Identifying Bioactive Phytochemicals in Spent Coffee Grounds for Cosmetic Application Through Global Metabolite Analysis

Home , Astragalin, Flavanonol, Formononetin, Genistin, Isorhamnetin, Karlina

IDENTIFYING BIOACTIVE PHYTOCHEMICALS IN SPENT COFFEE GROUNDS FOR COSMETIC APPLICATION THROUGH GLOBAL METABOLITE ANALYSIS

______

A Thesis Presented to the Faculty of the Graduate School University of Missouri-Columbia

______

In Partial Fulfillment of the Requirements for the Degree

Master of Science

______

JIHYUN PARK

Dr. Chung-Ho Lin, Thesis Supervisor

DECEMBER 2017

The undersigned, appointed by the dean of the Graduate School,

have examined the thesis entitled:

IDENTIFYING BIOACTIVE PHYTOCHEMICALS IN SPENT COFFEE GROUNDS FOR COSMETIC APPLICATION THROUGH GLOBAL METABOLITE ANALYSIS

presented by

JIHYUN PARK

a candidate for the degree of Master of Science and hereby

certify that in their opinion it is worthy of acceptance

Professor Chung-Ho Lin

Professor Shibu Jose

Professor Gary Stacey

Professor Minviluz Stacey

ACKNOWLEDGEMENTS

I would like to thank my advisor, Professor Chung-Ho Lin about his guidance with huge effort, and the Center for Agroforestry, without their considerable assistant, I could not have done this research. Conducting this project provided me a great opportunity to meet and work with many people who have a comprehensive mind and passion. It is very pleasure for me to offer thanks to them.

In addition, my thesis committee members, Dr. Shibu Jose, Gary Stacey, and Bing Stacey, were the most responsible professors in their field and they leading me to focus on this research. They realized the novelty of my thesis topics paying attention to this project and encouraging me to accomplish the experiment on time.

Also I would like to say thanks to Nahom Taddese Ghile who gave me the first help beginning this project and the laboratory members Van Ho, Danh Vu, Phuc Vo, who provided assistance in several steps of my experiment with familiarity.

Special thanks to many capable professionals Zhentian Lei, Lloyd Summer, Michael Greenlief, Brian P. Mooney in metabolomics center and department of chemistry.

Lastly I would like to thank to my family and friends who continuously supported me with their love.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... ii LIST OF TABLES ...... v LIST OF FIGURES ...... vi Chapter 1. Identifying the bioactive compounds in SCG through global metabolomics analysis (XCMS) ...... 1 1. Introduction ...... 2 2. Objectives ...... 11 3. Materials and Methods ...... 11 3.1. Sample preparation ...... 12 3.2. LCMS Analysis ...... 12 3.3 Global metabolite profiling using XCMS and METLIN Library ...... 15 4. Results ...... 18 4.1. Multi-group Comparison between three SCG cultivars using low resolution data ...... 19 4.2. Identified bioactive compounds through global metabolomics analysis using high resolution data ...... 28 5. Discussion ...... 42 5.1. Global metabolomics approach (XCMS Online) vs traditional analytical techniques ...... 42 5.2. Phenolic acids and caffeine in SCG ...... 45 5.3. Glycosylated form of flavonoids and their activity ...... 47 5.4. Flavonoids (flavonols, flavones, isoflavones, flavanones, anthocyanidins, and flavanols) in SCG for cosmetic application ...... 48 6. Conclusion ...... 54 Chapter 2. Isolation of the antioxidant and antimicrobial compounds with bioassay-guided fractionation ...... 56 1. Introduction ...... 57 2. Objectives ...... 67 3. Materials and Methods ...... 67 3.1. Explore new application (antioxidant and antimicrobial activities) through bioassay-guided purification ...... 70 3.2. Determination of anti-oxidant activities ...... 73

iii

3.3. Anti-bacterial analysis ...... 73 3.4. High resolution (HR) metabolomics analysis...... 74 3.5. Statistical analysis ...... 76 4. Results ...... 76 4.1 Antioxidant study ...... 76 4.2 Antibacterial activity against Methicillin-resistant Staphylococcus aureus (MRSA) ...... 89 5. Discussion ...... 93 5.1. Antioxidant activity of SCG for cosmetic application ...... 93 5.2. Antimicrobial activity of SCG for cosmetic application ...... 96 5.3. Importance of anti-inflammatory and anti-carcinogenesis activity in cosmetics 99 5.4. High resolution metabolomics analysis followed by bioassay guided purification ...... 100 6. Conclusion ...... 101 References ...... 103 Appendix A...... 123 Appendix B...... 147 1. Introduction ...... 147 2. Objectives ...... 150 3. Materials and Methods ...... 150 4. Results and discussion ...... 151

LIST OF TABLES

Table Page

1.1. Major compounds exist in natural resources that have biological function related to skin care. The list of bioactive compounds have been reported, showing antioxidant, Anti-fungal, skin-whitening, anti-inflammatory, anti-aging, and antimicrobial properties or other skin-care applications……………………………………………………...... 8

2.1. Antioxidant activity of SCG methanol extracts of three cultivars including Ethiopian Yirgacheffe, Costa Rican Tarrazu, and Hawaiian Kona ...... 79

2.2. Identified compounds in purified fraction 14 through XCMS Online HR metabolomics analysis. Each row indicates feature, compound name, m/z, retention time, maximum intensity and its known bioactivity. Using XCMS Online platform, chemical profiling was compiled...... 85

2.3. FeCl2 concentration (µM) of bioactive compounds which were reported in previous study (unpublished data by Nahom Taddese Ghile and Lin) ...... 88

Appendix A. Phenolic acids, flavonoids, and other bioactive compounds including coffee alkaloids, phenylpropanoids, stilbenoids, and anthraquinones extracted from three different coffee wastes (Ethiopian Yirgacheffe, Costa Rican Tarrazu, and Hawaiian Kona). Maximum intensity of peaks detected by XCMS online comparing the relative intensity between the three cultivars of SCG are shown ...... 123

LIST OF FIGURES Figure Page

1.1. Bioactive compounds related to UV protection efficacy...... 7

1.2. Flow chart of the SCG metabolomics approach project using Waters Xevo UPLC- MS for non-VOC analysis and XCMS global metabolomics platform...... 14

1.3. Flow chart of metabolomics data processing………………………………………..15

1.4a. Box-plot generated from interactive cloud plot ...... 21

1.4b. Composite data base after mouse click on interactive cloud plot. Seven compounds were matched with METLIN database based on ion chromatogram of each circle. Compounds in composite DB linked to METLIN library it will show compound formula, structure, and spectrum ...... 22

1.4c. Multi-group cloud plot. Metabolite feature which varies significantly (p < 0.01) from each groups are presented on cloud plot depending on their retention time (x-axis) and m/z (y-axis). Each circle indicates each compound feature. The color intensity of circle represents statistical significance (p-value). The size of the circle indicates feature intensity. If you move the mouse on each circle, p-value, q-value, m/z, RT, Max intensity will pop-up. With mouse click each circle, Box-plot, EIC, Post-HOC, and Composite database (METLIN) are displayed under the cloud plot. Compounds in composite DB are linked to METLIN library showing compound formula, structure, and spectrum...... 23

1.5a, Interactive principal component analysis. A Scores plot shows the correlation between submitted samples. Three clusters represent each SCG cultivars which includes Hawaiian Kona (ESI(-)_K1), Costa Rican Tarrazu (ESI(-)_T1, and Ethiopian Yirgacheffe (ESI(-)_Y1) ...... 25

1.5b. Principal Component Analysis (PCA) results: SCG samples from three different coffee cultivars (Ethiopian Yirgacheffe, Costa Rican Tarrazu, Hawaiian Kona) clustered by each groups. Individuals are projected in a reduced dimension space defined by component 1 (x-axis) and component 2 (y-axis). Each color corresponds to a specific cluster...... 26

1.5c. Non-metric multi-dimensional Scaling of different coffee cultivars (Ethiopian Yirgacheffe, Costa Rica Tarrazu, and Hawaiian Kona) shows that the chemical profiling is significantly different between three cultivars………………………………………...27

1.6. Chemical structures of phenolic acid...... 29

1.7a. Extracted ion chromatogram of Ethiopian Yirgacheffe (E) and Hawaiian Kona (H). The peak represents relative intensity of quercetin pentaacetate in SCG from Yirgacheffe (E), and Kona (H) ...... 39

1.7b. Extracted ion chromatogram of Yrigacheffe (E) and Costa Rican Tarrazu (C). The peak represents relative intensity of candidate compounds such as cyanidin-3-(6'- malonylglucoside) or apigenin 7-O-(6-O-malonylglucoside) at RT 34.86 ...... 40

1.8. Cloud Plot results: A Cloud Plot displays features whose intensities are altered between sample groups. Up-regulated features are represented as circles on the top of the plot and down-regulated features are represented as circles on the bottom of the plot, where the size corresponds to the (log) fold change of the feature. Circles with black outlines indicate hits in the METLIN database. The lighter or darker the color of the circle relates directly to the significance of the p-value. 46 features were generated by pairwise comparison between Ethiopian Yirgacheffe and Costa Rican Tarrazu showing the relationship between retention time and m/z. Green color represents upregulated compounds which means abundant in Tarrazu and red color circle represents down regulated compounds which means the compounds exist more in Yirgacheffe. With a mouse click, EIC, Box and Whisker, and Spectrum appears under the cloud plot ...... 41

2.1. Common bioactive compounds in SCG ...... 66

2.2. Flow chart of the SCG project using bioassay guided purification………………....69

2.3. Flow chart of the bioassay-guided purification for identifying antioxidant activity and antimicrobial assay. Same approach was applied for antimicrobial test...... 72

2.4. Standard curve of FRAP assay. Each dot in X axis represents around 0, 31.25, 62.5, 125, 250, 500, 1000 FeCl2 concentration of standards (STD; n = 7)…………………….78

2.5. Reducing power of methanol extract (10mg/mL) from three SCG cultivars including Ethiopian Yirgacheffe, Costa Rican Tarrazu, and Hawaiian Kona were compared. Concentration of FeCl2 was calculated from FRAP value (absorbance in 560nm) using y = 0.002x - 0.0054. Hawaiian Kona had highest antioxidant compared to other cultivars. Means in the same column followed by different letters (A-B) represent statistical significance (P <0.05). Sextuplicates were analyzed using one-way analysis of variance (ANOVA) followed by pairwise comparisons of means using Tukey’s range test was performed to assess the difference in three cultivars (significance level P < 0.05). R: The R Project for Statistical Computing 3.3.1 was used for statistical computing...... 80

2.6. Antioxidant activities of the fractions from flash chromatography (1:10 dilution). The fraction number 4 showed highest antioxidant activity as 475.16 µM FeCl2 condition among 43 fractions so F4 was injected into HPLC for further purification...... 81

2.7. 26 fractions was obtained from HPLC and tested through FRAP assay. Fraction 14 (vial 12) was revealed as bioactive fractions with high antioxidant activity ...... 82

2.8. Total ion chromatogram (TIC) of fraction 14 purified by HPLC. The TIC was generated by XCMS Online positive ion mode. The highest peak intensity 618,624 was shown at RT 11.83 with m/z 679.9 and cafestol was the best possibility among other compound candidates ...... 86

vii

2.9. SCG samples (Number from 47 to 55) had strong anti-microbial activity which is Ethiopian Yirgacheffe, Costa Rican Tarrazu, Hawaiian Kona. The size of inhibition zone was measured three times for all samples. Among three cultivars, Kona showed the highest anti-microbial activity having 8.85 mm inhibition zone...... 90

2.10. Antimicrobial activities of the E (Ethiopian Yirgacheffe), C (Costa Rican Tarrazu), H (Hawaiian Kona). Using zone of inhibition assay for assessing anti-microbial activity against S. aureus, we found the following order of efficiency among coffee Arabica cultivars: Kona (8.85) > Yirgacheffe (8.19) >> Tarrazu (5.76). DMSO did not show any inhibitory activity. Inhibition zone: 8 mm or greater — good antibacterial activity; 6–7 mm — moderate antibacterial activity; 4–5 mm — weak antibacterial activity; 2–3 — very weak antibacterial activity. Mean values followed by different letters (A and B) represent statistical different (P <0.05). Zone of inhibition is expressed as the diameter (mm). Values represent means ± standard deviations of inhibition zones from experiments, each using triplicate plates and was measure three times. Means with the same superscripts (A, and B) are not significantly different (P < 0.05) from each other. One-way analysis of variance (ANOVA) followed by pairwise comparisons of means was performed to assess the difference in three cultivars (significance level P < 0.05). R: The R Project for Statistical Computing 3.3.1 was used………………………………..91

2.11. Antimicrobial activity of 43 fractions using flash chromatography from Ethiopian Yirgacheffe. Fraction number 7 and 12 showed highest zone of inhibition value. The rest of fraction number over 20 that have no activity doesn’t represent in the chart ...... 92

Appendix B1. Mode of action of anti-pigmentation ...... 148

Appendix B2. Melanin synthesis pathway...... 149

Appendix B3. Calibration curve of the positive control (arbutin). Each dot in X axis represents around 0, 1.1, 11, 110, 1100 µg/ml concentration of arbutin (STD; n = 5) ....152

viii

Chapter 1. Identifying the bioactive compounds in SCG through global metabolomics analysis (XCMS)

Abstract: Annually, more than 6 million tons of spent coffee grounds (SCG) are generated worldwide. The present study explores the possible use of spent coffee grounds as the raw materials for cosmetics industry. The main objective of this project are to investigate the chemical profiles and identify the bioactive compounds for cosmetics application through global metabolite analysis. The compounds extracted from SCG of

Ethiopia coffee (Yirgacheffe), Costa Rican coffee (Tarrazu) and Hawaiian coffee (Kona

Blend) were analyzed by ultra-high pressure liquid chromatography coupled with mass spectrometry (UPLC-MS). The ion chromatograms were submitted to XCMS platform operated by Center for Metabolomics at the Scripps Research Institute. The peak detection, peak grouping, spectra extraction, and retention alignment were processed by

XCMS. The spectra were annotated and the compounds were identified and categorized by integration with METLIN, the world's largest metabolite database. Multivariate and univariate statistical analysis including PCA and cloud-plot were performed by XCMS to compare the chemical profiles between the three coffee cultivars. These analyses indicated that each cultivar showed a specific cluster. Over 300 compounds related to anti-oxidant, anti-inflammatory, anti-tyrosinase and anti-tumor for skin care application were identified by XCMS. Therefore, the presence of bioactive compounds in SCG makes it a potential source of raw material for cosmetic application (e.g., anti-oxidant, anti-inflammatory, skin-whiting, and anti-aging).

