiside (3.42)[33], uridine (3.43)[34] and cytidine (3.44)[35] (Figure 3.2B).

The structures were elucidated by 1D-, 2D-NMR and MS spectroscopic data which were confirmed by the comparison with those in the literature. Detailed purifications and structure elucidations were introduced in the following sections.

3.2 Bioassay-guided extraction and isolation The TCM material was purchased from local Chinese pharmacy (Figure 3.3). The fruit of the plant material was sequentially extracted with hexane, dichloromethane (DCM) and methanol to afford hexane, DCM and methanol extracts respectively. Three extracts were then tested their inhibitory activity against mushroom tyrosinase.

Figure 3.3 Purchased TCM Xanthium strumarium L.

120

100

40 Inhibition/%

0 + - Hexane DCM extract Methanol extract extract Samples Figure 3.2 Inhibitory activity of hexane, DCM, methanol extract against tyrosinase (concentration for all extracts: 1mg/mL)

As shown in Figure 3.2, at the concentration of 1 mg/mL, it showed that the methanol extract of Xanthium possessed the most of inhibitory activity against tyrosinase with almost 68.7% inhibition. Compared to the methanol extract, the hexane extract showed little inhibition and the DCM extract showed about 3.7% inhibition against tyrosinase. According to the biological activity, methanol extract was chosen for further chemical investigations.

Figure 3.4A Semi-preparative HPLC UV trace of methanol extract of Xanthium with gradient MeOH/H2O (0.1% TFA). Timed fractions (1 min x 90) were collected. (Black at 210 nm, green at 280 nm, pink at 254 nm, blue at 320 nm)

120

100

Inhibition/% 40

0 + - A B C D E F G Fraction

Figure 3.4B Inhibitory activity of seven fractions from semi-preparative HPLC chromatography (concentration for all fractions: 1 mg/mL)

The methanol extract of Xanthium was fractionated by C18 semi-preparative HPLC chromatography eluting with gradient MeOH/H2O (0.1 % TFA) (Figure 3.4A). Based on peaks distribution, seven fractions were obtained according to the separation (Fraction A: 1-5 min, Fraction B: 6-31 min, Fraction C: 32-39 min, Fraction D: 40-49 min, Fraction E: 50-69 min, Fraction F: 70-75 min, Fraction G: 76-90 min). To trace the inhibitory activity against tyrosinase, seven fractions were tested by mushroom tyrosinase assay. As shown in Figure 3.4B, among all fractions, fraction A had the greatest inhibitory activity, almost 61.2% inhibition against tyrosinase. Fraction B-E showed approximately 43.2%, 21.1%, 57.6% and 27.6% inhibition respectively. The last two fractions, F and G showed minimal inhibitory activity against tyrosinase.

1 Figure 3.5 H NMR (DMSO-d6, 800MHz) spectra of fractions A-E

1H NMR spectra were obtained for the five active fractions from A to E (Figure 3.5).

Fraction B, C, D, E shared similar proton signals in aromatic region (δH 6.5-8.5), which indicated that each of five fractions contained compounds with similar aromatic

1 scaffold or vinyl groups. H NMR spectra for fraction A contained proton signals at δH 3- 5, which suggested that the major compounds in fraction A were sugars and/or polysaccharides. The main interest of this project was to isolate small molecules rather than structure class such as sugar. Therefore, fraction A was not further investigated.

3.2.1 Further purification for fraction B As shown in Figure 3.6, fraction B was further purified by Luna C18 HPLC, eluting with gradient MeOH/H2O (0.1% TFA). Five sub-fractions with the same peak group were

42 combined and tested against tyrosinase (Fraction B1: 1-6 min, Fraction B2: 7-10 min, Fraction B3: 11-12 min, Fraction B4: 13-14 min, Fraction B5: 15-30 min).

Figure 3.6 Semi-preparative HPLC UV trace of Fraction B with gradient MeOH/H2O (0.1%TFA). Timed fractions (1 min x 30) were collected. (Black at 210 nm, green at 280 nm, pink at 254 nm, blue at 320 nm) 100 90 80 70 60 50

40 Inhibition/% 30 20 10 0 kojic acid DMSO B1 B2 B3 B4 B5 Fraction

Figure 3.7 Inhibitory activity of combined fractions after HPLC chromatography of Fraction B (concentration for all fractions: 0.5 mg/mL)

As indicated in Figure 3.7, peaks with retention time between 7-10 min showed strong inhibitory activity with almost 70% inhibition against tyrosinase. The second active region with retention time of 13-14 min showed 57% inhibitory activity. Based on the activity data, further purification was carried out to pure compounds from the two active fractions. Figure 3.8A showed HPLC trace of B2 and Figure 3.8B showed the results of separation for B4. Fractions that potentially contained pure compounds were analyzed by 1H NMR. The structures were elucidated by 1D-, 2D-NMR and MS spectroscopic data. Based on the chromatograms, 2 compounds, uridine (3.43) and cytidine (3.44) were isolated from B2 and 2 compounds, caffeoylcholine (3.41) and 4- caffeoylquinic acid (3.36) were isolated from B4.

Figure 3.8B Semi-preparative HPLC UV trace of Fraction B4 with gradient MeOH/H2O (0.1% TFA). Timed fractions (1 min x 60) were collected. (Black at 210 nm, green at 280 nm, pink at 254 nm, blue at 320 nm)

Uridine (3.43) Uridine (3.43) was obtained as white powder (1.4 mg, percentage yield of 0.014%) from the fraction at 7 min in Figure 3.8A. The positive ion LC-MS showed a pseudomolecular ion at m/z = 245 ([M+H] +), which supported the molecular weight of 244 Da. As shown in 1H NMR spectra (Figure 3.9) and 2D-NMR data (Table 3.9), the doublet at δH 7.85 was coupled to the doublet at δH 5.64 (J= 8.1 Hz). Carbons that attached to these protons were resonating at δC 141.0 and 101.4 respectively, which was indicated by HSQC spectrum. Additionally, the uracil pyrimidine moiety was assigned by the HMBC correlations from proton at δH 7.85 to the carbons at δC 151.0 and 163.0. Further HMBC correlation from δH 7.85 to δC 87.2 suggested that the ribose moiety was attached to the uracil. The NMR data of uridine was consistent with those reported in the literature. The stereochemistry was also consistent with the literature[36] by the comparison of the coupling constants of the protons on the ribose ring, as shown in Table 3.9.

1 Figure 3.9 H NMR (DMSO-d6, 800 MHz) spectra of uridine (3.43)

Table 3.9 NMR data for uridine (3.43) Position 13C* (δ) 1H (δ, multiplets, J in Hz) HMBC COSY 1-NH 11.33 (s) 2 163.0 - 4 141.0 7.85 (d, 8.1Hz) 2, 5, 6, 1’ 5 5 101.4 5.64 (d, 8.1Hz) 4 4 6 151.0 - 1’ 87.2 5.75 (d, 5.5Hz) 2, 4, 2’ 2’ 2’ 73.1 4.01 (q, 5.5Hz) 1’ 1’, 3’ 3’ 69.3 3.95 (q, 5.1Hz) 1’ 2’, 4’ 4’ 84.8 3.83 (q, 3.5Hz) 3’ 3’, 5’ 5’ 60.6 3.53 (ddd, 12.1, 5.1, 3.4Hz) 4’ 3.60 (ddd, 12.1, 5.1, 3.4Hz) 4’ 2’-OH - 5.43 3’-OH - 5.20 5’-OH - 5.17 1 Spectra were recorded at 800MHz for H using DMSO-d6 as solvent. *Carbon chemical shift resourced from 2D NMR spectra.

Cytidine (3.44)

1 Figure 3.10 H NMR (DMSO-d6, 800 MHz) spectra of Cytidine (3.44)

Cytidine (3.44) was obtained as white powder (1.1 mg, percentage yield of 0.01%) from the fraction at 8 min in Figure 3.8A. The molecular weight was determined to be

243 Da based on m/z = 244 ([M+H]+) from LC-MS data. As shown in 1H NMR (Figure

3.10) and 2D-NMR data (Table 3.10), protons at δH 5.74 represented two protons, a doublet of doublet (7.5Hz and 3.9Hz), which showed COSY correlations to protons at

δH 7.83 and 3.81 respectively. The chemical shifts of these protons at δH 3-5 were similar to those in uridine (3.43), which indicated the similar ribose scaffolds.