1. Introduction

In USA alone, more than 4000 tons of coffee are consumed every day. Annually, more than 6 million tons of spent coffee grounds (SCG) are generated worldwide (Mussatto et al., 2011b). Conventionally, the uses of the SCG has been only limited to home gardening

(Cruz et al., 2012), compost (Adi and Noor, 2009) and landfills. Recent studies have suggested that SCG could be an excellent source of high value phenolic compounds which have wide applications in personal care products (Mussatto et al., 2011a). For examples, commercialized extracts which contain chlorogenic acids, major compounds in

SCG, with proanthocyanidins, quinic acids, and ferulic acid have shown positive results for facial skin care purpose due to their antioxidant, anti-aging, anti-inflammatory activities (Esquivel and Jiménez, 2012). Several compounds such as phenolic acids that have shown strong anti-microbial effects were also found in the extracts (Table 1.1).

Table 1.1 shows variety of major compounds in extracts reported in the previous studies that have biological function related to skin care. Among these, SCG contain large amounts of phenolic acids such as caffeoylquinic acids, dicaffeoylquinic acids, feruloylquinic acids, and p-coumaroylquinic acids (Zuorro and Lavecchia, 2012) showing strong antioxidant, anti-fungal, skin-whitening, anti-aging, and antimicrobial properties or other skin-care applications. Especially, antioxidant reaction, as a result of free radical scavenging activity, was attributed to the presence of these phenolic compounds in SCG.

This is because the brown pigments such as phenolic polymers generated from roasting process protect cells from oxidative damage (Ramalakshmi et al., 2009). Additionally, several flavonoids in SCG are known to be responsible for the antioxidant activity

(Mussatto et al., 2011a). Figure 1.1 shows flavonoids such as cyanidin, resveratrol,

2 quercetin, apigenin, hesperetin with strong antioxidant effects against UV for reduced skin damage. These flavonoids also possess anti-aging and anti-melanogenesis efficacy as well (Arung et al., 2011; Kitagawa et al., 2011; Wang et al., 1999). In addition to phenolic acids and flavonoids, caffeine, one of the major bioactive alkaloid in SCG contributes significantly in reducing interfacial tension between water and oil phase, important in defining the emollient characteristics of cosmetics.(Campos-Vega et al.,

2015). Likewise, galactomannans acts as excellent stiffeners and emulsion stabilizers in cosmetics as a crude flour form (Silveira and Bresolin, 2011). This polysaccharide mostly remains as insoluble material bound to the SCG matrix during coffee brewing (Campos-

Vega et al., 2015). Due to considerable amount of these bioactive compounds, it provides an excellent opportunity to utilize the organic-based SCG as a potential valuable raw material in cosmetics industry. Although several cosmetics applications of SCG have been reported and identified, the bioactive compounds in SCG have never been systematically studied or characterized. To date, most of the bioactive compounds were isolated and identified through the conventional labor-intensive and time-consuming bioassay-guided fractionation followed by targeted analysis and NMR structural elucidation.

Recent advances of metabolomics science, mass spectrometry, computation capacity, and compound libraries have allowed systematic identification and characterization of bioactive phytochemicals. Over the last decades, untargeted global metabolomics analysis has been utilized as a powerful tool for natural product research through processing raw spectra information acquired from mass spectrometry such as liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass

3 spectrometry (GC-MS). Currently, several freely available or commercial metabolomics software platforms such as Metab, Mzmatch, MZmine and XCMS have emerged.

However, each software platform has some technical limitations due to installation requirement, or requirement of command line interface which demand advanced computer skills, or lack of the data analytical tools in the package. For example, both

Metab and Mzmatch do not provide any statistics analytical tools, such as principal component analysis (PCA) in the software package, and therefore users may need additional statistical analysis. For the accurate mass LC/MS data, XCMS shows higher detection percentages than MZmine (Coble and Fraga, 2014). The technical challenges of several platforms and user unfriendly interfaces often require scientists or clinicians collaborate closely with bioinformatics laboratory for data processing (Tautenhahn et al.,

2012). One way to solve these problems is the development of user-friendly interfaces with strong analytic algorithms and comprehensive libraries and database.

As a solution, XCMS Online, an open sourced and web-based metabolomics platform was launched with powerful analyzing algorithms and comprehensive MS and MS/MS metabolite database through improvement of XCMS R packaging version. Newly developed online platform allows easy file uploading, parameter setting, and result interpretation (Tautenhahn et al., 2012). As a result, XCMS Online has been known as the “World’s most cited metabolomics software” cited in literature over 1000 times (Ubhi et al, 2014). It has detection algorithm, quality assessment, metabolite identification

(Wen et al., 2017) coupled with METLIN database.

Compared to other platforms, XCMS Online offers a more comprehensive package that integrates the signal deconvolution, library matching, and statically analysis

4 including interactive cloud plot (univariate) and interactive PCA (multivariate). XCMS

Online includes peak detection, profile alignment, feature annotation, and exploratory statistical analyses (Libiseller et al., 2015) with customized parameter setting. According to its peak detection algorithm, it covers both low-resolution, and high-resolution using matchedFilter method and centWave method respectively (Smith, 2013). Interactive platform can monitor the statistical output of both univariate (cloud plot) and multivariate

(PCA) in real time by adjusting various parameters (Gowda et al., 2014). For identifying distinct metabolites between two classes, it considers which intensities are significantly different from sample class to sample class and rank the metabolites by p-value (Smith et al., 2006). These statistical results are valuable when researchers are looking for distinct metabolites between two treatments (basically control versus sample) or visualized cluster between treatments. The seemly integration of METLIN library into XCMS allows rapid identifying list of candidates of compounds for each extracted ion chromatogram based on accurate mass (Smith et al., 2006). The METLIN and METLIN mobile now include more than 961,820 molecules providing information for each compound such as name, systematic name, structure, elemental formula, mass, CAS number, KEGG ID and link, HMDB ID and link, PubChem ID and link, commercial availability and direct search options on the molecule itself (https://metlin.scripps.edu/).

With the development of non-targeted metabolomics algorithm and computational capacity, analytical instrument such as high resolution mass spectrometry enable high- throughput chemical profiling in crude samples for natural product research. Compare to targeted analysis, non-targeted analysis requires high resolution mass spectrometer

(HRMS) to screen (Un)known chemicals in the samples (Ruff et al., 2015). Several

5 concerns about non-targeted analysis due to massive amount of data processing, are being solved by recent advancement in computation power, deconvolutions software algorithms and the design of mass spectrometers. Since HRMS can measure molecular weights more accurately with several decimal, compounds with similar mass can be distinguished by

HRMS that provides precision of molecular formula to be narrowed down to only a few possibilities. There are several mass spectrometers that have been widely used for untargeted analysis; time-of-flight (TOF) or orbitrap (Ruff et al., 2015). In the interest of screening unknown or less common compounds, liquid chromatography coupled to mass spectrometry (LC-MS) with time-of-flight (TOF) analyzers (TOF-MS or QTOF-MS) is most suitable, due to its accurate mass analysis, resolving power, enhanced selectivity and high sensitivity (García-Reyes et al., 2007). A large number of targets could be screened simultaneously without sensitivity loss and unknown peaks can be identified through accurate mass evaluation. Besides TOF-MS having 5000 FWHM (term of mass- resolving power) (Petrovic et al., 2006), Orbitrap-MS is the most recent instrument for

HRMS-based non-target screening providing much higher mass-resolving with 280,000

FWHM (Kaufmann, 2014).

These innovational technologies especially XCMS Online metabolomics platform could be extremely helpful to reduce time spent in processing the huge data sets for health-promoting compound identification in each SCG cultivars and the SCG cultivar comparison analysis in our study. These qualitative and quantitative analysis could quickly help translate the benchtop experiments into real-world practical application in the cosmetic industry.

1. Cyanidin 2. Resveratrol 3. Quercetin

4. Apigenin 5. Hesperetin 6. 3,5-Di-O-caffeoylquinic acid

Figure 1.1. Bioactive compounds related to UV protection efficacy

1 Cyanidin Anti-oxidant, Anti-inflammatory, Anti-cancer, Skin protection, anti-aging

2 Resveratrol Anti-inflammatory, antioxidant, anti-aging, photo protection

3 Quercetin Anti-bacterial, anti-oxidant, skin protection, Anti-melanogenesis

4 Apigenin Anti-inflammatory, antioxidant and anticancer, UVA/UVB-induced skin carcinogenesis

5 Hesperetin Photoprotective, Anti-inflammatory, antioxidant

6 3,5-Di-O-caffeoylquinic acid Antioxidant, anti-tyrosinase

Table 1.1. Major compounds exist in natural resources that have biological function related to skin care.

The list of bioactive compounds have been reported, showing antioxidant, Anti-fungal, skin-whitening, anti-inflammatory, anti-aging, and antimicrobial properties or other skin-care applications.

Compound name Biological function 1 Quercetin1,2 Anti-bacterial, anti-oxidant, skin protection, anti- melanogenesis 2 Resveratrol2,3,16 Anti-inflammatory, antioxidant, anti-aging, photo protection 3 Cyanidin4,5 Antioxidant, anti-inflammatory, anti-cancer, skin protection, anti-aging 4 Apigenin6,7 anti-inflammatory, antioxidant and anticancer, UVA/UVB-induced skin carcinogenesis 5 Curcumin8,9 Anti-inflammatory, photo-protective, anti-cancer 6 Delphinidin10,11 Anti-proliferative, anti-inflammatory, antioxidant, anti- mutagenic, anti-angiogenic 7 Malvidin12 Antioxidant, anti-inflammatory 8 Pelargonidin13, 14 Anti-inflammatory, antioxidant 9 Isorhamnetin14 Anti-inflammatory, anticancer 10 Kaempferol14 Anti-inflammatory ,hypoglycemic and antioxidant 11 Daidzein14 Anti-inflammatory, anti-carcinogenic, anti-inflammatory 12 Genistein14 Anti-inflammatory, antioxidant and anticancer 13 Caffeic acid15 Antioxidant 14 Chlorogenic Anti-inflammatory, antioxidant, anti-aging, anti-skin acid15,16,41 cancer, antimicrobial 15 Luteolin17, 30 UV protective, antioxidant, anti-inflammatory, antimelanogenic 16 Tangeritin18 Anti-inflammatory and anti-skin cancer 17 Myricetin19 UVB-induced skin cancer, anti-wrinkle 18 Proanthocyanidins20 Skin lightening, antioxidant 19 Hesperetin21 Photo protective, anti-inflammatory, antioxidant 20 Naringenin14,21, 47 Anti-inflammatory, anti-oxidant and anti-tumor, photoprotective 21 Gallic acid22,23 Anti-skin tumor, antioxidant, anti-tyrosinase 22 Vanillic acid24 Antioxidant, anti-tyrosinase 23 Protocataechuic Antioxidant acid25 24 Dicaffeoylquinic Antioxidant, anti-tyrosinase acid26 25 Tannic acid22, 27 Anti-skin tumor, anti-bacterial 26 p-Coumaric acid28 Anti-pigmentation

27 Taxifolin29, 30 Anti-skin carcinogenesis, anti-tyrosinase 28 Luteolin30 Anti-melanogenic 29 Naringin31 Skin photo protection, antioxidant, anti-carcinogenic 30 Chrysin32 Skin protection, anti-inflammation and anti-oxidation 31 Rutin33 Anti-inflammation, anti-oxidation 32 Ethylguaiacol34 Antimicrobial 33 Vinylguaiacol34 Antimicrobial 34 Cinnamic acid35 Anti-tyrosinase, antioxidant and antimicrobial 35 Vanillin36 Antioxidants, antibacterial, antimicrobial anti- inflammatory 36 Glycitin38,39,40 Anti-photo aging, anti-inflammatory, anti-aging 37 Genistin42 Prevent UV-induced skin damage 38 Caffeine44,45,46 Antioxidant, anti-skin cancer, antimicrobial 1: (Ramos et al., 2006)

2: (Choquenet et al., 2008)

3. (Alarcon De La Lastra and Villegas, 2005)

4: (Choi et al., 2010)

5: (Wang et al., 1999)

6: (Svobodová et al., 2003)

7: (Kris-Etherton et al., 2004)

8: (Saraf and Kaur, 2010)

9:(García-Bores and Avila, 2008)

10: (Chamcheu et al., 2015)

11: (Bin Hafeez et al., 2008)

12: (Bognar et al., 2013)

13: (Noda et al., 2002)

14: (Hämäläinen et al., 2007)

15: (Sato et al., 2011)

16: (Kitagawa et al., 2011)

17: (Wölfle et al., 2011)

18: (Yoon et al., 2010)

19: (Jung et al., 2010)

20: (Yamakoshi et al., 2003)

21: (Bonina et al., 1996)

22: (Perchellet et al., 1992)

23: (Kim, 2007)

24: (Chou et al., 2010)

25: (Tseng et al., 1996)

26: (Iwai et al., 2004)

27: (Akiyama et al., 2001)

28: (Seo et al., 2011)

29: (Oi et al., 2012)

30: (An et al., 2008)

31: (Ren et al., 2016)

32: (Wang et al., 2011)

33: (Choi et al., 2014)

34: (Banister and Cheetham, 2001):

35: (Kong et al., 2008)

36: (Makni et al., 2011)

37: (Jin et al., 2012)

38: (Seo et al., 2014)

39: (García-Lafuente et al., 2009)

40: (Kim et al., 2015)

41: (Zhao et al., 2010)

42: (Russo et al., 2006)

43: (Arung et al., 2011)

44: (Devasagayam et al., 1996)

45: (Lu et al., 2002)

46: (Almeida et al., 2006)

47: (An et al., 2010)

2. Objectives

Over the past several decades, most of the bioactive compounds identified in coffee and its by-products were through the labor-intensive conventional bio-assay guided purification schemes followed by structural analysis (e.g., NMR and MS/MS). Although high-resolution and tandem mass spectrometry have been utilized for structural elucidation process, the process often requires manual database searching for compound identification. Due to these technical limitations and time-consuming processes, it has been challenging to systematically characterize the bioactive compounds in coffee extracts. Recent advances in metabolomics science, mass spectrometry, computation capacity, and compound libraries have made this task possible.

The goal of this project is to investigate the chemical profiles of SCG derived from three coffee cultivars through global metabolite analysis and gauge the potential use of SCG from these cultivars cosmetics application.

The specific objectives of this study are as following:

1. Identify and characterize the health-promoting polyphenols, and other health- promoting secondary metabolites in SCG using global metabolomics approach (XCMS).

2. Compare the chemical profiles between the SCG extracts from Ethiopia coffee

(Yirgacheffe), Costa Rican coffee (Tarrazu) and Hawaiian coffee (Kona)

3. Materials and Methods

Figure 1.3 illustrated the flowchart of the SCG research using a non-targeted metabolomics approach. In this scheme, the compounds were extracted, and analyzed by

UPLC-MS and UPLC-HRMS (Bruker maXis Q-TOF), followed by XCMS Online analysis coupled with METLIN compound database. The list of identified bioactive compounds through this global metabolomics approach were compiled.

3.1. Sample preparation

SCG from three Arabica coffee cultivars including Ethiopian Yirgacheffe, Costa Rican

Tarrazu, and Hawaiian Kona were selected for this study. Roasted and ground coffee were purchased from the local Lakota Coffee (Columbia, Missouri). After brewing for 5 minutes, SCG were collected. Twenty-five grams of wet SCG (78% water, about 5.5 g

DW equivalent) were homogenized with blender and the compounds were extracted with

200 ml methanol (HPLC-grade, purchased from Fisher Scientific, Pittsburg, PA, USA).

The extract was then sonicated for 60 mins and filtered through Whatman phase separators (silicone treated filter paper, diameter 125mm). Each SCG extracts from three cultivars was prepared as independent triplicates.

3.2. LCMS Analysis

3.2.1. LC-Low Resolution conditions: The LCMS analysis was performed by a

Waters Ultra-High Performance Liquid Chromatography system coupled with Waters

Xevo TQ triple quadrupole mass spectrometer (HPLC-MS/MS). The compounds were separated by a Phenomenex (Torrance, CA) Kinetex C18 (100mm x 4.6 mm; 2.6 µm particle size) reverse-phase column. The mobile phase consisted of 10mM ammonium acetate and 0.1% formic acid in water (A) and 100% acetonitrile (B). The gradient conditions were 0–0.5 min, 2% B; 0.5–7 min, 2–80% B; 7.0–9.0 min, 80–98% B; 9.0–

10.0 min, 2% B; 10.0–15.0 min, 2% B at a flow rate of 0.5 ml/min. The MS system was operated using electrospray ionization in the either negative ion mode (ES-) or positive

(ES+) with capillary voltage of 1.5 kV (ES-). The ionization source was programmed at

150C and the desolvation temperature was programmed at 750C.

Figure 2.2. Flow chart of the SCG metabolomics approach project using Waters Xevo UPLC-MS for non-

VOC analysis and XCMS global metabolomics platform.

3.2.2 The LC-HRMS conditions: LC-MS analysis was performed on a Bruker maXis impact quadrupole-time-of-flight mass spectrometer coupled to a Waters

ACQUITY UPLC system. Separation was achieved on a Waters C18 column (2.1x 100 mm, BEH C18 column with 1.7-um particles) using a linear gradient of 95%:5% to

30%:70% eluent A:B (A: 0.1% formic acid, B: acetonitrile) over 30 min. Between 30-

33min, the a linear gradient was increased from 70% to 95% B, and maintained at 95%B for 3 min. The percentage of B was maintained at 5% from 36-40 min. The flow rate was

0.56 mL/min and the column temperature was 60 C. Mass spectrometry was performed in the negative (or positive) electrospray ionization mode with the nebulization gas pressure at 43.5 psi, dry gas of 12 l/min, dry temperature of 250C and a capillary voltage of

4000V. Mass spectral data were collected from 100 to 1500 m/z and were auto-calibrated using sodium format after data acquisition.

3.3 Global metabolite profiling using XCMS and METLIN Library

Figure 1.3. Flow chart of Metabolomics Data Processing.

XCMS Online data processing includes peak detection, retention time correction, alignment, annotation, and statistical analyses (Figure 1.3). First, determining the boundaries, centres and intensities of the two-dimensional peak signals in the LC/MS raw data is called feature detection (Tautenhahn et al., 2008). According to its peak detection algorithm, it covers both low-resolution, and high-resolution using matchedFilter method and centWave method respectively (Smith, 2013). After peak detection, XCMS Online uses nonlinear methods to compensate for retention time drifts between samples

(Tautenhahn et al., 2012). All total ion chromatogram (TIC)s should be aligned to compare compound quantities between fractions after retention time correction,. XCMS

Online uses the OBI-Warp (Ordered Bijective Interpolated Warping) retention time alignment method based on dynamic time warping with a one-to-one (bijective) smooth warp-function, especially for ESI-LC-MS data (Smith et al., 2006). Next step is ion annotation. By annotating at least two ions, the molecular mass can be calculated which is necessary to search compound libraries (Kuhl et al., 2011). After preprocessing, statistical analyses were conducted. For identifying distinct metabolites between two classes, it considers which intensities are significantly different from sample class to sample class and rank the metabolites by p-value (Smith et al., 2006). XCMS Online provides statistical result of data processing such as interactive cloud plot, principle component analysis (PCA), and non-metric multidimensional scaling (MDS).

3.3.1. Low resolution data processing through XCMS

The ion chromatograms generated from MS (.cdf) files including triplicate samples for each SCG cultivars were submitted to XCMS platform operated by Center for

Metabolomics at the Scripps Research Institute. The peak detection, peak grouping, spectra extraction, retention alignment, were processed by XCMS (Figure 1.3). The job was defined a multi-group job for comparative analysis to distinguish chemical profiling between three cultivars and the nearest instrument option, HPLC / Single Quad – HPLC with ~60 min gradient, Single Quadrupole MS was selected. For feature detection, the parameters were set as the following, mass tolerance of 30 ppm in consecutive scans, peak width of 10-60 s, S/N threshold of 6, and matched Filter algorithm was selected for low resolution data analysis. Both positive and negative polarity were analyzed. For the retention time alignment, the peakgroups method with m/z width 0.25 s, 5 (50%) of minimum fraction of samples was selected to make at least one of the sample groups valid. In every XCMS data processing, ‘plant’ was selected as a biosource. Post-hoc analysis in multi-group job was performed. Multivariate analysis and principle component analysis (PCA) were performed by XCMS to compare the chemical profiles between the three coffee cultivars. The spectra were annotated and the compounds were identified and categorized by the integration of the METLIN, the world's largest metabolite database. Each identified compound was assigned and related to its biological pathway through XCMS biological pathway/network analysis.

3.3.2. High resolution data processing through XCMS

Pairwise job was selected for two-group comparison study. First, we compared among cultivars without control. For example, Ethiopian Yirgacheffe versus Costa Rican

Tarrazu, Costa Rican Tarrazu versus Hawaiian Kona, Hawaiian Kona versus Ethiopian

Yirgacheffe. In second two-group comparison analysis, we used control (MeOH) with

17 each cultivar such as control versus Ethiopian Yirgacheffe, control versus Costa Rican

Tarrazu, control versus Hawaiian Kona. Each file group was submitted as triplicates. In the instrument selection, UPLC/Bruker QTOF POS was marked for high resolution data analysis, which automatically activates centWave algorithm, a newly published algorithm designed for centroid and high-resolution data analysis for both two-group comparison study. Parameters in high resolution analysis were set as follows: maximal tolerance of

10 ppm between successive measurements, peak width of 5-20 s, S/N threshold of 6 and obiwarp algorithm for retention time correction and alignment (width of overlapping m/z slices 0.015, minimum fraction of samples 0.5 for group validation). To visualize differences between two samples in pairwise job, 0.001 of p-value threshold and 1.5 of fold change was selected to be considered highly significant.

4. Results

XCMS Online allows for visualization of statistical results such as interactive cloud plot and principal component analysis (PCA). Both functions could be applied to multi-group and two-group comparison. Through multi-job using low resolution data, three clusters in

PCA and non-metric multidimensional scaling were used to closely examine and differentiate the chemical profiles and composition between the SCG extracts from

Ethiopian Yirgacheffe, Costa Rican Tarrazu, and Hawaiian Kona. Pairwise job using high resolution data enables comparisons between cultivars and provides statistical analysis as well. Lastly, pairwise analyses were performed to identify the bioactive compounds and evaluate their relative concentrations between the three cultivars. The results of pairwise analyses were compiled in table 3.1 (see Appendix A.) sorted by their compound groups

18 including phenolic acids, flavonoids and other compounds with their activities such as antioxidant, anti-inflammatory, and other biological effects relevant to skin protection.

4.1. Multi-group Comparison between three SCG cultivars using low resolution data

XCMS provides information about a comparison of features extracted from the chromatograms using principal component analysis (PCA), non-metric multi-dimensional scaling and metabolomic cloud plot (Figure 1.4c) to visualize statistical results.

For multi-job comparison analysis using XCMS, if more than three groups should be analyzed, multiple group analysis is more suitable. It enables the comparison between three SCG cultivars based on their metabolite features when they show significant difference. To distinguish each group as a specific cluster, Post-HOC multiple comparison was selected in parameter setting (Gowda et al., 2014).

Multi-group cloud plot allows visualization of metabolite features if they vary significantly among each group. Cloud Plot is showed based on their retention time and m/z ratio. In figure 1.4c, statistical significance threshold (Kruskal–Wallis p-value) was entered as 0.01 in main panel. In addition to p-value, feature intensity, m/z, retention time, and circle radius ratio range were selected as 29706525, 993, 15, 0.1 default values, respectively. With each selected circle, the box-plot, EIC, Post-HOC, and METLIN were illustrated under the cloud plot. Among features, ID: 280 was selected which includes quercetin 3,7-di-O-sulfate which is our candidate of interest in composite database.

Figure 1.4a represents the sulfate conjugate of quercetin which is significantly upregulated in the Hawaiian Kona (ESI(_)_K1) among other cultivals such as CostaRican

Tarazzu (ESI(_)_T1), and Ethiopian Yirgacheffe (ESI(_)_Y1). Therefore, multi-group

19 analysis helps to compare relative feature intensity among different groups (in this case, among different cultivars)

Intensity

Cultivar

Figure 1.4a. Box-plot generated from interactive cloud plot

The relative concentration between three cultivars for ID 280 (in interactive cloud plot) were represented by Box-plot. ID 280 includes quercetin 3,7-di-

O-sulfate which is our candidate of interest in composite database. Each SCG extracts was shown in x-axis and its relative concentration was indicated

in y-axis as signal intensity.

Figure 1.4b. Composite data base after mouse click on interactive cloud plot. Seven compounds were matched with METLIN database based on ion

chromatogram of each circle. Compounds in composite DB linked to METLIN library it will show compound formula, structure, and spectrum.

Figure 1.4c. Multi-group cloud plot. Metabolite feature which varies significantly (p < 0.01) from each groups are presented on cloud plot depending on

their retention time (x-axis) and m/z (y-axis). Each circle indicates each compound feature. The color intensity of circle represents statistical significance

(p-value). The size of the circle indicates feature intensity. If you move the mouse on each circle, p-value, q-value, m/z, RT, Max intensity will pop-up.

With mouse click each circle, Box-plot, EIC, Post-HOC, and Composite database (METLIN) are displayed under the cloud plot. Compounds in

composite DB are linked to METLIN library showing compound formula, structure, and spectrum.

XCMS Online provides interactive multivariate principal component analysis (PCA). The results of the PCA analysis suggests that SCG samples from three different coffee cultivars (Ethiopian Yirgacheffe, Costa Rican Tarrazu, and Hawaiian Kona) clustered by each group indicating chemical composition of cultivars is significantly different each other (Figure 1.5). The multivariate besides univariate analysis (cloud plot) are often required, when the differences between groups cannot be distinguished by simple separating method like univariate tests (Gowda et al., 2014). PCA is a major statistical tools in untargeted metabolomics for chemical profiling (Bartel et al., 2013). PCA reduces the dimensionality of the data while remaining as much of variation so it can detect differences between the groups. In figure 1.5, the three SCG cultivars were divided into each cluster by PCA. XCMS Online also provides scree plot, scores plot and loading plot in Interactive PCA tab (Figure 1.5a). The scree plot represents how many principal components were used in PCA. The scores plot visualizes cluster of submitted samples.

Three clusters are examined through PCA which means three SCG cultivars have distinct compounds contents compared to each other. Non-metric Multidimensional Scaling

(NMDS) plot also shows variation in the composition of SCG metabolite profiles.

Figure 1.5a, Interactive principal component analysis. A Scores plot shows the correlation between submitted samples. Three clusters represent each SCG cultivars which includes Hawaiian Kona

(ESI(-)_K1), Costa Rican Tarrazu (ESI(-)_T1, and Ethiopian Yirgacheffe (ESI(-)_Y1).

Figure 1.5b. Principal Component Analysis (PCA) results: SCG samples from three different coffee cultivars (Ethiopian Yirgacheffe, Costa Rican Tarrazu, Hawaiian Kona) clustered by each groups.

Individuals are projected in a reduced dimension space defined by component 1 (x-axis) and component 2

(y-axis). Each color corresponds to a specific cluster.

Figure 1.5c. Non-metric multi-dimensional Scaling of different coffee cultivars (Ethiopian Yirgacheffe,

Costa Rica Tarrazu, and Hawaiian Kona) shows that the chemical profiling is significantly different between three cultivars.

4.2. Identified bioactive compounds through global metabolomics analysis using high resolution data

4.2.1 Two -group comparison study between control and each cultivar

Through XCMS Online, around 200 bioactive compounds were identified including 22 phenolic acids, 161 flavonoids 2 coffee alkaloids, 1 stilbenoid, and 3 anthraquinones (see table 3.1). Free forms of these identified compounds have bioactivity related to skin care such as antioxidant, anti-microbial, anti-inflammatory, anti-skin cancer and skin whitening effect. In table 3.1, each row represents compound name, m/z, retention time, relative concentration between cultivars (Yirgacheffe, Tarrazu, and Kona) and bioactivity of the free form. Comparison of maximum peak intensity among the three SCG cultivars was accomplished by combination of 3 two-group comparison analysis. Several compounds which shares same molecular weight with the same m/z and max intensity indicates candidate compounds of certain ion chromatogram. Most of the compounds are derivatives and glycosides of well-known anti-oxidant, anti-inflammatory, anti-bacterial or anti-photo aging compounds. These bioactive compounds include the derivatives or free form of cinnamic acid, hydroxycinnamic acid, glycosidic form of isorhamnetin, hesperetin, theaflavin derivatives, glycosidic form of chrysin, epigallocatechin coumaroylquinic acid, coumaric acid glycosides, caffeoylquinic aicd, dicaffeoylquinic acid, glycosidic form of quercetin, kaempferol, cyanidin, pelargonidin, malvidin, apigenin, fujikinetin, genistin, isovitexin, viexin, puerarin, daidzein, dihydroxyflavone, free form of glycitin, obtusifolin glycosides, formononetin glycosides and procyanidin glycosides were identified in SCG by XCMS Online.