Compared to the proton signals in uridine (3.43), two exchangeable protons at δH 7.24 and 7.02 were assumed to be amine group, which was also proved by the MS measurement. The data was comparable to those reported in the literature.[35]

Table 3.10 NMR data for cytidine (3.44) Position 13C* (δ) 1H (δ, multiplets, J in Hz) HMBC COSY - 2-NH2 7.24 (brs) & 7.02 (brs) 2 165.9 - 3 94.0 5.74 (d, 7.5Hz) 4, 6 4 4 141.6 7.83 (d, 7.5Hz) 1’, 2, 3, 6 3 6 156.1 - 1’ 89.4 5.74 (d, 3.9Hz) 6, 4’ 2’ 2’ 84.5 3.81 (q, 3.9Hz) 4’ 1’, 3’ 3’ 69.9 3.92 (t, 4.5Hz) 1’ 2’, 4’ 4’ 74.4 3.92 (t, 4.5Hz) 2’ 3’, 5’ 5’ 61.1 3.54 (dd, 5.3 & 3.5Hz) 4’ 4’ 3.52 (dd, 5.3 & 3.5Hz) 4’ 4’ 2’-OH - 5.33 (d, 3.9Hz) 3’-OH - 5.08 (d, 4.5Hz) 5’-OH - 5.16 (t, 5.3Hz) 1 Spectra were recorded at 800 MHz for H using DMSO-d6 as solvent. *Carbon chemical shift resourced from 2D NMR spectra.

Caffeoylcholine (3.41)

1 Figure 3.11 H NMR (CD3OD-d4, 800 MHz) spectrum of caffeoylcholine (3.41)

Caffeoylcholine (3.41) was obtained as light yellow hygroscopic solid from the fraction B4 at 2 min shown in Figure 3.8B (5.2 mg, percentage yield of 0.05%). According to the 1H NMR spectrum in Figure 3.11, there were two vinyl doublets which were coupled to each other (δH 7.53 and 6.32). The coupling constants of three aromatic signals (δH 7.08, 7.02 and 6.78) indicated the positional relationship on one aromatic ring. The skeleton of the caffeoyl group was confirmed by HMBC correlations (Table 3.11) from proton at δH 7.53 to carbons resonated at δC 115.3 and 122.1, from proton at δH 6.32 to carbon at δC 166.4. Additionally, there were three chemical-equivalent methyl singlets at δH 3.16, which indicated that these methyl groups were attached to the same heteroatom. The molecular weight was determined to be 266 Da by LC-MS data. Therefore, the heteroatom was calculated to be nitrogen and the compound was positively charged. The data obtained was comparable with those reported in the literature.[37]

Table 3.11 NMR data for caffeoylcholine (3.41) Position 13C* (δ) 1H (δ, multiplets, J in Hz) HMBC COSY 1 125.8 2 115.3 7.08 (d, 2.1Hz) 3, 4, 6 3 146.1 4 149.2 5 116.1 6.78 (d, 8.1Hz) 1, 3, 4 6 6 122.1 7.02 (dd, 8.1 & 2.1Hz) 2, 4 5 7 146.7 7.53 (d, 15.9Hz) 1, 2, 6, 8, 9 8 8 113.7 6.32 (d, 15.9Hz) 1, 9 7 9 166.4 10 58.1 4.56 (m, 2H) 1, 9 11 64.3 3.71 (m, 2H) 10, 12 12 53.6 3.16 (s, 9H) 11, 12 1 Spectra were recorded at 800 MHz for H using CD3OD-d4 as solvent. *Carbon chemical shift resourced from 2D NMR spectra.

4-caffeoylquinic acid (3.36) 4-caffeoylquinic acid (3.36) was obtained as light yellow powder from the fraction at 20 min shown in Figure 3.8B (0.4 mg, percentage yield of 0.04%). As shown in Figure

1 3.12, H NMR at δH 6-8 gave the characteristic signals of caffeoyl group. As indicated by the HSQC correlation in Table 3.12 from δH 4.82 to δC 77.0, the proton signal was overlapped with the water peak. The two methylene groups on quinic acid were at δH

2.21, 2.08, 2.18 and 2.01. the other two proton signals were shown at δH 4.29-4.31.

The attachment of caffeoyl group was indicated by the HMBC correlation from δH 4.82 to δC 166.3. The pseudomolecular weight of the structure was determined to be 354 Da as the LC-MS data showed a molecular ion at m/z = 353 ([M-H]-). The data obtained from 1D- and 2D-NMR was comparable with those in the literature.[38]

1 Figure 3.12 H NMR (CD3OD, 800 MHz) spectrum of 4-caffeoylquinic acid (3.36)

Table 3.12 NMR data for 4-caffeoylquinic acid (3.36) Position 13C* (δ) 1H (δ, multiplets, J in Hz) HMBC COSY

1 75.3 2 41.4 2.21&2.08 (m) 1, 6, 7 3 3 63.9 4.31 (m) 1, 4 4 4 77.0 4.82 (overlapped) 3, 8 3, 5 5 68.4 4.29 (m) 1, 3, 6 4 6 40.0 2.18&2.01 (m) 1, 4, 5, 7 7 175.9 8 166.3 9 113.9 6.38 (d, 15.3Hz) 8, 10, 11 10 10 145.5 7.65 (d, 15.3Hz) 8, 12, 16 9 11 126.1 12 113.7 7.07 (d, 2.1Hz) 13, 14, 16 13 148.5 14 121.5 15 115.0 6.79 (d, 8.1Hz) 13, 14 16 16 121.8 6.97 (dd, 2.1 & 8.1Hz) 8, 11, 14 15 1 Spectra were recorded at 800 MHz for H using CD3OD as solvent. *Carbon chemical shift resourced from 2D NMR spectra.

3.2.2 Further purification for fraction C As shown in Figure 3.13A, fraction C was further purified by Luna C18 HPLC eluting with gradient MeOH/H2O (0.1% TFA). Similarly, mushroom tyrosinase assay was run for the combined fractions to find the active fractions (Fraction C1: 1-11 min, Fraction C2: 12-22 min, Fraction C3: 23 min, Fraction C4: 24-25 min, Fraction C5: 26-60 min). Based on the HPLC chromatogram, the peak at 23 min was relatively pure, the peak at 24-25 min was not pure enough and given the concentration of the fraction was very low, the fraction was not further purified. As shown in Figure 3.13B, peak with retention time at 23 min showed moderate inhibition with almost 27% inhibitory activity against tyrosinase, followed by peaks at 24-25 min with 21% inhibitory percentage. Since the fraction at 23 min was relatively pure, the chemical structure was elucidated using 1D- and 2D-NMR without any further purifications. The rest of peaks showed little inhibition against tyrosinase which were not considered for further purifications. Based on the 1H NMR, 3-caffeoylquinic acid (3.35) was isolated from C3.

Figure 3.13A Semi-preparative HPLC UV trace of Fraction C with gradient MeOH/H2O (0.1% TFA). Timed fractions (1 min x 60) were collected. (Black at 210 nm, green at 280 nm, pink at 254 nm, blue at 320 nm) 100 90 80 70 60 50 40

inhibition/% 30 20 10 0 kojic acid DMSO C1 C2 C3 C4 C5 Fraction

Figure 3.13B Inhibitory activity of combined fractions after HPLC chromatography of Fraction C (concentration for all fractions: 0.5 mg/mL) 49

3-caffeoylquinic acid (3.35)

1 Figure 3.14 H NMR (CD3OD, 800 MHz) spectrum of 3-caffeoylquinic acid (3.35)

Table 3.14 NMR data for 3-caffeoylquinic acid (3.35) Position 13C* (δ) 1H (δ, multiplets, J in Hz) HMBC COSY

1 74.1 2 37.1 1.94 (m) 1, 6, 7 3 3 71.3 5.07 (td, 9.0 & 4.4Hz) 1, 4, 8 2, 4 4 70.9 3.57 (dd, 8.5 & 3.2Hz) 3, 6 3 5 68.9 3.92 (dt, 5.5 & 3.2Hz) 1, 3, 6 6 1.79 (m) 6 37.6 2.02 (m) 1, 4, 5, 7 5 7 175.6 8 166.3 9 114.8 6.15 (d, 15.9Hz) 8, 10, 11 10 10 145.6 7.42 (d, 15.9Hz) 8, 12, 16 9 11 126.3 12 115.3 7.03 (d, 2.1Hz) 13, 14, 16 13 145.8 14 148.7 15 116.4 6.77 (d, 8.1Hz) 13, 14 16 16 121.8 6.98 (dd, 2.1 & 8.1Hz) 8, 11, 14 15 1 Spectra were recorded at 800 MHz for H using CD3OD as solvent. *Carbon chemical shift resourced from 2D NMR spectra.