Figure 1.6. Chemical structures of phenolic acid.

Many types of phenolic acids ,caffeine and flavonoids (flavonols, flavones, isoflavones, flavanones, anthocyanidins and flavanols) were identified in SCG through XCMS

Online.For example, cinnamic acid, hydrocinnamic acid, p-coumaroylquinic acid, p- coumaric acid glucosides, vanillic acid derivatives, 1,4-Dicaffeoylquinic acid

(Isochlorogenic acid), chlorogenic acid were detected as a phenolic acid group.

Phenolic acids: For the cinnamic acids, a wide range of derivatives were identified as described in Table 3.1. Among them, hydrocinnamic acid was detected at RT 23.79 and

RT 14.21. Generally, Costa Rican Tarrazu has highest concentration of hydrocinnamic acid among the three cultivars. Likewise, this cultivar shows the highest amount of 4-Methoxy-

(2E)-cinnamic acid at RT 9.73. The detection of O-Methoxyhydrocinnamic acid in SCG is reported for the first time from this study. However, its bioactivity has not been carefully examined. The ion chromatogram showed that O-methoxyhydrocinnamic has m/z of

181.0848 with the RT 6.03 min. The relative signal intensities of O- methoxyhydrocinnamic were 29,768, 34,924, and 27,494 for Ethiopian Yirgacheffe, Costa

Rican Tarrazu and Hawaiian Kona respectively. P-coumaroylquinic acid, only appeared in

Hawaiian Kona variety, has been known for its strong antioxidant, and antibacterial activity

(Pereira et al., 2007). As candidate compounds, trans-p-coumaric acid 4-glucoside and coumaric acid at RT 4.92, m/z 349.0898 were only identified in cultivar Tarrazu. Vanillic acid 5-(4-carboxy-2-methoxyphenoxy)-, was first identified by this study having m/z

357.0585, RT 3.47. The free form of vanillic acid has been known for its strong bioactivity including antioxidant (Han et al., 1981), antibacterial and anti-fungal activity (Aziz et al.,

1998). Relative concentration of vanillic acid was the highest in Ethiopian Yirgacheffe as

68,432, 14,468, 20,036 in each cultivars (4.73 times higher than Costa Rican Tarrazu

varieties). Epigallocatechin 3-O-vanillate is first discovered from our research. Through

XCMS Online data processing, the compound was identified at RT 5.92 with m/z 457.1109 in Costa Rican Tarrazu and Hawaiian Kona. In case of caffeoylquinic acid, only Costa

Rican possesses 1,4-Dicaffeoylquinic acid and Cis-5-Caffeoylquinic acid with m/z

517.1335 and 355.1025 respectively. On the other hand, 1-feruloyl-5-caffeoylquinic acid concentration appear higher in Ethiopian Yirgacheffe than Costa Rican Tarrazu.

Isochlorogenic acid A (3,5-Di-O-caffeoylquinic acid), a compound with anti-oxidant and anti-tyrosinase activity so that it has value in and cosmetics (Iwai et al., 2004) showed highest concentration in Costa Rican Tarrazu varieties. In XCMS, the compound was identified with m/z 539.1149 at RT 6.67. XCMS also detected an anti-pigmentation compound from the methanol extract of SCG, which was identified as 4,5-푂- dicaffeoylquinic acid (4,5-diCQA).

Flavonoids: Quercetin, kaempferol, and isorhamnetin exist in glycosylated forms in SCG samples with various types of attached sugars, such as glucose, rhamnose, galactose, arabinose, and rutinose. As flavones, apigenin, chrysin, and hydroxyluteolin derivatives and dihydroxyflavone appeared in XCMS Online results table. Glycosylated flavanones such as naringin, and hesperetin glycosides were also detected. Isoflavones are commonly found in legume plants (e.g., soybean) (Amaral et al., 2006). However, XCMS Online detected the three major soybean isoflavones including genistein, daidzein, and naringenin attached with variety of glycosides in SCG samples. Also, isoflavones such as retusin, formononetin and fujikinetin glycosides were identified as candidate compounds. In addition, antocyanins including malvidin, cyanidin, pelargonidin glycosides, and glucopyranosylprocyanidin were detected through XCMS Online chemical profiling.

Relative concentration of flavonoids among SCG cultivars:

A remarkable difference in the concentrations of 8-C-ascorbylepigallocatechin 3-gallate

(m/z 633.1122, RT 35.86) were found between the SCG varieties. Previously, 8-C- ascorbylepigallocatechin 3-gallate was identified in oolong tea with no known bioactivities. Through XCMS Online chemical profiling, this compound was found in SCG extracts. The highest concentration was found in Ethiopian Yirgacheffe, approximately

3.24 times higher than that in Costa Rican Tarrazu.

We found that caffeine, was the most abundant compounds in Ethiopian Yirgacheffe cultivar followed by isorhamnetin glycosides, isorhamnetin 3-(3'''-ferulylrobinobioside),

E-cinnamic acid and isorhamnetin 3-glucosyl-(1->2)-galactoside-7-glucoside. Especially, isorhamnetin 3-glucosyl-(1->2)-galactoside-7-glucoside was found only in Ethiopian

Yirgacheffe SCG. In case of E-cinnamic acid detected at RT 16.76, Hawaiian Kona showed the highest concentration of the compound among the three cultivars. Costa Rican Tarrazu had the highest concentration of dicaffeoyquinic acid.

Flavonols such as quercetin and kaempferol are the most dominant flavonoids in foods and often exist in glycosylated forms (Manach et al., 2004). A wide range of quercetin glycosides including 3-sophoroside-7-glucuronide, 3-gentiobioside-7-glucuronide, and 7- glucuronoside 3-sophoroside were detected as candidates for certain ion peak having m/z

820.2114, and RT 4.57 in SCG samples. The maximum intensity generated by XCMS

Online and high resolution SCG data files was 3,178 2,780 or 2,312 for Ethiopian

Yirgacheffe, Costa Rican Tarrazu, and Hawaiian Kona respectively. Likewise kaempferol exists in many types of glycosylated forms including 7-rhamonoside, diacetylrhamnoside,

p-coumarylrhamnoside, coumaroylrutinoside, galloylrutinoside, coumarylrobinobioside, - galloyl-alpha-L-arabinopyranoside, arabinofuranosyl-(1->6)-glucoside. In the case of kaempferol with coumarylrobinobioside and diacetylrhamnoside moieties, the highest amount was found in Costa Rican Tarrazu. However, the levels of the glycosylated kaempferol with 7-rhamnoside (at RT 32.85) were around 2.7 times higher in Ethiopian

Yirgacheffe than Costa Rican Tarrazu. The levels for the rest of kaempfrerol derivatives at

RT 19.09 and RT 35.68 are the highest in Hawaiian Kona especially for latter case, the gap is 71% between Costa Rican Tarrazu. Isorhamnetin 5-glucoside that can be found in

Artemisia capillaris (Wu et al., 2001) also isolated in SCG Ethiopian Yirgacheffe.

Information about bioactivity related to cosmetics of Isorhamnetin 3-(4''-sulfatorutinoside) was rare likewise isorhamnetin 5-glucoside. Basically, contents of isorhamnetin glycosides are higher in Ethiopian Yirgacheffe. Among flavones, types of apigenin glycosides were noticeable in table 3.1. In case of apigenin 7-O-(6-O-malonylglucoside), it exists only in

Ehiopian Yirgacheffe at RT 34.86. Apigenin 8-C-(6''-acetylgalactoside) and apigenin 7- rhamnoside were not detected in Tarrazu and the concentration was rarely different between Yirgacheffe and Kona. Specially, the concentration of apigenin contained with acetylalloside in Ethiopian Yirgacheffe is 2.71 times higher than Tarrazu. The 7-O-beta-

D-glucoside of genistein was found at RT 5.57 only in Kona. Isoflavone glycosides, retusin

7-O-glucoside which has been reported exists in plant Pterocarpus marsupium; leguminosae (Schmid et al., 2003) was identified in SCG Ethiopian Yirgacheffe varieties with m/z 464.1547, RT 23.9. Formononetin 7-O-glucoside-6''-malonate (formononetin 7-

O-(6''-malonylglucoside) was detected as candidates at RT 5.83 and RT 6.67 accompanied by isochlorogenic acid (dicaffeoyquinic acid) showing relatively high concentration in

Tarrazu. Isoflavone glycosides, fujikinetin 7-O-laminaribioside and fujikinetin 7-O- glucoside were detected with quercetin 3-(4''-acetylrhamnoside)-7-rhamnoside at RT 32.85 and with 8-C-beta-D-glucofuranosylapigenin 2''-O-acetate at RT 5.83 respectively. In former case, the concentration of this compound in Yirgacheffe is 2.71 times higher than concentration in Tarrazu.

Among compounds described in table 3.1, anthraquinone glycosides such as physcion 8- glucoside were only detected in Yirgacheffe. The concentration of another anthraquinone derivatives chrysophanol 1-tetraglucoside was 57% higher in Kona compared to Tarrazu.

Lastly, various types of dihydroxyflavone (DHF) were selected in XCMS chromatogram peak at RT 20.28 with m/z 255.0655. The concentration of the compound in variety

Ethiopian Yirgacheffe is 23 % higher than variety Costa Rican Tarrazu.

Biological function

Our analytical results have identified several major compounds having bioactivity related to skin care application such as antioxidant, anti-inflammatory, anti-microbial and other biological functions (see table 3.1). For example, among the 200 identified compounds,

133 antioxidant, 132 anti-inflammatory, 13 anti-microbial, 15 anti-pigmentation, and 69 anti-cancer agents were found. Additionally, 30 compounds which have been reported as components for skin protection against UV or anti-aging were also detected through XCMS

Online.

Antioxidant effect: Several phenolic acids with antioxidant effect were found in SCG through XCMS Online; caffeoyl quinic acids (Iwai et al., 2004), chlorogenic acids (CGA)

(Sato et al., 2011), cinnamic acid derivatives (Kong et al., 2008), and vanillic acids (Chou et al., 2010) are well-known as their antioxidant. SCG also contains apigenin, chrysin, quercetin, hesperetin, naringenin, kaempferol, cyanidin, pelargonidin, isovitexin and malvidin glycosides and the free from of these flavonoids have strong antioxidant activity

(Saija et al., 1995) (Wu et al., 2011) (Torel et al., 1986) (Rice-Evans et al., 1996). Flavone tectochrysin in Ethiopian Yirgacheffe at RT 8.42 and the flavanol theaflavin which were detected in SCG by XCMS Online platform have been studied that it has antioxidant activity (Lee et al., 2003b) (CHEN and HO, 1995). Also, there is one interesting fact that

7,8-dihydroxyflavone selected in XCMS chromatogram peak at RT 20.28 with m/z

255.0655 has an antioxidant effect inhibiting cell-death from hydrogen peroxide (Chen et al., 2011). Isoflavones, formononetin derivatives containing malonate are detected at RT

6.67 and the free form is known as antioxidants (Mu et al., 2009). Interestingly, puerarin

4'-O-glucoside were reported as antioxidant as glycosides itself (Xiao et al., 2011). Among several kinds of procyanidin included in SCG sample, procyanidin B2 (epicatechin-(4β-8)- epicatechin) are known as antioxidant (Sakano et al., 2005). Coffee alkaloid caffeine which is most abundant in SCG are also have antioxidant activity (Devasagayam et al., 1996).

Lastly, eugenol and resveratrol are also studied widely as antioxidant (Bendre et al., 2016)

(Mahal and Mukherjee, 2006).

Anti-inflammatory effect: In table 3.1, it is revealed that a lot of compounds that have anti- inflammatory effect with anti-oxidant activity. For example, the isoflavones daidzein and genistein, the flavonols isorhamnetin, kaempferol and quercetin, the flavanone naringenin, and the anthocyanin pelargonidin existed in SCG samples as glycosylated forms and their free forms are known as anti-inflammatory compounds inhibiting NF- κB activation (Rice-

Evans et al., 1996). Additionally, other flavonoids such as malvidin (Huang et al., 2014), chrysin (Wu et al., 2011), isovitexin (Lv et al., 2016), and hydroxykempferol (Williams et al., 1999) 7,8-dihydoxyflavone, (Park et al., 2012), procyanidin B1 (Xing et al., 2015) are important anti-inflammatory plant flavonoids and were detected by XCMS Online with attached with glycosides. Other types of compounds including phenylpropanoid, stilbenoid, and anthraquinones were found in SCG with anti-inflammaotry activity (e.g. eugenol, resveratrol, chrysophanol, physcion) (Bendre et al., 2016) (García-Sosa et al.,

2006). Puerarin 4'-O-glucoside antioxidant introduced in above, has anti-inflammatory effect as well (Xiao et al., 2011).

Other activities related to skin care: Among phenolic acids, E-cinnamic acid (RT 16.76), vanillic acid (RT 3.47), and dicaffeoylquinic acid (RT 6.67) has anti-tyrosinase effect which means skin whitening efficacy (Lin et al., 2011) (Chou et al., 2010) (Iwai et al.,

2004). Likewise, coumaric acid which was detected at RT 4.92 also has anti-pigmentation effect (An et al., 2010) while chlorogenic acid (5-Caffeoylquinic acid) affects to photo aging (Kitagawa et al., 2011). In case of caffeic acid, there has been some reports that it increases the chances of cancer. However, the International Agency for Research on

Cancer (IARC), convened by the World Health Organization has just concluded that there’s no strong evidence that coffee increases the chances of cancer which have mooted in the past (Loomis et al., 2016). In addition, the compound protects phospholipidic membranes from UV light-induced peroxidation (Bonina et al., 2002) and showed a significant protection to the human skin against UVB-induced erythema (Saija et al., 2000).