3-caffeoylquinic acid (3.35) was obtained as dark brown amorphous powder from the fraction at 23 min shown in Figure 3.14 (2.1 mg, percentage yield of 0.02%). The LC-MS data showed a molecular ion at m/z = 353 ([M-H]-), which indicated that the molecular weight was 354 Da. As shown in 1H NMR spectra (Figure 3.14) and 2D-NMR data (Table 50

3.14), the aromatic signals of 3.35 was similar to that of 3.41, which proved the presence of caffeoyl group. Other signals (Table 3.14) showed the scaffold of quinic acid. The data obtained from 1D- and 2D-NMR was comparable with those in the literature. Additionally, the stereochemistry was confirmed by the comparison of the coupling constants indicated in the table with those in the literature.[39]

3.2.3 Further purification for fraction D As shown in Figure 3.15A, fraction D was further purified and 60 fractions were combined based on the peaks (Fraction D1: 1-10 min, Fraction D2: 11-12 min, Fraction D3: 13-14 min, Fraction D4: 15-24 min, Fraction D5: 25-34 min, Fraction D6: 35-60 min). To determine the activity, mushroom tyrosinase assay was run for all combined fractions. According to Figure 3.15B, among six fractions, fraction D3 and D5 showed evident inhibitory activity against tyrosinase with inhibition percentage of almost 54.7% and 51.1% respectively. Compared to kojic acid, other fractions showed little inhibition against tyrosinase, which were similar to the negative control.

Figure 3.15A Semi-preparative HPLC UV trace of Fraction D with gradient MeOH/H2O (0.1% TFA). Timed fractions (1 min x 60) were collected. (Black at 210 nm, green at 280 nm, pink at 254 nm, blue at 320 nm)

100 90 80 70 60 50 40

inhibition/% 30 20 10 0 kojic DMSO D1 D2 D3 D4 D5 D6 acid Fraction

Figure 3.15B Inhibitory activity of combined fractions after HPLC chromatography of Fraction D (concentration for all fractions: 0.5 mg/mL)

1H NMR of fraction D3 indicated a pure compound, the structure of the compound in Fraction D3 was determined to be xanthiside (3.42). For fraction D5, further purification was carried out using Luna C18 column HPLC chromatography with gradient MeOH/H2O (0.1% TFA) and the HPLC trace was shown in Figure 3.16A. Similarly, fractions were combined based on the peak separation (Fraction D5-1: 1-4 min, Fraction D5-2: 5-22 min, Fraction D5-3: 23-25 min, Fraction D5-4: 26-30 min, Fraction D5-5: 31-60 min). As shown in Figure 3.16B, fraction D5-1, D5-2 and D5-5 showed little inhibitory activity against tyrosinase, fraction D5-4 showed almost 12.5% inhibition and fraction D5-3 showed the most inhibition, almost 65.4%. 1H NMR spectra for both fractions, D5-3 and D5-4, indicated two individual pure compounds, 1,3-dicaffeoylquinic acid (3.37) and 1,4-dicaffeoylquinic acid (3.38).

Figure 3.16A Semi-preparative HPLC UV trace of Fraction D5 with gradient MeOH/H2O (0.1% TFA). Timed fractions (1 min x 60) were collected. (Black at 210 nm, green at 280 nm, pink at 254 nm, blue at 320 nm)

100 90 80 70 60 50

40 Inhibition/% 30 20 10 0 kojic acid DMSO D5-1 D5-2 D5-3 D5-4 D5-5 Fraction

Figure 3.16B Inhibitory activity of combined fractions after HPLC chromatography of Fraction D5 (concentration for all fractions: 0.5 mg/mL)

Xanthiside (3.42)

1 Figure 3.17 H NMR (CD3OD-d4, 800 MHz) spectrum of xanthiside (3.42)

Table 3.17 NMR data for xanthiside (3.42) Position 13C* (δ) 1H (δ, multiplets, J in Hz) HMBC COSY 1 166.1 2 42.2 3 142.6 4 129.6 5 170.6 6 121.5 6.69 (s) 1, 2, 4, 10 7 28.3 3.51 (s) 3, 8 8 163.5 9 26.1 1.49 (d, 11.9Hz) 1, 2, 3 10 66.6 4.52 & 4.78 (dd, 15.9 &1.6Hz) 1, 2, 6, 1’ 1’ 102.5 4.40 (d, 7.8Hz) 10, 2’, 3’ 2’ 2’ 76.6 3.31 (m) 1’ 1’ 3’ 73.7 3.31 (m) 2’ 4’ 70.3 3.31 (m) 5’ 5’ 76.7 3.41 (m) 6’ 5’ 6’ 61.8 3.69 & 3.90 (m) 4’, 5’ 6’ 1 Spectra were recorded at 800 MHz for H using CD3OD-d4 as solvent. *Carbon chemical shift resourced from 2D NMR spectra.

Xanthiside (3.42) was obtained as light yellow powder from the fraction at 13 min shown in Figure 3.15A (0.9 mg, percentage yield of 0.009%). As shown in Figure 3.17,

1 H NMR showed characteristic signals for the glucose moiety (δH 3-4.5), two methyl singlets were also observed at δH 3.51 which were chemically equivalent. Two methylene groups were observed at δH 3.69 and 3.90, δH 4.52 and 4.78 based on HSQC correlations to the carbons at δC 61.8 and 66.6 respectively from Table 3.17. the

53 proton at δH 6.69 was the only signal in aromatic region responsible for a vinyl proton.

According to the HMBC correlations from δH 4.40 to δC 66.6 and δH 6.69 to δC 66.6, the other partial structure was attached to the glucose by the methylene carbon at δC 66.6. The stereochemistry of the structure was indicated by the coupling constants of proton signals on glucose (Table 3.17). The LC-MS data showed a pseudomolecular ion at m/z = 402 ([M+H]+), which indicated that the molecular weight was 401 Da, which indicated there was at least one nitrogen atom. Except for the nitrogen and all the carbons and hydrogens, mass of 32 Da was left that responsible for a sulphur atom. The structure was confirmed by comparison with those data reported in the literature.[40]

1,3-dicaffeoylquinic acid (3.37)

1 Figure 3.18 H NMR (CD3OD-d4, 800 MHz) spectrum of 1,3-dicaffeoylquinic acid (3.37)

1,3-dicaffeoylquinic acid (3.37) was obtained as dark green powder from the fraction at 23 min shown in Figure 3.16A (0.5mg percentage yield of 0.005%). The positive ion LC-MS showed a molecular ion at m/z = 515 ([M-H]-), which suggested the molecular weight of 516 Da. As shown in Figure 3.18, there were two sets of signals of caffeoyl groups at δH 6.5-8.0. The rest of proton signals were similar to those of 3.35, which indicated the quinic acid ring. The attachment of the second caffeoyl group was suggested by the HMBC correlation from δH 5.37 to δC 167.4 and 166.1 in Table 3.18.

54 the stereochemistry was determined by the comparison of the coupling constants on quinic acid with those in the literature.[41]

Table 3.18 NMR data for 1,3-dicaffeoylquinic acid (3.37) Position 13C* (δ) 1H (δ, multiplets, J in Hz) HMBC COSY 1 79.6 2.89&2.31 (dd, 16.1, 2 31.5 3 3 3.3Hz) 3 71.6 5.37 (q, 3.4Hz) 1, 8a/8b 2 4 66.4 4.23 (m) 5, 6 5 5 73.8 3.63 (dd, 9.5, 3.4Hz) 1, 4 4, 6 2.53&1.84 (dd, 13.7, 6 40.0 1, 2, 5, 7 5 11.2Hz) 7 173.0 8a/8b 167.4/166.1 9a/9b 113.6/113.9 6.12/6.19 (d, 15.9Hz) 16a/16b, 8a/8b 10a/10b 8a/8b, 12a/12b, 10a/10b 146.4/145.8 7.49/7.47 (d, 15.9Hz) 9a/9b 16a/16b 11a/11b 125.8/125.1 13a/13b, 14a/14b, 12a/12b 114.7/114.0 6.93/6.82 (d, 2.1Hz) 16a/16b 13a/13b 145.3/145.0 14a/14b 148.3/147.9 15a/15b 121.6/120.7 6.75/6.59 (d, 8.2Hz) 11a/11b, 14a/14b 16a/16b 16a/16b 115.1/115.2 6.64/6.52 (dd, 8.2, 2.1Hz) 12a/12b, 14a/14b 15a/15b 1 Spectra were recorded at 800 MHz for H using CD3OD-d4 as solvent. *Carbon chemical shift resourced from 2D NMR spectra.

3,5-dicaffeoylquinic acid (3.38)

1 Figure 3.19 H NMR (CD3OD-d4, 800 MHz) spectrum of 3,5-dicaffeoylquinic acid (3.38) 55

3,5-dicaffeoylquinic acid (3.38) was obtained as light green powder from the fraction at 29 min shown in Figure 3.16A (0.4 mg, percentage yield of 0.004%). The molecular weight was determined to be 516 Da based on the positive ion at m/z =515 ([M-H]-) from LC-MS. As shown in Figure 3.19, the characteristic proton signals at δH 6.5-8 suggested the scaffold of the caffeoyl groups. The HMBC correlations in Table 3.19 from δH 5.40 to δC 168.9 and δH 5.39 to δC 168.1 indicated the attachments of two caffeoyl groups to the quinic acid. The 1D- and 2D-NMR data was comparable to those from the literature.[42]

Table 3.19 NMR data for 3,5-dicaffeoylquinic acid (3.38)

Position 13C* (δ) 1H (δ, multiplets, J in Hz) HMBC COSY 1 74.9 2.16&2.31 (dd, 13.1, 2 37.8 3 3 3.3Hz) 3 72.3 5.40 (m) 8a 2 4 71.0 4.01 (dd, 7.6, 3.3Hz) 5, 6 5 5 72.2 5.39 (t, 6.7Hz) 8b 4, 6 36.1 2.24 (m) 1 5 7 178.1 8a/8b 168.9/168.1 9a/9b 115.3 6.26/6.34 (d, 15.9Hz) 8a/8b 10a/10b 10a/10b 146.8 7.60/7.58 (d, 15.9Hz) 8a/8b, 12a/12b 9a/9b 11a/11b 128.0 12a/12b 116.9 7.06 (d, 2.1Hz) 13a/13b, 16a/16b 13a/13b 147.8 14a/14b 149.3 15a/15b 115.7 6.77 (d, 8.2Hz) 11a/11b 16a/16b 16a/16b 113.0 6.95 (dd, 8.2, 2.1Hz) 12a/12b, 14a/14b 15a/15b 1 Spectra were recorded at 800 MHz for H using CD3OD-d4 as solvent. *Carbon chemical shift resourced from 2D NMR spectra.

3.2.4 Further purification for fraction E

Figure 3.20A Semi-preparative HPLC UV trace of Fraction E with gradient MeOH/H2O (0.1% TFA). Timed fractions (1 min x 25) were collected. (Black at 210 nm, green at 280 nm, pink at 254 nm, blue at 320 nm)

100 90 80 70 60 50 40

inhibition/% 30 20 10 0 kojic acid DMSO E1 E2 E3 E4 Fraction

Figure 3.20B Inhibitory activity of combined fractions after HPLC chromatography of Fraction E (concentration for all fractions: 0.5 mg/mL)

As shown in Figure 3.20A, HPLC chromatogram was obtained by Luna C18 column, eluting with gradient MeOH/H2O (0.1% TFA). Based on the peak’s distribution, four fractions were obtained and tested against tyrosinase (Fraction E1: 1-10 min, Fraction E2: 11-12 min, Fraction E3: 13-14 min, Fraction E4: 15-25 min).

The biological results shown in Figure 3.20B indicated that peaks with retention time 11-12 min, 13-14 min showed moderate inhibitory activity against tyrosinase, 36% and 39% respectively. However, the other two fractions, E1 and E4 showed little inhibition with less than 5% inhibition. 1H NMR spectra for fraction E2 and E3 indicated two individual pure compounds, 1,5-dicaffeoylquinica acid (3.39) for E2 and 4,5- dicaffeoyquinic acid (3.40) for E3.

1,5-dicaffeoylquinic acid (3.39)

1 Figure 3.21A H NMR (CD3OD-d4, 800 MHz) spectrum of 1,5-dicaffeoylquinic acid (3.39)

Table 3.21 NMR data for 1,5-dicaffeoylquinic acid (3.39) Position 13C* (δ) 1H (δ, multiplets, J in Hz) HMBC COSY 1 79.6 2 34.5 2.45 (m) 3 3 3 67.9 4.29 (m) 2 4 71.2 3.78 (m) 5 5 70.3 5.38 (td, 8.4, 4.1Hz) 8a/8b 4, 6 6 36.1 2.07&2.57 (m) 5, 7 2, 5 7 173.2 8a/8b 166.3 9a/9b 114.7 6.27/6.29 (d, 15.9Hz) 16a/16b, 8a/8b 10a/10b 8a/8b, 12a/12b, 10a/10b 145.8 7.59/7.60 (d, 15.9Hz) 9a/9b 16a/16b 11a/11b 125.5 13a/13b, 14a/14b, 12a/12b 115.3 7.06 (d, 8.0Hz) 16a/16b 13a/13b 146.1 14a/14b 148.9 15a/15b 116.2 6.77 (d, 2.1Hz) 14a/14b, 16a/16b 16a/16b 16a/16b 121.8 6.97 (dd, 2.1, 8.0Hz) 12a/12b, 14a/14b 15a/15b 1 Spectra were recorded at 800 MHz for H using CD3OD-d4 as solvent. *Carbon chemical shift resourced from 2D NMR spectra.

1,5-dicaffeoylquinic acid (3.39) was obtained as light yellow powder from the fraction at 11 min shown in Figure 3.20A (0.9 mg, percentage yield of 0.009%). Since there was a pseudomolecular ion at m/z = 517 ([M+H])+from LC-MS, the molecular weight of 3.39 was 516 Da. As shown in Figure 3.21, proton signals in the aromatic region indicated two caffeoyl groups. The two sets of vinyl groups in the caffeoyl groups were at δH 6.21 and 7.47, 6.22 and 7.48. The aromatic protons of two caffeoyl groups were overlapped with each other. The HMBC correlation of proton (Table 3.21) at δH 5.21 to δC 166.3 suggested the second caffeoyl group was attached to the 5-position on the quinic acid ring. The stereochemistry of the compound was determined by the comparison of coupling constants on the quinic acid with those reported in the literature.[31]

4,5-dicaffeoylquinic acid (3.40) 4,5-dicaffeoylquinic acid (3.40) was obtained as light green powder from the fraction at 13 min shown in Figure 3.22 (0.9 mg, percentage yield of 0.005%). The molecular weight was determined to be 516 Da based on the positive ion at m/z =515 ([M-H]-) from LC-MS. As shown in Figure 3.22, 1H NMR spectra indicated the characteristic proton signals of caffeoyl groups at δH 6.5-8. The HMBC correlation in Table 3.22 from

δH 5.07 to δC 168.3 and δH 5.66 to δC 168.3 indicated that the two caffeoyl groups were attached to the protons at δH 5.07 and 5.66 on the quinic acid ring. The NMR data was comparable with those reported in the literature.[43]

1 Figure 3.22 H NMR (CD3OD-d4, 800 MHz) spectrum of 4,5-dicaffeoylquinic acid (3.40)

Table 3.22 NMR data for 4,5-dicaffeoylquinic acid (3.40) Position 13C* (δ) 1H (δ, multiplets, J in Hz) HMBC COSY 1 75.4 2.26&1.98 (dd, 14.5, 2 37.8 3 3 2.2Hz) 3 67.7 4.28 (dt, 3.2, 2.8Hz) 1 2 4 73.0 5.07 (dd, 10.8, 3.2Hz) 5, 8b 5 5 70.0 5.66 (dt 10.4, 6.6Hz) 1, 8a 4 6 39.9 2.16 (m) 1, 2, 5, 7 2, 5 7 181.9 8a/8b 168.3 9a/9b 115.2 6.25/6.17 (d, 15.9Hz) 16a/16b, 8a/8b 10a/10b 10a/10b 147.3 7.58/7.48 (d, 15.9Hz) 8a/8b 9a/9b 11a/11b 128.0 12a/12b 115.4 6.99 (d, 8.0Hz) 13a/13b, 14a/14b 13a/13b 146.6 14a/14b 149.9 15a/15b 116.1 6.72 (d, 2.2Hz) 16a/16b 16a/16b 16a/16b 123.5 6.88 (dd, 2.2, 8.0Hz) 12a/12b, 14a/14b 15a/15b 1 Spectra were recorded at 800 MHz for H using CD3OD-d4 as solvent. *Carbon chemical shift resourced from 2D NMR spectra.