Flavonoids which have been widely studied for skin protection properties are found in

SCG. Among them, apigenin, genistin, genistein, daidzein, chrysin, eriodictyol, hesperetin,

and naringenin have been reported as photo protective agents against UV (Wang et al.,

2011) (Bonina et al., 1996) (An et al., 2010). Especially quercetin are notable for having anticancer, anti-melanogenesis, skin protection, and antibacterial effects (Svobodová et al.,

2003). Additionally, genistin detected in Kona at RT 5.57 was also reported to have protective effect on UV-induced DNA damage in cell-free condition. This compound reduces the M14 cell viability due to anti-carcinogenic effects (Russo et al., 2006).

Flavanones naringin was studied as photo protection, and anti-carcinogenic compound

(Ren et al., 2016) mostly found in Yirgacheffe in this study. Cyanidin also has multiple health-promoting effects such as anticancer, skin protection, and antiaging (An et al.,

2010). Glycitin (García-Bores and Avila, 2008) which is a natural isoflavone, was detected as a free form in SCG at RT 23.9 and it is known as antioxidant, anti-allergic agent which has been reported to increase dermal fibroblast cell proliferation and migration linked to anti aging effect (Kim et al., 2015). The protective effect on UV-induced skin photo aging was also reported by Seo and Park. (Seo et al., 2014). Both 6-hydroxykaempferol and 6- methoxyeriodictyol identified in SCG using metabolomics analysis has anti-tyrosinase by itself (Gao et al., 2007) (Lee et al., 2006). Compared to other flavonoids whether free form or glycosylated form, a great amount of hesperetin laurate which has anti-fungal effect strongly inhibiting the mycelial growth of P. expansum (Razzaghi-Abyaneh and Rai, 2013) was shown in variety Kona at RT 36.05. The compound also acts as skin conditioning agent

(Slavtcheff et al., 2005). In addition, it is interesting that caffeine has anti-skin cancer (Lu et al., 2002) and antimicrobial (Fardiaz, 1995) with emollient properties. Besides, several compounds including guaiacol, chrysophanol, and physcion shares anti-microbial activities and three anthroquinones (chrysophanol, soranjidiol, and physcion) listed at RT 35.68, RT

20.28, and RT 23.9 respectively have anti-skin cancer effects (García-Sosa et al., 2006;

Vittar et al., 2014). Eugenol detected through XCMS Online in SCG are well known aromatic compound responsible for “woodsy” flavor in coffee. It has multiple bioactivities such as antimicrobial, and anticancer with antioxidant and anti-inflammatory effects

(Bendre et al., 2016). Lastly, resveratrol is well known for anti-aging and photo protection effects (Alarcon De La Lastra and Villegas, 2005). Such reports about bioactivity can make SCG as a promising raw material in the dermatology field.

4.2.2 Two-Group Comparisons between SCG cultivars

Comparing two groups is the basic strategy in every research to determine certain bioactive metabolite features that one group has significant difference compared to the other. Here, pairwise analyses were performed to compare the chemical profiles and the concentrations of the bioactive compounds between Ethiopian Yirgacheffe vs. Costa

Rican Tarrazu, Costa Rican Tarrazu vs. Hawaiian Kona, and Hawaiian Kona vs.

Ethiopian Yirgacheffe. For each group, data files (.cdf) from three independent replicates were uploaded and processed by XCMS. More than 11,932 features generated from comparison between Ethiopian Yirgacheffe and Hawaiian Kona, and several of them were identified as the potential bioactive compounds for skin-care application. For example, strong anti-inflammatory compounds quercetin pentaacetate was identified with m/z 530.1253, with retention time (RT) 11.70 min. The concentration of quercetin pentaacetate was significantly higher (p = 0.00147) in Ethiopian Yirgacheffe with 2.3 folds higher as compared to that in Hawaiian Kona. In another experiment, Ethiopian versus Costa Rican Tarrazu, significant feature (p = 0.02081) detected at RT 34.86. The candidate compounds were cyanidin-3-(6'-malonylglucoside), apigenin 7-O-(6-O-

malonylglucoside) with m/z 558.0949, four times higher concentration in Ethiopian rather than Costa Rican Tarrazu (Figure 1.7b). The maximum intensity was 242,958.

Comparison between two groups were also visualized by cloud plots. 21 features were acquired from Ethiopian versus Hawaiian Kona. Nine and 46 features were detected from

Costa Rican Tarrazu versus Hawaiian Kona and Ethiopian Yirgacheffe versus Costa

Rican Tarrazu respectively. Similar to multi-job interactive cloud plot, users can view m/z, RT, p-value, fold change, max intensity, and METLIN library matches by moving mouse. Interactive cloud plot enables filtering p-value and fold change by adjusting the bar in main panel. By zooming, mouse can detect every circle without any inconvenience even though the circles are overlapped.

Figure 1.7a. Extracted ion chromatogram of Ethiopian Yirgacheffe (E) and Hawaiian Kona (H). The peak represents relative intensity of quercetin pentaacetate in SCG from Yirgacheffe (E), and Kona (H).

Figure 1.7b. Extracted ion chromatogram of Yrigacheffe (E) and Costa Rican Tarrazu (C). The peak represents relative intensity of candidate compounds such as cyanidin-3-(6'-malonylglucoside) or apigenin

7-O-(6-O-malonylglucoside) at RT 34.86.

Figure 1.8. Cloud Plot results: A Cloud Plot displays features whose intensities are altered between sample groups. Up-regulated features are represented as circles on the top of the plot and down-regulated features are represented as circles on the bottom of the plot, where the size corresponds to the (log) fold change of the feature. Circles with black outlines indicate hits in the METLIN database. The lighter or darker the color of the circle relates directly to the significance of the p-value. 46 features were generated by pairwise comparison between Ethiopian Yirgacheffe and Costa Rican Tarrazu showing the relationship between retention time and m/z. Green color represents upregulated compounds which means abundant in Tarrazu and red color circle represents down regulated compounds which means the compounds exist more in

Yirgacheffe. With a mouse click, EIC, Box and Whisker, and Spectrum appears under the cloud plot.

5. Discussion

5.1. Global metabolomics approach (XCMS Online) vs traditional analytical techniques

Over the past decades, scientists have been relying on traditional analytical techniques such as targeted analysis to identify compounds in samples. Although bioassay-guided fractionation followed by targeted metabolite analysis and NMR structural elucidation has been broadly used in identifying novel bioactive natural compounds, these techniques are time-consuming research. For example, bioassay-guided fractionation involves multiple steps such as repetitive preparative-scale fractionation and assessment of biological activity until pure compound with the bioactivity of interest have been isolated

(Marston and Hostettmann, 2009). However, natural compound research has been improved by modern separation and structural elucidation techniques such as HPLC-MS and NMR (Seger et al., 2005; Steinmann and Ganzera, 2011). Probing the compound composition, content, and structure in time and space often need accurate, sensitive, selective, stable, fast, automated, high-throughput processes (Ju, 2013). These characteristics are the eternal goals of analytical chemistry and are challenges even in modern industry.

Since metabolomics is a powerful tool for profiling of largely unknown chemicals, global metabolomics analysis has been used as a modern approach for identification of bioactive compounds in natural products such as cooked rice, small berries and whole grain cereals (Heuberger et al., 2010) (Badjakov et al., 2008) (Gani et al., 2012). In these papers, chemical comparison and profiling were achieved globally and rapidly by untargeted metabolomics-driven strategy (UMDS) trying to meet these requirements

described above. By introducing the combination of modern data processing online platform (XCMS) and global compound data bank (METLIN), we overcome challenges of traditional analytical chemistry which often are time consuming and labor intensive to distinguish the bioactive compounds between sample treatments or groups. XCMS provides PCA, non-metric multi-dimensional scaling, and extracted ion chromatogram results (Figure 1.5 and 1.7) showing significant differences between sample groups based on their chemical profiling. These results showing distinct characteristics between cultivars matched with our expectation the three varieties of coffee are likely comprised of different metabolites. Since they have distinct and typical characteristics in flavor profile coming from different growing conditions such as geographic regions, elevation, soil composition and temperature, we anticipated different chemical profiling results for the three varieties.

Table 1.2. Growing condition of coffee cultivars (Ethiopian Yirgacheffe, Costa Rican Tarrazu, and

Hawaiian Kona)

Species Growing Harvest Roasting

Altitude period

Ethiopian Coffea Arabica 1,700 – 2,200 M October – Medium

Yirgacheffe December

Costa Rican Coffea Arabica 1,200 – 1,800 M December – Medium

Tarrazu February

Hawaiian Kona Coffea Arabica 250 – 1000 M August – Medium

January

For example, Ethiopian Yirgacheffe coffee is grown at elevations from 1,700 to 2,200 meters above sea level in southern Ethiopia where coffees grow slowly due to the altitude

(the town of Yirgacheffee in Ethiopia’s Sidamo area), allowing additional time for the tree to deliver nutrients to the coffee and develop the best flavors. They were harvested from October to December. Compared to this, Costa Rican Tarrazu coffees were grown in lower regions which is called Tarrazu Coffee harvested between December and

February. Lastly, the Hawaiian Kona coffee is grown in the Kona district in Hawaii. The farms are on the volcanic land with mild temperatures and located in lower altitude than some of other Arabica coffees. Their harvesting time is from late August to late January.

Although shade and altitude are major factors known to have benefit on coffee quality, there are contradictory results obtained from literatures on the chemical composition related to these influence. In case of total lipid content, it was reported to decrease or not to be influenced by altitude increases. Chlorogenic acid contents were shown inconsistent results in several studies (Guyot et al., 1996). Additionally, the bigger factor than elevation is reported as maturation rate of coffee fruit on the plant. In usual, low temperature slow down the ripening process and leads to higher accumulation of aromatic compounds (Joët et al., 2010). During the roasting process, not only aromas but also chemical composition of coffee will change. At medium roast level, the coffee will create a balance between acidity and body showing a darker brown color. According to Vignoli

(2011), increasing roasting will degrade 5-CQA and form high molecular weight polymers, which is called melanoidins. However, they concluded that the antioxidant activities under different conditions were equal and the activity depends more on chemical composition than roasting degree (Vignoli et al., 2011). In terms of antibacterial

activities, Daglia (1994) found that 1) the lack of this activity in green coffee and 2) degree of roasting affected the activities in roasted coffee extracts (Daglia et al., 1994a).

To quantify the contents of specific compounds, high resolution data from UPLC Bruker

Q-TOF was submitted to XCMS Online. Compare to low- resolution data which provides visualized statistical results (PCA, and nonmetric-dimensional scaling), high resolution data generates the relative intensity providing more detailed and numerical results of compound content (table 3.1). From the XCMS Online result table of pairwise job submission, maximum intensity was summarized in table 1.2 which shows significantly different concentration between three cultivars. Additionally in table 3.1, 200 bioactive compounds are compiled to compare the relative intensity of compounds which have bioactivity related to skin care (phenolic compounds). While traditional target analysis would take much more time and labor, this study has clearly demonstrated the advantages of global non-targeted analysis over the traditional method for systematic analysis of the bioactive compounds in SCG. Various compounds were detected in coffee samples in

XCMS result table, from simple linear and branched molecule structures to complex glycosylated forms.

5.2. Phenolic acids and caffeine in SCG

From XCMS Online result, 22 of the phenolic acids were detected including cinnamic acids, vanillic acids, and caffeoylquinic acids. Their concentration were relatively higher than other groups of compounds. Specially, the concentration of cinnamic aicd was most abundant followed by caffeine in the SCG. The majority of the phenolic acids identified in this study were in free form. Phenolic acids including gallic acid, p-coumaric, caffeic,

and ferulic acids are rarely found in the free form, except in processed food that has undergone freezing, sterilization, or fermentation (Raman et al., 1995). The bound forms are glycosylated derivatives or esters of quinic acid. Especially caffeoylquinic acids

(CQAs) which is found in high concentrations in coffee can be formed by combining caffeic acid and quinic acid (El Gharras, 2009). Their biological function have been widely studied for the anti-oxidant and anti-pigamentation effects for potential skin whitening agent. According to the Bravo, even though 5-CQA is the most abundant caffeoylquinic acids in SCG, the total amount of CQA including 3-CQA, 4-CAQ, and 5-

CQA is less than the total Di-CQA including 3,4-diCQA, 3,5-diCQA, and 4,5-diCQA .

Similar aspect was detected in XCMS Online result table (table 3.1). Interestingly, in

Yrgacheffe, CQA was not detected although several types of Di-CQA such as 1,3- diCQA, 1,4-diCQA, 1,5- diCQA, 3,4- diCQA, 3,5- diCQA, 4,5- diCQA, were identified.

The reason why diCQA contents are higher than monoCQAs in SCG is that monoCQAs are extracted during coﬀee brewing, whereas the diCQAs need more time to be extracted

(Bravo et al., 2012). Other research on this phenomenon explained that diCQAs are extracted slowly compared to monoCQAs from coffee (Blumberg et al., 2010; Ludwig et al., 2012). In addition, a previous study revealed that caffeine and chlorogenic acids contents, including diCQA and monoCQA in arabica SCG were slightly higher or similar with reported concentrations in arabica coffee brew (Monente et al., 2015). Therefore, spent coffee extracts which pass through the brewing process can be considered as a better source of diCQAs and caffeine (compare to coffee brew) which are used in cosmetic fields due to their antioxidant and antimicrobial effects (Almeida et al., 2006;

Devasagayam et al., 1996).

Our results are not consistent with the studies previously conducted by Belguidoum

(Belguidoum et al., 2014). In their studies, the concentration of cinnamic acid was much less than chlorogenic acids in roasted coffee and green coffee samples whereas a higher amount of E-cinnamic acid was detected by XCMS Online compared to other phenolic acids such as coumaric acid, vanilic acid derivatives and (di)-caffeoylquinnic acids.

When the levels of the bioactive phenolic acids were compared between the cultivars, the levels of E-cinnamic acid (the third highest concentration in SCG samples) in variety

Kona has almost 1.4 times higher than Tarrazu, suggesting the SCG from the variety

Hawaiian Kona is an possible source for antioxidant and anti-pigmentation products.

However, other anti-pigmentation compounds such as coumaric acid derivatives, vanillic acid derivatives, and (di)-caffeoylquinic acids were produced as highest amount in

Tarazzu even though their concentrations were much lesser compared to E-cinnamic acid

(minimum 4.48 times) (see table 3.1).