3.3 Inhibitory activity of the pure compounds against tyrosinase All 10 pure compounds were dissolved in DMSO then diluted in 0.1 M potassium phosphate buffer (pH 6.8) to make final concentration of 0.5 mg/mL. The inhibitory activity of the compounds was tested by mushroom tyrosinase assay.

120 100 80 60

40 Inhibition/% 20 0 kojic 3.35 3.36 3.37 3.38 3.39 3.40 3.41 3.42 3.43 3.44 acid pure compounds

Figure 3.23 Inhibitory activity of isolated compounds against tyrosinase under concentration of 0.5mg/mL

As shown in Figure 3.23, compared to kojic acid (0.5 mg/mL), cytidine (3.44) showed the greatest inhibition against tyrosinase, almost 92%. However, its analog, uridine (3.43) showed only 4% inhibition. The second most active compound was 1,3- dicaffeoylquinic acid (3.37), with almost 87% inhibition. Among all caffeoylquinic acids and their derivatives, 3.37 was the only one showed significant inhibitory activity against tyrosinase at the concentration of 0.5 mg/mL. Xanthiside (3.42) was the third most active compound with 53% inhibition against tyrosinase.

To determine the IC50 value of the three active compounds, dose-response curves were constructed using 7 different concentrations. As shown in Figure 3.24, dose- response curve for cytidine (3.44) contained concentrations ranging from 0.0036-0.9 mM (0.0036, 0.018, 0.09, 0.15, 0.25, 0.45 and 0.9 mM respectively). Cytidine (3.44) showed a dose-dependent inhibitory activity against tyrosinase with an IC50 value of 0.18 mM. Similarly, for 1,3-dicaffeoylquinic acid (3.37), 7 concentrations contained

0.05, 0.1, 0.2, 0.41, 0.68, 0.82 and 1.36 mM were used and the IC50 value was determined to be 0.29 mM. Last but not least, xanthiside (3.42) showed a dose- dependent inhibitory activity against tyrosinase with an IC50 value of 2.47 mM under 7 doses containing 0.05, 0.1, 0.2, 0.41, 0.68, 0.82 and 1.36 mM.

100 120 100 100 80 80 80 60 3.44 60 3.37 60 3.42

40 40 40

inhibition/% Inhibition/% inhibition/% 20 20 20

0 0 0 0 0.5 1 1.5 0 2 4 6 0 0.5 1 -20 Concentration/mM Concentration/mM concentration/mM Figure 3.24 Dose-response curve of the inhibition of cytidine (3.44), 1,3- dicaffeoylquinic acid (3.37), xanthiside (3.42) against tyrosinase.

3.3 Conclusions and discussions Out of 42 readily available TCM materials, Xanthium was chosen for further chemical and biological investigations based on the inhibitory activity against mushroom tyrosinase. Bioassay-guided chemical investigations resulted in the isolation of 10 pure compounds including 6 caffeoylquinic acids: 3-caffeoylquinic acid (3.35), 4- caffeoylquinic acid (3.36), 1,3-dicaffeoylquinic acid (3.37), 3,5-dicaffeoylquinic acid

(3.38), 1,5-dicaffeoylquinic acid (3.39), 4,5-dicaffeoylquinic acid (3.40), two alkaloids: uridine (3.43) and cytidine (3.44), one choline: caffeoylcholine (3.41) and one glycoside: Xanthiside (3.42). The structures of the pure compounds were determined by 1D- and 2D-NMR as well as LC-MS. The inhibitory activity tests of the pure compounds showed that three of them possessed significant inhibition against tyrosinase. Compound 3.44 showed an IC50 value of 0.18 mM, 3.37 showed an IC50 value of 0.29 mM and 3.43 showed IC50 value of 4.98 mM against tyrosinase. Although the purification processes were based on activity results, some of the pure compounds showed little inhibitory activity at the concentration of 0.5 mg/mL. They might not represent the inhibitory activity of the active fractions where they were isolated. This may be due to the fact that there were still minor compounds in the active fractions that may be responsible for the inhibitory activity. Based on the low concentration of the minor compounds, they were not isolated.

3 2.47 2.5 2 1.5 0.87 1 IC50 /mM 0.29 0.5 0.0344 0.18 0 positive positive (kojic Cytidine 1,3-DCQ Xanthiside (arbutin) acid) Compounds

Figure 3.25 Comparison of IC50 values of cytidine (3.44), 1,3-dicaffeoylquinic acid (3.37) and xanthiside (3.42) to kojic acid and arbutin.

IC50 values of the three active compounds were determined. As shown in Figure 3.25, compared to kojic acid (0.034 mM), the IC50 values of 3.44 and 3.37 were 0.18 mM and 0.29 mM respectively, almost 10 times of kojic acid. 3.42 was the least active compound and it had an IC50 value of 4.98 mM. However, compared to arbutin (0.87 mM), both 3.44 and 3.37 were more active. The comparisons suggested that the three pure compounds showed moderate inhibitory activity against tyrosinase. Given the objective of this project was to discover natural materials against hyperpigmentation, it was accepted that moderate inhibitors for tyrosinase may be better suited. The reason for that may be the complete inhibition of tyrosinase may cause severe side effects such as hypopigmentation and skin irritation.[44] 62

The demands for inhibitory efficiency of tyrosinase inhibitors are different due to various fields. Kojic acid has been used as positive control mainly for screening for tyrosinase inhibitors in food industries. As far as the food industry is concerned, food spoilage such as the browning of food and beverage from plant sources and melanosis of shrimps are primarily caused by the enzymatic oxidation of tyrosinase, which leads to the destruction of essential amino acids, disruption of digestibility and nutritional value of food.[45] The enzymatic oxidation may even result in formation of toxic compounds that brings economic loss to the food industry. Therefore, those IC50 values of tyrosinase inhibitors applied in food industries should be low enough that can completely inhibit the activity of tyrosinase in plants to reduce the economic loss. For instance, sulphite was once utilized widely in agriculture including marine products to prevent enzymatic browning as it possesses effective inhibitory activity against tyrosinase.[46] Nevertheless, FDA (Food and Drug Administration of the United States) prohibited the use of sulphite preparation in food additives due to health problems.[47] In the late 1980s, plenty of tyrosinase inhibitors from natural resources have been isolated while only few of them can be further introduced to produce skin-lightening agents or treating pigment-related skin diseases due to safety and stability. One of the positive controls used in this project, kojic acid, inhibits the activity of tyrosinase of different Aspergillus, neurospora crassa, mushrooms, some plants and crustaceans.[48- 50] In addition, kojic acid can effectively inhibit the formation of coloured products, which has been applied in cosmetics and even as a reference compound.[51, 52] According to the research on safety tests, kojic acid has been proved that can cause dermatitis or even skin cancer, which is the reason that some countries prohibit using kojic acid in cosmetics.[53, 54] Therefore, it is highly significant to hunt for natural and safe tyrosinase inhibitors applied in pigment-related diseases.

As illustrated in chapter 1, too much inhibition of human tyrosinase could cause hypopigmentation, which indicated that the inhibitory activity of tyrosinase applied in cosmetics or skin diseases ought to be moderate and without skin irritation. Based on the mass aesthetics, global women, especially Asian women, have considered pale complexion as a standard of beauty.[55] The huge demand of depigmenting products has attracted a large number of investigations on finding appropriate tyrosinase

63 inhibitors from natural products. For instance, arbutin, another positive control in this project, is without doubt the ingredients most often used in cosmetics industry for formulating skin-lightening products.[56] The use safety of alpha arbutin (for concentrations up to 2% in facial creams and 0.5% for body lotions) and of beta arbutin (for concentrations up to 7% in facial creams) were assessed by the Scientific Committee on Consumer Safety (SCCS).[57, 58] Another popular depigmenting agent is aloesin, which showed inhibitory activity against human tyrosinase with an IC50 value of 0.7 mM.[59] The efficacy of aloesin is proportional and it is evidently safe to apply on human skin.[60] Considering these results, the compounds using as depigmenting agents should have moderate inhibitory activity against tyrosinase in case of some side effects due to the over inhibition of tyrosinase. As far as the project is concerned, three active compounds with moderate inhibitory activity against tyrosinase are worth of further investigations as potential depigmenting agents. For future plans, these three compounds will be tested via cell-based assay and the inhibitory activity against human tyrosinase will be investigated.

3.4 Experimental General instrument HPLC were carried out on a Thermo Scientific UtiMate 3000 System with gradient

MeOH/H2O elution. A C18 column for crude (5m, 250 x 21.20 mm) and a C18 Luna column (5m, 250 x 10.00 mm) were used for semi-preparative HPLC chromatography.

NMR spectra were carried out on Brucker Avance HDX 800 MHz spectrometer under

25℃ and the solvent signals referenced were CD3OD-d4 (δH 3.31, δC 49.0) and DMSO-d6

1 13 (δH 2.50, δC 39.50). Both 1D- ( H and C) and 2D-NMR (COSY, HSQC, HMBC and ROESY) spectra were analysed through MestReNova (Version 11.0.2-18153).

LC-MS were carried out on Thermo UtiMate 3000 System with gradient MeOH/H2O elution. A C18 column (2.6m, 150 x 2.1 mm) was used.

Bioassay-guided extraction and isolation

TCM material was purchased from Hong Ren Tang Herbal, Sunnybank, Australia. The dry fruits of Xanthium Strumarium L. (98.5 g) were grounded and extracted by immersing in hexane on the shaker (2 x 600 mL, 2 h each). The residue was then extracted by DCM (2 x 600 mL, 2 h each) and methanol (2 x 600 mL, 2 h each) sequentially. The two extracts for each solvent were combined and evaporated to afford hexane extract, DCM extract and methanol extract respectively. The general steps of bioassay-guided extraction and isolation were shown in the flow chart below:

The hexane, DCM and methanol extracts were dissolved in DMSO then diluted in 0.1 M potassium phosphate buffer to make the final concentration of 1 mg/mL. The detailed process of tyrosinase assay was illustrated in the experimental part in Chapter 2. Briefly, 80 L freshly made L-tyrosine solution, 100 L hexane/DCM/methanol extract solution were mixed in a 96-well plate. Then 20 L mushroom tyrosinase solution was added to each mixture, which was then incubated for 35 min under 37 ℃. The absorbance of the microplate was obtained under 475nm UV wavelength using microplate reader. The calculation of inhibition percentage and the data analysis were also included in the experimental part in Chapter 2.

The methanol extract (10.003g) was dissolved in 2 mL methanol and applied on cotton and then fractionated by HPLC with C18 column for crude at a flow rate of 9 mL/min eluting with MeOH/H2O (5:95) (0.1% TFA) for 10 min, followed by a linear gradient to 50% MeOH (0.1% TFA) within 50min, then followed by a linear gradient to 100% MeOH 65

(0.1% TFA) within 30 min. Seven fractions were obtained based on the HPLC chromatogram: Fraction A: 1-5 min, Fraction B: 6-31 min, Fraction C: 32-39 min, Fraction D: 40-49 min, Fraction E: 50-69 min, Fraction F: 70-75 min, Fraction G: 76-90 min. The seven fractions were sampling and made into solutions at the concentration of 1 mg/mL with 1% DMSO, then tested against tyrosinase.

The dried fraction B was further purified by HPLC using C18 Luna column at a flow rate of 4 mL/min eluting with a linear gradient from MeOH/H2O (5:95) (0.1% TFA) to 24% MeOH within 30 min. Five fractions were obtained based on the HPLC chromatogram: Fraction B1: 1-6 min, Fraction B2: 7-10 min, Fraction B3: 11-12 min, Fraction B4: 13-14 min, Fraction B5: 15-30 min. The five fractions were sampling and made into solutions at the concentration of 0.5 mg/mL with 1% DMSO, then tested against tyrosinase.

The dried fraction B2 was further purified by HPLC using C18 Luna column at a flow rate of 4 mL/min eluting with a linear gradient from MeOH/H2O (7:93) (0.1% TFA) to 15% MeOH within 30 min. Subfraction 7 and 8 were relatively pure. Based on the NMR data, the two fractions were determined to be uridine (3.43) and cytidine (3.44).

The dried fraction B4 was further purified by HPLC using C18 Luna column at a flow rate of 4 mL/min eluting with a linear gradient from MeOH/H2O (15:82) (0.1% TFA) to 24% MeOH within 60 min. Subfraction 2 and 20 was rather pure and determined to be caffeoylcholine (3.41) and 4-caffeoylquinic acid (3.36) based on the NMR data.

The dried fraction C was further purified by HPLC using C18 Luna column at a flow rate of 4 mL/min eluting with a linear gradient from MeOH/H2O (25:75) (0.1% TFA) to 30% MeOH within 60 min. Five fractions were combined based on the peak distribution: Fraction C1: 1-11 min, Fraction C2: 12-22 min, Fraction C3: 23 min, Fraction C4: 24-25 min, Fraction C5: 26-60 min. The five fractions were sampling and made into solutions at the concentration of 0.5 mg/mL with 1% DMSO, then tested against tyrosinase. The subfraction 23 was rather pure and determined to be 3-caffeoylquinic acid (3.35).

The dried fraction D was further purified by HPLC using C18 Luna column at a flow rate of 4 mL/min eluting with a linear gradient from MeOH/H2O (31:69) (0.1% TFA) to 40% 66

MeOH within 60 min. Six fractions were combined based on the HPLC chromatogram: Fraction D1: 1-10 min, Fraction D2: 11-12 min, Fraction D3: 13-14 min, Fraction D4: 15- 24 min, Fraction D5: 25-34 min, Fraction D6: 35-60 min. The six fractions were sampling and made into solutions at the concentration of 0.5 mg/mL with 1% DMSO, then tested against tyrosinase. The subfraction 13 was rather pure and determined to be xanthiside (3.42).

The dried fraction D5 was further purified by HPLC using C18 Luna column at a flow rate of 4 mL/min eluting with a linear gradient from MeOH/H2O (35:65) (0.1% TFA) to 40% MeOH within 60 min. Five fractions were combined based on the HPLC chromatogram: Fraction D5-1: 1-4 min, Fraction D5-2: 5-22 min, Fraction D5-3: 23-25 min, Fraction D5-4: 26-30 min, Fraction D5-5: 31-60 min. The subfractions 23 and 29 was determined to be 1.3-dicaffeoylquinic acid (3.37) and 3.5-dicaffeoylquinic acid (3.38) respectively.

The dried fraction E was further purified by HPLC using C18 Luna column at a flow rate of 4 mL/min eluting with a linear gradient from MeOH/H2O (40:60) (0.1% TFA) to 65% MeOH within 25 min. Four fractions were combined based on the HPLC chromatogram: Fraction E1: 1-10 min, Fraction E2: 11-12 min, Fraction E3: 13-14 min, Fraction E4: 15-25 min. The subfractions 11 and 13 was determined to be 1.5- dicaffeoylquinic acid (3.39) and 4,5-dicaffeoylquinic acid (3.40) respectively.

To construct dose-response curves for the three active compounds, cytidine (0.0036, 0.018, 0.09, 0.15, 0.25, 0.45 and 0.9 mM respectively), 1,3-dicaffeoylquinic acid (0.05, 0.1, 0.2, 0.41, 0.68, 0.82 and 1.36 mM) and xanthiside (0.05, 0.1, 0.2, 0.41, 0.68, 0.82 and 1.36 mM), 7 different concentrations of solutions for each compound were freshly prepared. The detailed experimental was described in chapter 2.

1 13 3-caffeoylquinic acid (3.35): dark brown powder, H and C data (CD3OD-d4) see Table 3.14; (-)-LRESIMS at m/z 353 [M-H]-.

1 13 4-caffeoylquinic acid (3.36): light yellow powder, H and C data (CD3OD-d4) see Table 3.12; (-)-LRESIMS at m/z 353 [M-H]-. 67

1 13 1,3-dicaffeoylquinic acid (3.37): dark green powder, H and C data (CD3OD-d4) see Table 3.18; (+)-LRESIMS at m/z 517 [M+H]+.

1 13 3,5-dicaffeoylquinic acid (3.38): light green powder, H and C data (CD3OD-d4) see Table 3.19; (+)-LRESIMS m/z at 517 [M+H]+.

1 13 1,5-dicaffeoylquinic acid (3.39): light yellow powder, H and C data (CD3OD-d4) see Table 3.21; (+)-LRESIMS at m/z 517 [M+H]+.