5.3. Glycosylated form of flavonoids and their activity:

In this study, we identified a 139 glycosylated flavonoids with potential applications for personal care product industry. In cosmetics, flavonoids are becoming popular due to their antioxidant, anti-aging activities related to fibroblast proliferation (Kim et al., 1997) or reduction of collagen breakdown (Schmid and ZÜLLI, 2002). Most of the flavonoids in plants are found in glycosylated form so they can be dissolved in water phase and kept in plant vacuoles (Manach et al., 2004). The sugar is often glucose or rhamnose, but galactose, arabinose, xylose, glucuronic acid also can be attached (Manach et al., 2004).

As flavonoids do not exist in plants as aglycones except for catechins, consistent with our

findings that most of the compounds were in the glycosylated form (Table 3.1). Even though glycosides are generally inactive form in humans because of low cellular uptake due to inadequate size for enzyme binding sites and receptors (Schmid et al., 2003) most cosmetic ingredients containing flavonoids are glycoside forms (Schmid et al., 2003).

Therefore, after topical application, they should pass through enzymatic hydrolysis at their target location to become biologically active aglycones (the molecular from without sugar residues) (Schmid et al., 2003). For example, one study revealed that, the antioxidant capacity of soy bean flavonoids increased 33 % after enzymatic treatment

(Sansanelli et al., 2014). Other studies using mulberry leaves demonstrated that improved bioactivity is due to enzyme reaction, hydrolysis of the glycoside polyphenols rutin, soquercitrin, and astragalin into the aglycone polyphenols quercetin and kaempferol producing a remarkable increse in anti-oxidant and anti-tyrosinase inhibition activities by 219.5% and 230.9% respectively. Additionally, a significantly improved skin permeability (513%) was observed by the reaction due to β-glucosidase enhancing hydrophobicity (239.41%) and reduced molecular weight (from 387.3 to 291.4), which are critical factors for absorption of bioactive ingredients (Do et al., 2009). Enzymatic treatment is more important in cosmetic application rather than nutraceutical intake because the latter undergoes hydrolysis of glycoside to aglycones by fermentation during the manufacturing of certain foods or by intestinal bacterial metabolism (Manach et al.,

2004).

5.4. Flavonoids (flavonols, flavones, isoflavones, flavanones, anthocyanidins, and flavanols) in SCG for cosmetic application

Through the global metabolomics analysis, more than 180 bioactive flavonoids, including flavonols, flavones, isoflavones, flavanones, anthocyanidins, and flavanols were identified. These compounds are strong anti-bacterial, anti-oxidant, skin protection, anti- melanogenesis, and anti-inflammatory agents for skin care applications (Ramos et al.,

2006) (Choquenet et al., 2008) (Hämäläinen et al., 2007). According to Manach, flavonols are the most abundant flavonoids in foods, and the main representatives are quercetin and kaempferol (Manach et al., 2004). These compounds are also found in SCG as well with relatively higher concentrations than other flavonoids.Compared to flavonols such as quercetin and kaempferol, flavones which are mostly glycosylated apigenin, luteolin, and chrysin are less abundant in fruit or vegetables (Manach et al., 2004). As expected, these three flavones existed in SCG attached to several sugars such as glucose, galactose, glucopyranose, coumaryl glucose, rhamnose and malonylglucose. Another type of flavonoids, flavanones (eriodictyol, hesperetin, naringenin) are generally glycosylated at position 7 (Tomás‐Barberán and Clifford, 2000). These flavanones can be found in tomatoes, mint and but are only present at high concentration in citrus fruits.

XCMS Online identified naringenin and naringenin-7-O-glucodside in SCG. Although isoflavones are known as major compounds which can be found in leguminous plants

(Tomás‐Barberán and Clifford, 2000), daidzein 4'-O-glucoside, daidzein 6''-O-acetate, daidzein 7,4'-di-O-glucoside, genistin (genistein-7-glucoside) and genistein 7-O- rutinoside were identified by XCMS Online in SCG as well. Isoflavones glycosides are sensitive to heat. However, XCMS Online detected them in SCG that passed through the roasting process. In case of flavanols such as catechin and epicatechin the most common compounds in tea, exist in both monomer and polymer forms (El Gharras, 2009). The

ratio of monomers and polymers would change based on the fermentation (heating) process producing more complex polyphenols known as theaflavins (dimers) and thearubigins (polymers) (Cloughley, 1980). Therefore, black tea that is a more highly processed product has more complex polyphenols compared to green tea due to oxidation process leading to orange-brown color of black tea. Similarly, the theaflavins and its derivatives were detected in SCG which passed through the roasting process (see table

3.1). However, the monomer form of flavonoids such as 8-C-Ascorbylepigallocatechin 3- gallate still remained in SCG. In summary, some polyphenols such as quercetin are found in all plant products including fruit, vegetables, cereals, leguminous plants, tea, wine, etc

(Pandey and Rizvi, 2009). However, other flavonoids can be found only in certain foods.

For example, flavanones are in citrus fruit at high concentration, while isoflavones are in high concentration in soy protein. In SCG, flavonoids including flavanone, flavanonol, anthocyanidins, and isoflavones were found by METLIN database matches based on their accurate mass.

Our results reveal a existence of many kinds of flavonoids in SCG, which is consistent with the studies previously conducted by Mussatto (Mussatto et al., 2011a). In their studies, they also have reported high levels of total flavonoids and antioxidant properties in SCG extracted using methanol. They agreed that SCG could be of great interest for application in food and pharmaceutical fields due to antioxidant activities from phenolic compounds in SCG. Another research claimed that rutin, and quercetin were not detected in studied coffee samples (Belguidoum et al., 2014). The likely reason why quercetin was

not detected in their study is that the compound exist as glycosylated form not as free form as shown in our study.

When the levels of the bioactive flavonoids were compared among the varieties, the levels of 8-C-ascorbylepigallocatechin 3-gallate in variety Ethiopian Yirgacheffe was almost 3.24 times higher than Costa Rican Tarrazu. Although this compound was reported to have anti-HIV and anti-diabetic effects (Lee et al., 2003a) (Fukui and

Iwashita, 2013), dermatological activity has not been analyzed. Interestingly, the concentration of dihydroxyflavone which has strong anti-cancer effects was highest in

Yirgacheffee suggesting the SCG from the variety Yirgacheffe is the most possible candidate among three cultivars for skin protection products against UV-induced cancer.

Additionally, the antiaging agent, naringenin chrysin, which can provide UV protection, was detected only in Yirgacheffe. Likewise, the levels of the antiaging compound, glycosylated cyanidin, was 2.7 times higher in cultivar Yirgacheffe compared to Costa

Rican Tarrazu. The concentration of isorhamnetin, which has antioxidant and anti- inflammatory activity, was highest in Kona about 20 times than found in the Tarrazu cultivar. However, the concentration of other bioactive compounds such as quercetin, apigenin, and keampferol varied between the three cultivars.

The levels of 6-methoxyeriodictyol, which has skin-whitening activity, was significant only in Tarazzu, whereas the other anti-pigmentation agent such as 6-hydroxykaempferol

3,6-diglucoside 7-glucuronide concentration were not present. However, the concentration gap between cultivars is quite small.

Anti-oxidants, including hesperetin derivatives, were primarily found in Hawaiian Kona mostly. Of note is hesperetin 5,7-laurate which is used for emollient in skin conditioning products.

Only three flavonols (quercetin, myricetin, and kaempferol) and two flavones (apigenin and luteolin) are well studied with regard to the positive effects of dietary flavonoid intake. Bioactive compounds which exist in SCG have never been systematically studied using metabolomics approach for skin care application. Through this study, variety of flavonoids were detected in SCG, even though further studies is need to measure absolute quantity of these bioactive compounds for SCG to become a potential raw material in cosmetic formulation. Basically, the idea was that SCG is a by-product of roasted coffee which has variety of bioactive compounds including polyphenols (Nichols and Katiyar,

2010) and SCG may still retain those compounds. Several types of antioxidant compounds has more uses than aroma including caffeic acid melanoidins and vitamins, are generated through the roasting process (Yanagimoto et al., 2004). In many cases, plant extracts contained natural antioxidants such as polyphenols, flavonoids, flavanols, stilbens, and terpenes are used in commercial cosmetics. Some commercial products contain pure natural compounds such as quercetin, and resveratrol in their formulation

(Kusumawati and Indrayanto, 2013). These compounds were detected in SCG as well.

The bioflavonoids, including quercetin and hesperidin, are known as anti-inflammatory compounds inhibiting histamine release and mast cell degranulation. Quercetin (found in high levels in SCG through XCMS) inhibits phospholipase A2 (Lanni and Becker, 1985) and lipoxygenase enzymes (Kato et al., 1983). Other compounds such as tannins are also

major component in soothing compresses from their astringent, and cooling efficacy

(Graf, 2000). Several cosmetic products and soothing compresses includes bioflavonoids such as quercetin, hesperidin, and kaempferol which are the best known therapeutically useful bioflavonoids (Kenner and Requena, 2001). In addition, flavonoids such as apigenin and luteolin are major components in chamomile and documented to have anti- inflammatory and soothing effects in creams. This study observed that SCG contains several main active compounds with antioxidant or anti-inflammatory activities used in real cosmetic formulations. Since the early 2000s, a lot of studies have revealed that flavonoids have photo protective effect on skin due to their strong anti-oxidant activity.

From XCMS analysis, we found that SCG contains various bioactive compounds such as anti-oxidant, anti-inflammatory anti-tumor, and anti-tyrosianse. These results support that the idea that SCG can be used as a source of major active materials in cosmetic industry.

In addition to identifying health-promoting compounds, our study has importance in terms of using novel untargeted method to identify these compounds in SCG.

As described in table 3.1, several listed bioactive compounds have being known as agents with strong antioxidants, antimicrobial, antitumor and anti-tyrosinase activities. In addition, relative intensity of each compound was summarized through XCMS pairwise analysis.

Even though the compounds were qualified and the relative concentrations of these compounds were analyzed between three cultivars through computational technology, synergistic or antagonist effects between compounds that may affect their human health- promoting efficacy was not determined in this study. Therefore, the results can’t show which cultivar is better among the three varieties in antioxidant, anti-inflammatory, anti-

microbial or anti-tyroisnase activity. Several kinds of bioassays that are listed above will address this limitation and allow the selection of one variety with the highest efficacy in vitro and in vivo. For example, synergy effect between p-coumaric acid and ferulic acid has been reported by Maillard and Berset (1995) in antioxidant activity assays. Also, caffeic acids, which are abundant in SCG, can be affected by the presence of certain compound such as ascorbic acid. Compared to synergy effects, antagonist effects in antioxidant activity have been reported between ellagic acid and catechin, which were also identified in SCG through XCMS. Many studies are focusing on natural bioactive compounds from residual sources for cosmetics and these effects have been reported.

Synergistic effects of phenols from grape seeds and pomace polyphenols have been reported. Even though flavonoids and phenolic compounds interact with each other

(Meyer, Heinonen et al., 1998), the choice of the right combination of compounds for their stability and synergistic effect in cosmetics should be achieved to meet customer’s demands. That is why crude extracts of the three SCG cultivars should be tested through bioassay besides purified compounds even though the relative intensity was qualified for each compounds through metabolomics approach.

6. Conclusion

This project utilizes a modern metabolomics approach to identify the bioactive compounds in SCG and explore the new applications for skin-care application. The findings from this project will help turn the low value renewable materials into high- value products. It will not only benefit consumers (organic based, safe, effective), producers (low cost and abundant materials), and the environment (green, sustainable),

but also foster the development of rural economies. According to Industry Trends Report on Organic Personal Care Maket size, organic based personal care market size will reach

USD 25.11 billion by 2025. Therefore, the vision of this project completely aligns to global trends and provide the organic-based products without concerns of toxicity. In this study, methanol was used as a solvent to extract SCG. However, for practical large-scale process, supercritical fluid extraction (SFE) method should be used to reduce the extraction cost, toxic solvent use, and produce large quantity of the bioactive extracts. In this SCG project, there are more than 200 bioactive compounds were identified including phenolic acids, flavonoids, coffee alkaloids, stilbenoids, and anthraquinones. The toxicity of the bioactive compounds identified in SCG should be further evaluated.

Chapter 2. Isolation of the antioxidant and antimicrobial compounds with bioassay-guided fractionation

Abstract: Today, the importance of using organic-based anti-oxidant and anti-microbial compounds for cosmetic formulation is emerging. Oxidative stress is closely linked to skin aging processes. Antioxidants are valuable antiaging component by helping the skin fight against free radicals produced during oxidative stress. In addition, multiple antimicrobial ingredients are included to protect the products against the growth of micro-organisms such as bacteria and fungi. By formulating anti-microbial compounds in cosmetics, the products can be used to treat skin conditions associated with microbial infection, such as acne. It will also help prolong the shelf-life of the products and keep the products from causing skin infections, corruption or discoloration of the product itself during the usage. The present research examines SCG as the bioactive ingredient having anti-oxidant and anti-microbial activity for cosmetics formulation. To create value-added product from SCG, the objectives of this project are to 1) isolate and identify the bioactive compounds in the SCG of three cultivars (Ethiopian Yirgacheffe, Costa Rican

Tarazzu, and Hawaiian Kona) using bioassay guided-purification, and 2) explore their application in cosmetics. The compounds in SCG were extracted and tested by FRAP antioxidant assay, and zone of inhibition anti-microbial assay. The bioactive compounds were then isolated by bioassay-guided flash chromatography followed by HPLC purification for the structural analysis by high resolution mass spectrometer (HRMS). In anti-oxidant assay, the SCG extracts from Hawaiian Kona showed the highest activity among the three cultivars. Antimicrobial activity from the SCG extracts from Hawaiian

Kona and Ethiopian Yirgacheffe were much higher than Costa Rican Tarrazu. The

findings from bioassay and high-resolution analysis showed that SCG is an excellent potential raw material for cosmetic application (e.g., anti-oxidant, anti-microbial, skin- whitening, and anti-aging). High resolution analysis using systematic approach was conducted to identify the compound in three fractions which were purified through

HPLC. The SCG from Ethiopia coffee (Yirgacheffe) was fractionated by flash chromatography and HPLC for further purification and structural analysis. Fraction number 4 from flash chromatography showed the highest anti-oxidant activity and was analyzed further by HPLC. In addition, fractions 7 and 12 from flash chromatography showed the highest anti-microbial activity. After injecting fraction number 4 from Flash chromatography, 27 purified fractions were collected through HPLC and purified fraction

14 was selected for high resolution analysis. The results suggested that cafesterol, coffee alkaloid is the bioactive compound responsible for the observed strong anti-oxidant activities.

1. Introduction

Recent studies have suggested that SCG could be an excellent source of high value phytochemicals for cosmetics and skin-care application. Some of examples include chlorogenic acid (antioxidant), caffeic acid (antioxidant), isochlorogenic acid

(antioxidant, anti-inflammatory), protocatechuic acid (antioxidant, anti-fungal), 4,5- dicaffeoylquinic acid (pigmentation inhibitor) and quercetin (anti-fungal, anti-oxidant, skin protection, anti-melanogenesis) (Figure 2.1)

According to Mussatto, the content of the total phenolic compound in methanol extract of SCG is similar with the contents in raspberry, blackberry (Wang and Lin, 2000), and

almond shells. Also, they revealed that SCG includes chlorogenic acid, protocatechuic acid and flavonoids leading to antioxidant activity higher than the other fruits such as apple, peach, pear, and kiwi (Moure et al., 2007). Therefore, the coffee is laden with bioactive components including polyphenols and flavonoids with their antioxidant activities making them an attractive candidate for studies on naturally occurring “personal care products”

(Kilmartin and Hsu, 2003). In addition to antioxidant activity, the use of SCG is emerging as the trends in the dermatological field. For example, chlorogenic derivatives 4,5- dicaffeoylquinic acid have been known to be a pigmentation inhibitor in both melanocytes and the zebrafish vertebrate model (Russo et al., 2006). Specially, the inhibitory activities against melanogenesis of this compound was superior to arbutin, and kojic acid, which were well-known inhibitors against melanogenesis (Takagi et al., 2014). According to

Tabassum, the compound has no toxicity so that it can be considered as an attractive candidate for further development in pharmaceutical or cosmetic depigmentation

(Tabassum et al., 2016). Besides, SCG showed effectiveness in UV-induced photo-aging even though specific compounds have not been studied for this reaction (Choi et al., 2015).

However, the study considered caffeine and chlorogenic acids as health-promoting factors because they have protection efficacy against UV-induced skin damage (Huang et al.,

1997) (Kitagawa et al., 2011) and as most abundant compounds in SCG. Even though specific flavonoid contents except quercetin in SCG have not been reported yet, the total flavonoid concentration was estimated by colorimetric assay. Also, the effect of roasted coffee has been extensively studied and widely marketed with its antibacterial activity. The activity which is not present in green coffee was only shown in roasted coffee beans depends on the roasting degree (Daglia et al., 1994a). During the roasting process, various

types and concentrations of compounds were obtained having antibacterial activity against a wide range of bacteria (Daglia et al., 1994b) and they have been considered as nontraditional food antimicrobial agents (SHIBASAKI, 1982). Furthermore, a prior study reported that the lipid fraction of SCG including fatty acids such as Palmitic and Linoleic extracted with supercritical carbon dioxide showed improved skin lipids and hydration qualities (Ribeiro et al., 2013).

These antioxidant and antibacterial properties are extremely important in the cosmetic industry especially in natural cosmetics. The most important factor in the initiation of several skin disorders from dryness, wrinkling, localized pigment abnormalities including, hypopigmentation and hyperpigmentation to skin cancer is the chronic exposure of the skin to solar UV radiation (Lee et al., 2008) (Nichols and Katiyar,

2010). The exposure to UVA and UVB induce the generation of singlet oxygen and hydroxyl-free radicals. This oxidative stress can cause damage to cellular molecules, such as proteins, lipids and DNA (Lee et al., 2008) leading to premature aging of the skin (Nichols and Katiyar, 2010), called photo-aging, and multiple effects on the immune system (Ullrich, 1995). The increasing interest in flavonoids is due to their broad efficacy in pharmacological and dermatological field. The mechanism of action of flavonoids as antioxidants includes their ability to scavenge free radicals directly, activate antioxidant enzymes, chelate metals, reduce alpha-tocopheryl radicals, inhibit oxidases

(Procházková et al., 2011). Free radical scavenging ability is highly related to chemical structure of flavonoids such as high number of hydroxyl groups (Burda and Oleszek,

2001). In terms of activating antioxidant enzymes, flavonoids can induce detoxifying enzymes which regulate protective gene expression (Lee-Hilz et al., 2006). For examples,

flavonoids quercetin, myricetin containing 3-position OH induce detoxifying enzymes against toxic chemical reaction (Ferrali et al., 1997). Quercetin also interferes with the development of free radicals by chelating active form of metal (Ferrali et al., 1997). In addition, flavonoids including kaempferol, myricetin, quercetin, and catechin can reduce prooxidant such as alpha-tocopherol radicals by interacting with them even though the activity was varied each other (Zhu et al., 2000). Lastly, flavonoids (quercetin, and luteolin) inhibit the enzymes related to superoxide so that reducing oxidative injury in cells (Hanasaki et al., 1994).

Related to antimicrobial effect, antimicrobial ingredients should be formulated to protect the personal care products against the growth of micro-organisms such as bacteria, and fungi. By adding anti-microbial compounds in cosmetic formulation, the products can be safe from causing skin infections, or discoloration of the product itself during the usage. Furthermore, antimicrobial agents in the products make consumers can be protected from bacterial disease. Anti-microbial agents against gram positive bacteria such as Propionibacterium acnes, Staphylococcus aureus, and Staphylococcus epidermidis helps to prevent skin illnesses, from minor skin infections including pimples, impetigo, boils, cellulitis, folliculitis, carbuncles, and abscesses, to life-threatening diseases such as toxic shock syndrome, bacteremia, and sepsis (Hanasaki et al., 1994) .

These bacterial strains are the major causes of acne in human skin (Nishijima et al., 2000) and acne is the most common skin issue in cosmetics, affecting up to 50 million

Americans annually (Bickers et al., 2006). Numerous research groups have sought to elucidate the antibacterial mechanisms of selected polyphenols. Inhibition of nucleic acid synthesis is one of the modes of actions because B ring of the flavonoids can form

interaction with nucleotide bases in microorganism. Flavonoids such as myricetin and epigallocatechin, can affect RNA synthesis of S. aureus (Mori et al., 1987). Enzyme binding of flavonoids also contributes to antimicrobial activity by inhibiting DNA gyrase in bacteria (e.g. quercetin and apigenin) (Ohemeng et al., 1993). Glycosylated rutin are also inhibitor of nucleic acid synthesis in E.coli (Bernard et al., 1997). Another way to prevent bacterial infection is by reducing membrane fluidity of bacterial cells.

Epigallocatechin gallate shows strong antibacterial activity inducing aggregation of cell membrane (Ikigai et al., 1993). In addition, flavonoids are reported to inhibit bacterial respiratory electron transport chain, called energy metabolism (Haraguchi et al., 1998).

Even though non-flavonoids also shows antimicrobial activity, it is weaker than flavonoids. However, some phenolic acids including caffeic acid, ferulic acid and gallic acid are notified that they have antibacterial activity against both Gram-positive such as

S. aureus and L. monocytogenes and Gram-negative such as E. coli and Pseudomonas aeruginosa showing stronger activity compared to general antibiotics including gentamicin and streptomycin (Saavedra et al., 2010). It is critical to evaluate the antimicrobial activities of SCG against gram positive stains to explore its potential for cosmetics product formulation for acne treatment or product protection.

These days, a lot of studies are being conducted related to antioxidant derived from natural resources. Therefore, the bioactive compounds such as polyphenolic acids, flavonoids, vitamins having antioxidant activity are identified (Núñez-Sellés, 2005). To evaluate their ability to inhibit oxidation, several assays are developed. Based on research types, adequate tests should be selected between lipid peroxidation and electron/radical scavenging associated with oxidation inhibition. Since reactive oxygen species (ROS) are

hard to measure due to their extremely short half-lives, other oxidative stress products can be measured. As a result of oxidative stress, thiobarbituric acid reactive substances

(TBARS) are formed. Using thiobarbituric acid, TBARS can be detected and this test is called TBA assay (Hodges et al., 1999). Another antioxidant assay associated lipid peroxidation is beta-carotene bleaching assay; beta-carotene can react with peroxyl radical which is formed from lipids such as linoleic acid in the presence of ROS.

Antioxidant activity is measured by reduction amount of beta-carotene (Tsuchihashi et al., 1995). Conjugated diene assay is monitoring conjugated diene which occurs by the action related to ROS and oxygen (Moore and Roberts, 1998). This method has weakness because natural compounds have similar absorbance wavelength with conjugated diene

(Moon and Shibamoto, 2009). Antioxidant assays associated with electron and radical scavenging include DPPH (2,2-diphenyl-1-picrylhydrazyl), ABTS (2,2'-azino-bis(3- ethylbenzothiazoline-6-sulphonic acid)), and FRAP (ferric reducing antioxidant power) assay (MacDonald‐Wicks et al., 2006). DPPH•, one of few stable and commercially available radicals accepts hydrogen from an antioxidant based on the theory that antioxidant act as a hydrogen donor. Antioxidant activity of many natural compounds are evaluated by ABTS assay as well due to accurate and easy method especially for fruit and vegetable extracts (Sanchez-Moreno, 2002). By measuring ABTS radical cation which absorbs blue-green wavelength, antioxidant ability can be determined (Re et al., 1999). In

FRAP assay which gives fast, reproducible results, antioxidant activity can be observed by looking transformation of Fe3+-TPTZ complex to Fe2+ form by an antioxidant (Moon and Shibamoto, 2009). The FRAP has been recognized as one of the most accurate and reliable methods to evaluate the antioxidant activities of the natural product. Multiple

natural antioxidant studies using FRAP assay were showed consistent result with DPPH and ABTS assays (Katalinic et al., 2006).

In terms of antimicrobial assays, there are several methods to view the activities.

Very common assays include disk diffusion and agar dilution (Balouiri et al., 2016).

Since microbial resistance is getting more important in drug discovery, dermatology, and cosmetics, in vitro antimicrobial test is prevalently used. Both crude extracts and pure compounds could be evaluated for their application in personal care industry (Runyoro et al., 2006). Among several methods, agar disk diffusion testing is officially used in many clinics and laboratories (Balouiri et al., 2016). Once antimicrobial agent diffuses into plates and inhibits growth of microorganism, the zone of inhibition are measured (Ahmad and Beg, 2001). Even though the agar disk diffusion method cannot quantify the antimicrobial agent in agar plates, this is the most common method due to its low cost, simplicity, availability of vast number of testing samples in one plate and the ease of result interpretation (Korgenski and Daly, 1998). Other techniques including flow cytofluorometric method and time-kill test have pros and cons. For example, the former technique is useful to give reproducible results rapidly (Ramani and Chaturvedi, 2000) with information on cell damage of the tested microorganism but unlikely used due to required equipment.

Bioassay-guided purification has been often used as an approach to isolate the bioactive compounds in the natural extracts such as Chinese green tea (Si et al., 2006),

Brazilian propolis (Castro et al., 2009), and lemon balm (Awad et al., 2009). With this approach, the compounds were first separated by their hydrophobicity and polarity based on their interaction with the functional group on the surface of the solid-phase on the

liquid chromatography column. The eluted fractions were collected, concentrated and tested for their bioactivity until the lead compounds are isolated (Evans, 2009). Since plant extracts are being considered for their potential use as natural preservatives in dietary supplements or personal care products, it is needed to study biological properties of the compounds in extracts (Seeram et al., 2005). Unfortunately, purifying pure compound from crude extract for compound structure studies such as Nuclear Magnetic

Resonance (NMR) requires a lot of time and intensive labor using chromatographic technique. As a solution, high-resolution MS metabolomics and natural product libraries was introduced for fast identification of the compounds in the bioactive fractions without labor intensive purification required in NMR structural analysis. The “World’s most cited metabolomics software called XCMS Online” cited in literature over 1000 times (Ubhi et al, 2014), which includes adequate algorithm in peak detection, retention time correction, alignment, annotation, and statistical analyses was introduced for compound identification. Only basic steps are needed for metabolomics analysis after the fractionation of the extract using liquid chromatography coupled with mass spectrometer

(LC-MS). Once researchers obtain ion chromatogram of fractionated sample from MS, users can submit it to web-based metabolomics platform to accomplish chemical profiling and acquire structural information. According to its peak detection algorithm, it covers both low-resolution, and high-resolution using matchedFilter method and centWave method respectively (Smith, 2013). After peak detection, XCMS Online uses nonlinear methods to compensate for retention time drifts between samples (Tautenhahn et al., 2012). All total ion chromatogram (TIC)s should be in alignment to compare compound quantities between fractions after retention time correction (Christin et al.,

2008). If there is significant difference in the peak between submitted groups, one group can be distinguished from another group. XCMS Online uses minimum fraction method to process this step. For example, minfrac value 0.5 (50%) means, two groups can be considered as two distinct groups when they show at least 50% difference. In annotation step, information about at least two ions is need to calculate the molecular mass which is necessary to be matched with compound library (Kuhl et al., 2011) called METLIN. The fingerprint of compounds are matched with the database which has more than 961,820 molecules providing information for each compound such as name, systematic name, structure, elemental formula, and mass (https://metlin.scripps.edu/).

Through bioassay (e.g. antioxidant and antimicrobial assay) -guided purification method followed by global metabolomics approach, SCG can be assessed as potential raw material in cosmetic field. Especially, high resolution MS metabolomics was introduced to identify pure compound in fraction while reducing time and intensive labor.

1. Chlorogenic acid 2. Caffeic acid 3. Isochlorogenic acid

4. Protocatechuic acid 5. 4,5-dicaffeoylquinic acid 6. Quercetin

Figure 2.1. Common bioactive compounds in SCG

1 3-Caffeoylquinic (Chlorogenic) acid Antioxidant1

2 3,4-Dihydroxycinnamic (Caffeic acid) Antioxidant1

3 3,4-Di-O-caffeoylquinic acid (Isochlorogenic acid) Antioxidant1, Anti-inflammatory8

4 Dihydroxybenzoic acid (protocatechuic acid) Antioxidant2, Anti-fungal4

5 4,5-dicaffeoylquinic acid Pigmentation inhibitor3

6 Quercetin Anti-fungal4, Anti-oxidant5, Skin protection6, Anti-melanogenesis7

1: (Sroka and Cisowski, 2003)

2: (Li et al., 2011)

3: (Russo et al., 2006)

4: (Aziz et al., 1998)

5: (Boots et al., 2008)

6: (Choquenet et al., 2008)

7: (Arung et al., 2011)

8: (Park et al., 2009)

2. Objectives

In chapter 2, the bioassay- guided fractionation approach was used including anti-oxidant and anti-microbial assay to purify the bioactive compounds from SCG extraction.