1 13 4,5-dicaffeoylquinic acid (3.40): light green powder, H and C data (CD3OD-d4) see Table 3.22; (+)-LRESIMS at m/z 517 [M+H]+.

1 13 Caffeoylcholine (3.41): light yellow powder, H and C data (CD3OD-d4) see Table 3.11; (+)-LRESIMS at m/z 266 [M]+.

1 13 Xanthiside (3.42): light yellow powder, H and C data (CD3OD-d4) see Table 3.17; (-)- LRESIMS at m/z 400 [M-H]-.

1 13 Uridine (3.43): white powder, H and C data (DMSO-d6) see Table 3.9; (-)-LRESIMS at m/z 243 [M-H]-.

1 13 Cytidine (3.44): white powder, H and C data (DMSO-d6) see Table 3.10; (+)-LRESIMS at m/z 244 [M+H]+.

References 1. Fatiha, B., et al., Phenolic composition, in vitro antioxidant effects and tyrosinase inhibitory activity of three Algerian Mentha species: M. spicata (L.), M. pulegium (L.) and M. rotundifolia (L.) Huds (Lamiaceae). 2015. 74: p. 722- 730. 2. Li, H., et al., Inhibition of Melanogenesis by Xanthium strumarium L. 2012. 76(4): p. 767-771. 3. Peng, W., et al., Anti-allergic rhinitis effect of caffeoylxanthiazonoside isolated from fruits of Xanthium strumarium L. in rodent animals. 2014. 21(6): p. 824-829.

4. Kamboj, A. and A.K.J.I.J.o.G.P. Saluja, Phytopharmacological review of Xanthium strumarium L.(Cocklebur). 2010. 4(3). 5. Hong, S.-H., H.-J. Jeong, and H.-M.J.J.o.e. Kim, Inhibitory effects of Xanthii fructus extract on mast cell-mediated allergic reaction in murine model. 2003. 88(2-3): p. 229-234. 6. Hong, S.-H., et al., Processed Xanthii fructus increases cell viability of mast cell line, RBL-2H3. 2004. 4(1): p. 60-64. 7. Zhao, Y., et al., The effects of Fructus Xanthii extract on cytokine release from human mast cell line (HMC-1) and peripheral blood mononuclear cells. 2008. 30(3): p. 543-552. 8. Tao, L., et al., Concerted suppression of STAT3 and GSK3β is involved in growth inhibition of non-small cell lung cancer by xanthatin. 2013. 8(11): p. e81945. 9. Tao, L., et al., Xanthatin anti-tumor cytotoxicity is mediated via glycogen synthase kinase-3β and β-catenin. 2016. 115: p. 18-27. 10. Ramirez-Erosa, I., et al., Xanthatin and xanthinosin from the burs of Xanthium strumarium L. as potential anticancer agents. 2007. 85(11): p. 1160-1172. 11. woong Ahn, J., et al., Isolation of Cytotoxic Compounds from the Leaves of Xantbium strumarium L. 1995. 1(1): p. 1-4. 12. Bui, V.-B., et al., Sesquiterpene lactones from the aerial parts of Xanthium sibiricum and their cytotoxic effects on human cancer cell lines. 2012. 5(3): p. 685-689. 13. Ishwarya, S. and M.K.J.I.J.o.P.R. Singh, Evaluation of antioxidant and invitro cytotoxicity of crude fractions from the roots of Xanthiuim strumarium. 2010. 2(4): p. 2219-2223. 14. An, H.-J., et al., Xanthii fructus inhibits inflammatory responses in LPS- stimulated mouse peritoneal macrophages. 2004. 28(5): p. 263-270. 15. Kupiecki, F.P., C.D. Ogzewalla, and F.M.J.J.o.p.s. Schell, Isolation and characterization of a hypoglycemic agent from Xanthium strumarium. 1974. 63(7): p. 1166-1167. 16. Sridharamurthy, N., et al., In-vitro antioxidant and antilipidemic activities of Xanthium strumarium L. 2011. 5(3): p. 1362-1371. 17. Rad, J.S., et al., In-vitro antioxidant and antibacterial activities of Xanthium strumarium L. extracts on methicillin-susceptible and methicillin-resistant Staphylococcus aureus. 2013. 33(2): p. 109. 18. Lavault, M., et al., Antileishmanial and antifungal activities of xanthanolides isolated from Xanthium macrocarpum. 2005. 76(3-4): p. 363-366. 19. Han, C., et al., Indirect allelopathic effects of Xanthium italicum Morretti on soil properties and microbial communities. 2017. 41(2): p. 211-222. 20. Fan, W., et al., Traditional Uses, Botany, Phytochemistry, Pharmacology, Pharmacokinetics and Toxicology of Xanthium strumarium L.: A Review. 2019. 24(2): p. 359. 21. Jiang, H., et al., New phenylpropanoid derivatives from the fruits of Xanthium sibiricum and their anti-inflammatory activity. 2017. 117: p. 11-15. 22. Jiang, H., et al., HPLC-PDA Combined with Chemometrics for Quantitation of Active Components and Quality Assessment of Raw and Processed Fruits of Xanthium strumarium L. 2018. 23(2): p. 243. 23. Ismail, M.J.A.e., Extract of the plant Costus speciosus as a new acaricide for control of the two-spotted spider mite Tetranychus urticae Koch (Acari: Tetranychidae). 2017. 25(1): p. 148-156.

24. Kan, S., et al., Chemical constituents from the roots of Xanthium sibiricum. 2011. 25(13): p. 1243-1249. 25. Jiang, H., et al., Four new glycosides from the fruit of Xanthium sibiricum Patr. 2013. 18(10): p. 12464-12473. 26. Han, T., et al., New thiazinediones and other components from Xanthium strumarium. 2006. 42(5): p. 567-570. 27. Maas, M., F. Petereit, and A.J.M. Hensel, Caffeic acid derivatives from Eupatorium perfoliatum L. 2009. 14(1): p. 36-45. 28. Bajko, E., et al., 5-O-Caffeoylquinic acid: A spectroscopic study and biological screening for antimicrobial activity. 2016. 65: p. 471-479. 29. Albergamo, A., et al., Metabolite and mineral profiling of “Violetto di Niscemi” and “Spinoso di Menfi” globe artichokes by 1H-NMR and ICP-MS. 2017. 31(9): p. 990-999. 30. Clifford, M.N., et al., Discriminating between the six isomers of dicaffeoylquinic acid by LC-MS n. 2005. 53(10): p. 3821-3832. 31. Schram, K., et al., Mass spectrometry of 1, 3‐and 1, 5‐dicaffeoylquinic acids. 2004. 39(4): p. 384-395. 32. Pagani, F.J.P.m., Caffeic acid choline ester: a new plant constituent from Reseda luteola. 1983. 48(06): p. 128-128. 33. Mahmoud, A.A., et al., A new heterocyclic glucoside from the fruits of Xanthium pungens. 2005. 19(6): p. 585-589. 34. Granqvist, L. and P.J.T.J.o.o.c. Virta, 2′-O-[(4-CF3-triazol-1-yl) methyl] Uridine–A Sensitive 19F NMR Sensor for the Detection of RNA Secondary Structures. 2015. 80(16): p. 7961-7970. 35. Tsanakopoulou, M., et al., A new and potentially prebiotic α-cytidine derivative. 2017. 53(23): p. 3327-3329. 36. Li, M., et al., Uridine derivatives from the seeds of Lepidium apetalum Willd. and their estrogenic effects. 2018. 155: p. 45-52. 37. Uwai, K., et al., Inhibitory effect of the alkyl side chain of caffeic acid analogues on lipopolysaccharide-induced nitric oxide production in RAW264. 7 macrophages. 2008. 16(16): p. 7795-7803. 38. Sefkow, M., A. Kelling, and U.J.E.J.o.O.C. Schilde, First efficient syntheses of 1‐, 4‐, and 5‐caffeoylquinic acid. 2001. 2001(14): p. 2735-2742. 39. Catarino, M.D., et al., Characterization of phenolic constituents and evaluation of antioxidant properties of leaves and stems of Eriocephalus africanus. 2018. 11(1): p. 62-69. 40. Dai, Y.-H., et al., A new thiaziedione from the fruits of Xanthium sibiricum. 2008. 10(4): p. 303-305. 41. Wu, Q.-z., et al., Antitussive, expectorant, and anti-inflammatory activities of four caffeoylquinic acids isolated from Tussilago farfara. 2016. 54(7): p. 1117-1124. 42. Hong, S., et al., Antioxidant and anti-inflammatory activities of 3, 5- dicaffeoylquinic acid isolated from Ligularia fischeri leaves. 2015. 24(1): p. 257-263. 43. Chen, J., et al., Caffeoylquinic acid derivatives isolated from the aerial parts of Gynura divaricata and their yeast α-glucosidase and PTP1B inhibitory activity. 2014. 99: p. 1-6. 44. Nicolaidou, E. and A.D.J.C.i.d. Katsambas, Pigmentation disorders: hyperpigmentation and hypopigmentation. 2014. 32(1): p. 66-72.

45. Gonçalves, A.A., A.R.M.J.L.-F.S. de Oliveira, and Technology, Melanosis in crustaceans: A review. 2016. 65: p. 791-799. 46. He, Q. and Y.J.S.P.R. Luo, Enzymatic browning and its control in fresh-cut produce. 2007. 3(6): p. 1-7. 47. Lester, M.R.J.J.o.t.A.C.o.N., Sulfite sensitivity: significance in human health. 1995. 14(3): p. 229-232. 48. Kahn, V., et al., Effect of kojic acid on the oxidation of N-acetyldopamine by mushroom tyrosinase. 1997. 45(11): p. 4460-4465. 49. Whitaker, J.R., C.Y.J.E.b. Lee, and i. prevention, Recent advances in chemistry of enzymatic browning. 1995. 600: p. 2-7. 50. Chen, J.S., et al., Inhibitory effect of kojic acid on some plant and crustacean polyphenol oxidases. 1991. 39(8): p. 1396-1401. 51. İyidoǧan, N. and A.J.J.o.F.E. Bayındırlı, Effect of L-cysteine, kojic acid and 4- hexylresorcinol combination on inhibition of enzymatic browning in Amasya apple juice. 2004. 62(3): p. 299-304. 52. Crespo, M.I., et al., Inhibitory effects of compounds isolated from Lepechinia meyenii on tyrosinase. 2019. 53. Gillbro, J. and M.J.I.j.o.c.s. Olsson, The melanogenesis and mechanisms of skin‐lightening agents–existing and new approaches. 2011. 33(3): p. 210- 221. 54. Saeedi, M., et al., Kojic acid applications in cosmetic and pharmaceutical preparations. 2019. 110: p. 582-593. 55. Porcheron, A., et al., Influence of skin ageing features on Chinese women's perception of facial age and attractiveness. 2014. 36(4): p. 312-320. 56. Lukas, B., et al., Arbutin in marjoram and oregano. 2010. 121(1): p. 185-190. 57. Sccs and G.H. Degen, Opinion of the Scientific Committee on Consumer safety (SCCS)--Opinion on the safety of the use of alpha-arbutin in cosmetic products. Regul Toxicol Pharmacol, 2016. 74: p. 75-6. 58. Degen, G.H., Opinion of the Scientific Committee on Consumer Safety (SCCS)-- Opinion on the safety of the use of beta-arbutin in cosmetic products. Regul Toxicol Pharmacol, 2015. 73(3): p. 866-7. 59. Nesterov, A., J. Zhao, and Q.J.D.F. Jia, Natural tyrosinase inhibitors for skin hyperpigmentation. 2008. 33: p. 945-954. 60. Yimam, M., et al., In vivo safety evaluation of UP780, a standardized composition of aloe chromone aloesin formulated with an Aloe vera inner leaf fillet. 2014. 69(3): p. 390-397.

Appendix

Supporting data for 10 pure compounds (1H, COSY, HSQC, HMBC spectra and MS data)

3-caffeoylquinic acid (3.35)

Figure A1.1 1H NMR spectra for 3-caffeoylquinic acid (3.35)

Figure A1.2 COSY spectra of 3-caffeoylquinic acid (3.35)

Figure A1.3 HSQC spectra of 3-caffeoylquinic acid (3.35)

Figure A1.4 HMBC spectra of 3-caffeoylquinic acid (3.35)

Figure A1.5 ROESY spectra of 3-caffeoylquinic acid (3.35)

Figure A1.6 Mass spectrum of 3-caffeoylquinic acid (3.35)

4-caffeoylquinic acid (3.36)

Figure A2.1 1H NMR spectra for 4-caffeoylquinic acid (3.36)

Figure A2.2 COSY spectra of 4-caffeoylquinic acid (3.36)

Figure A2.3 HSQC spectra of 4-caffeoylquinic acid (3.36)

Figure A2.4 HMBC spectra of 4-caffeoylquinic acid (3.36)

Figure A2.5 Mass spectrum of 4-caffeoylquinic acid (3.36)

1,3-dicaffeoylquinic aicd (3.37)

Figure A3.1 1H NMR spectra of 1,3-dicaffeoylquinic aicd (3.37)

Figure A3.2 COSY spectra of 1,3-dicaffeoylquinic aicd (3.37)

Figure A3.3 HSQC spectra of 1,3-dicaffeoylquinic aicd (3.37)

Figure A3.4 HMBC spectra of 1,3-dicaffeoylquinic aicd (3.37)

Figure A3.5 Mass spectrum of 1,3-dicaffeoylquinic aicd (3.37)

3,5-dicaffeoylquinic acid (3.38)

Figure A4.1 1H NMR spectra of 3,5-dicaffeoylquinic acid (3.38)

Figure A4.2 COSY spectra of 3,5-dicaffeoylquinic acid (3.38)

Figure A4.3 HSQC spectra of 3,5-dicaffeoylquinic acid (3.38)

Figure A4.4 HMBC spectra of 3,5-dicaffeoylquinic acid (3.38)

Figure A4.5 Mass spectrum of 3,5-dicaffeoylquinic acid (3.38)

1,5-dicaffeoylquinic aicd (3.39)

Figure A5.1 1H NMR spectra of 1,5-dicaffeoylquinic aicd (3.39)

Figure A5.2 COSY spectra of 1,5-dicaffeoylquinic aicd (3.39)

Figure A5.3 HSQC spectra of 1,5-dicaffeoylquinic aicd (3.39)

Figure A5.4 HMBC spectra of 1,5-dicaffeoylquinic aicd (3.39)

Figure A5.5 Mass spectrum of 1,5-dicaffeoylquinic aicd (3.39)

4,5-dicaffeoylquinic acid (3.40)

Figure A6.1 1H NMR spectra of 4,5-dicaffeoylquinic acid (3.40)

Figure A6.2 COSY spectra of 4,5-dicaffeoylquinic acid (3.40)

Figure A6.3 HSQC spectra of 4,5-dicaffeoylquinic acid (3.40)

Figure A6.4 HMBC spectra of 4,5-dicaffeoylquinic acid (3.40)

Figure A6.5 Mass spectrum of 4,5-dicaffeoylquinic acid (3.40)

Caffeoylcholine (3.41)

Figure A7.1 1H NMR spectra of Caffeoylcholine (3.41)

Figure A7.2 COSY spectra of Caffeoylcholine (3.41)

Figure A7.3 HSQC spectra of Caffeoylcholine (3.41)

Figure A7.4 HMBC spectra of Caffeoylcholine (3.41)

Figure A7.5 Mass spectrum of Caffeoylcholine (3.41)

Xanthiside (3.42)

Figure A8.1 1H NMR spectra of Xanthiside (3.42)

Figure A8.2 COSY spectra of Xanthiside (3.42)

Figure A8.3 HSQC spectra of Xanthiside (3.42)

Figure A8.4 HMBC spectra of Xanthiside (3.42)

Figure A8.5 Mass spectrum of Xanthiside (3.42)

Uridine (3.43)

Figure A9.1 1H NMR spectra of Uridine (3.43)

Figure A9.2 COSY spectra of Uridine (3.43)

Figure A9.3 HSQC spectra of Uridine (3.43)

Figure A9.4 HMBC of Uridine (3.43)

Figure A9.5 ROESY spectra of Uridine (3.43)

Figure A9.6 Mass spectrum of Uridine (3.43)

Cytidine (3.44)

Figure A10.1 1H NMR spectra of Cytidine (3.44)

Figure A10.2 1H NMR spectra of Cytidine (3.44)

Figure A10.3 HSQC spectra of Cytidine (3.44)

Figure A10.4 HMBC spectra of Cytidine (3.44)

Figure A10.5 ROESY spectra of Cytidine (3.44)

Figure A10.6 Mass spectrum of Cytidine (3.44)