The specific objectives of this study are described as following

1. Perform bioassay guided-purification to explore new skin-care application.

2. Isolate the bioactive compounds in SCG by flash liquid chromatography fractionation

followed by HPLC.

3. Perform FRAP anti-oxidant assay and zone of inhibition anti-microbial assay to

confirm and evaluate the activity.

4. Compare the difference in antioxidant and antimicrobial activities between selected

three cultivars.

5. Identify potential compounds in active fractions using high-resolution MS

metabolomics.

3. Materials and Methods

The health-promoting properties of coffee are often attributed to its rich phytochemistry, including caffeine, chlorogenic acid, caffeic acid, hydroxyhydroquinone (HHQ). Since coffee contains multiple bioactive compounds such as anti-oxidant, anti-inflammatory, anti-microbial, and anti-tumor, we need to isolate and identify each compound from whole extract by checking each fractions from chromatography. Liquid chromatography including Flash chromatography and HPLC enable to generate fractions from extract having different purpose. Flash chromatography helps to fractionate the extract into vials which contains around 100 compounds. Further steps are need to purify the one hundred

compounds using HPLC. In this stage, we can narrow down into only one or two compounds in one vial.

For this study, the SCG extract from three cultivars, including Ethiopian Yirgacheffe,

Costa Rican Tarrazu, Hawaiian Kona) were evaluated for their antioxidant and antimicrobial activities by FRAP antioxidant assay and zone of inhibition assay respectively.

As described in the Figure 2.3, the bioactive compounds in SCG were purified by HPLC after fractionated by flash liquid chromatography. The purified compounds were characterized by ultra-pressure liquid chromatogram (UPLC) coupled with high- resolution mass spectrometry (HRMS). The ion chromatograms were submitted to

XCMS, and the identified signals were annotated by comparing with the METLIN metabolomics library.

SCG extrac ts

Extraction Extract non-volatile compounds from SCG using methanol

Flash chromatography, HPLC Collect fractions from liquid chromatography

Antioxidant, antimicrobial assay Perform assays of each fraction to identify the active one

UPLC-HRMS analysis Identify non-volatile chemical composition in active fractions

Metabolomics approach (XCMS Online) Perform metabolite profiling in active fractions

Figure 2.2. Flow chart of the SCG project using bioassay guided purification

3.1. Explore new application (antioxidant and antimicrobial activities) through bioassay-guided purification

Anti-oxidant analysis

3.1.1. Sample preparation

SCG from three Arabica coffee cultivars including Ethiopian Yirgacheffe, Costa Rican

Tarrazu, and Hawaiian Kona were selected for this study. Roasted and grinded coffee were purchased from the local Lakota Coffee (Columbia, Missouri). After brewing for 5 minutes, SCG were collected. Twenty-five grams of wet SCG (78% water, about 5.5 g DW equivalent) were homogenized with blender and the compounds were extracted with 200 ml methanol (HPLC-grade, purchased from Fisher Scientific, Pittsburg, PA, USA). The extract was then sonicated for 60 mins and filtered through Whatman phase separators

(silicone treated filter paper, diameter 125mm). The methanol was evaporated by rotary evaporator. The phenolic compounds were extracted by a sequential water: chloroform liquid/liquid extraction with 200ml of chloroform (HPLC-grade, purchased from Fisher

Scientific, Pittsburg, PA, USA). Following the liquid/liquid extraction, the chloroform fraction was collected and the chloroform was removed by a rotary evaporator (BUCHI,

Rotavapor R 110). The powder form of final extract was loaded on the flash chromatography column packed with 34g of C18 resin (Specific surface area: 460-520 m2

/g, average particle size: 47-60 µm, average pore diameter: 60-87 Å)

3.1.2. Fractionation and bioassay-guided purification

The compounds in the extract was fractionated by a Jones Flash Master II Flash

Chromatography connected with a fraction collector. The mobile phases consist of

methanol and deionized water. Compounds were separated by a gradient described as following: 25% in the first 40 min, followed by a linear gradient from 25% to 50% solvent

B (Methanol) for 50 min, then followed by a linear gradient from 50% to 75% B for 30 min, and then remain 75% for 60 min before achieve the 100% conditions. The flow rate was maintained at 2.5 mL/min.

For further purification, the bioactive fractions collected from flash chromatography were injected into a Shimadzu SCL-10Avp HPLC system with a SPD-M10Avp photodiode array detector coupled with a Shimadzu FRC-10A fraction collector. The compounds were separated by a Phenomenex (Torrance, CA) Columbus C8 (250mm x 4.6 mm; 5 µm particle size) reverse-phase column. The mobile phases consist of ACN and deionized water by following gradient: 20% to 80% for 16 min with a linear gradient, then followed by a linear gradient from 80% to 98% solvent (ACN) for 2 min, and then remain 98% for

1 min before go back to 20% at 20 min. The gradient remain 20% for 9 min. The total running time was 30 min with 0.5 mL/min of the flow rate. The signals were monitored at wavelength of 220 and 254 nm. The injection volume was 50 µL. Each fraction was collected, and the fraction with the same retention time window were combined after injection of the extracts 30 times.

Figure 2.3. Flow chart of the bioassay-guided purification for identifying antioxidant activity and antimicrobial assay. Same approach was applied for antimicrobial test.

3.2. Determination of anti-oxidant activities

Each fraction was tested for the anti-oxidant activities using FRAP™ Detection Kit for anti-oxidant activity. All fractions were prepared with three dilution factors including

1:10, 1:100, and 1:1000 and raw extracts were prepared with two dilution factors including 1:25 and 1:50 to be within the range of the calibration curve (Figure 4). 20ul of samples or STDs into wells were pipetted in 96 well. 20ul of buffer was used as Zero

STD. After all samples was loaded, 75ul of Color Solution to each well was added using repeater pipet. The absorbance value at optical density 560 nm was read after incubating at room temperature for 30 minutes. FRAP assay uses antioxidants as reductants in redox-linked colorimetric method. Ferric (Ⅲ) [colorless] to Ferrous (Ⅱ) [blue] can be monitored by measuring absorbance at 560 nm. The absorption readings are related to the reducing power of the electron-donating antioxidants present in the test compound.

Hence the FRAP assay can rank the reducing power and the antioxidant potential of a wide range of test compounds. Calibration curve represents FeCl2 concentration (µM) in

X-axis and optical density (nm) in Y-axis. Seven standards solutions including 0, 31.25,

62.5, 125, 250, 500, and 1000 (µM) were used for the development of the calibration equation.

3.3. Anti-bacterial analysis

3.3.1. Sample preparation

Twenty-five grams of wet SCG (78% water, about 5.5 g DW equivalent) were homogenized with blender and extracted with 200 ml methanol. The extract was then sonicated for 60 mins and filtered using phase separators (silicone treated filter paper,

diameter 125mm). Methanol was evaporated by nitrogen evaporator and samples were redissolved in 0.5ml of DMSO (550 mg/ml). Raw extracts including Ethiopian

Yirgacheffe, Costa Rican Tarrazu, and Hawaiian Kona using same method as antioxidant assay for bioassay-guided purification were used for zone of inhibition assay.

Fractionated 43 samples through Flash chromatography dissolved in 100 µl of DMSO after evaporating methanol were also tested. With agar diffusion method, approximately one million cells from a single strain of Staphylococcus aureus were spread over an agar plate, then incubated in the presence of 10 µl of the extracts or fractions. If the bacterial strain is susceptible to the antimicrobial agent, then a zone of inhibition appears on the agar plate. If it is resistant to the antimicrobial agent, then no zone is evident. DMSO was used as a control. The size of the zones were measured to quantify and evaluate the antibacterial activities.

3.4. High resolution (HR) metabolomics analysis

The compounds in the purified bioactive fractions were analyzed by UPLC-HRMS and data was processed by XCMS Online. The signal was annotated by comparing the

METLIN metabolomics library. The purified fraction was analyzed by a Bruker maXis impact quadrupole-time-of-flight mass spectrometer coupled to a Waters ACQUITY

UPLC system. Separation was achieved on a Waters C18 column (2.1x 100 mm, BEH

C18 column with 1.7-um particles) using a linear gradient of 95%:5% to 30%:70% eluent

A:B (A: 0.1% formic acid, B: acetonitrile) over 30 min. Between 30-33min, the a linear gradient was increased from 70% to 95% B, and maintain at 95%B for 3 min. The percentage of B was maintained at 5% from 36-40 min. The flow rate was 0.56 mL/min

and the column temperature was 60 C. Mass spectrometry was performed in the negative

(or positive) electrospray ionization mode with the nebulization gas pressure at 43.5 psi, dry gas of 12 l/min, dry temperature of 250C and a capillary voltage of 4000V. Mass spectral data were collected from 100 to 1500 m/z and were auto-calibrated using sodium formate after data acquisition. The generated spectra were submitted to XCMS analysis.

The pairwise analysis between the control (ACN) and purified fraction was performed for the background subtraction. Parameters in high resolution analysis were: maximal tolerance of 10 ppm between successive measurements, peak width of 5-20 s, S/N threshold of 6 and obiwarp algorithm for retention time correction and alignment (width of overlapping m/z slices 0.015, minimum fraction of samples 0.5 for group validation).

To visualize differences between two samples in pairwise job, 0.001 of p-value threshold and 1.5 of fold change was selected to be considered highly significant. The bioactive compounds were identified by comparing the accurate mass of the molecular ions with the candidates in the METLIN database.

In metabolomics, two group comparison is very common way to assess difference between treated versus control. In the study, the control indicates ACN (purchased from Fisher

Scientific, Pittsburg, PA, USA), solvent which is used in UPLC Bruker Q-TOF (High resolution MS) system. In XCMS Online pairwise test, comparison between ACN and purified F13 was analyzed. This process was conducted in both positive and negative modes. Once UPLC/Bruker QTOF (HRMS) was marked, it automatically activates centWave algorithm, which is new published algorithm designed for centroid and high resolution data for both two-group comparison study. Parameters in high resolution analysis were: maximal tolerance of 10 ppm between successive measurements, peak width

of 5-20 s, S/N threshold of 6 and obiwarp algorithm for retention time correction and alignment (width of overlapping m/z slices 0.015, minimum fraction of samples 0.5 for group validation). To visualize differences between two samples in pairwise job, 0.001 of p-value threshold and 1.5 of fold change was selected to be considered highly significant.

3.5. Statistical analysis

All data including antioxidant and antimicrobial assay are presented as mean values

(±S.E). Values within columns followed by different superscript letters (A and B) are statistically different (P < 0.05). One-way analysis of variance (ANOVA) followed by pairwise comparisons of means was performed to assess the difference in three cultivars.

Pairwise comparison of means was analyzed by Tukey’s range test. R: The R Project for

Statistical Computing 3.3.1 was used for both statistical analysis. Samples were analyzed with 2 dilution factors and sextuplicate (n = 6) for antioxidant assay. In zone of inhibition assay, the diameter was measure three times for each samples with triplicate. However, the maximum and minimum value was eliminated for statistical analysis. Therefore,

Septuplicate (n = 7) were used for each cultivars in ANOVA.

4. Results

4.1 Antioxidant study

The FRAP analysis was successfully utilized to evaluate the antioxidant activities of raw

SCG extracts and the isolated fractions, suggesting the correlation coefficient of 0.9997 with the calibration range from 0-1000 µM of FeCl2 (Figure 2.4). The results of the study suggested that the SCG from Hawaiian Kona had highest antioxidant, about 23% higher

as compared to Yirgacheffe (Figure 2.5 and Table 2.1). The FRAP (Ferric Reducing

Antioxidant Power) assay measures the ability of an antioxidant to reduce ferric (III) to ferrous (II) in a redox-linked colorimetric reaction (Grover and Patni, 2013). Antioxidant induces the reaction producing ferrous (II) which has blue color from ferric (III). The reducing power of a compound serves as a significant indicator of its potential antioxidant activity. In investigated antioxidant results, the reducing power was higher in higher concentration of the extracts (Table 2.1). Interestingly, SCG extracts from Costa

Rican Tarrazu and Hawaiian Kona have significantly higher antioxidant activity (p <

0.05) compared to extracts from Ethiopian Yirgacheffe even though there was no significant difference between Tarrazu and Kona in 1:25 concentration. Same aspect was assessed in lower concentration (1:50). The antioxidant result value in absorbance was converted into value of FeCl2 concentration using equation y = 0.002x - 0.0054 generated by standard curve (Figure 2.4). Figure 2.5 shows FeCl2 concentration among three cultivars and Hawaiian Kona has highest reducing capacity.

Among the 43 fractions of the Ethiopian Yirgacheffe SCG extracts collected from the flash chromatography fractionation process, fraction number 4 showed the highest antioxidant activity (Figure 2.6). The bioactive compounds in the fraction 4 was further purified by the HPLC and 26 sub-fractions were obtained. After all fractions were tested through FRAP assay, sub-fraction 14 collected from HPLC separation showed the highest antioxidant activity among 26 fractions (Figure 2.7).

FRAP assay 2.5

y = 0.002x - 0.0054 2 R² = 0.9997

1.5

1 Series1

OD(nm) Linear (Series1)

0.5

0 0 200 400 600 800 1000 1200

-0.5 FeCl2 concentration (µM)

Figure 2.4. Standard curve of FRAP assay. Each dot in X axis represents around 0, 31.25, 62.5, 125, 250,

500, 1000 FeCl2 concentration of standards (STD; n = 7).

Table 2.1. Antioxidant activity of SCG methanol extracts of three cultivars including Ethiopian

Yirgacheffe, Costa Rican Tarrazu, and Hawaiian Kona

FRAP (abs. at 560 nm)

Dilution factors

Cultivar 1:25 1:50

Ethiopian Yirgacheffe 0.469±0.023a 0.310±0.008a

Costarican Tarrazu 0.553±0.008b 0.350±0.008b

Hawaiian Kona 0.551±0.016b 0.353±0.008b

Means in the same column followed by different letters (a-b) represent statistical significance (P <0.05).

Samples were analyzed with 2 dilution factors (1:25 and 1:50) and sextuplicate (n = 6). One-way analysis of variance (ANOVA) followed by pairwise comparisons of means using Tukey’s range test was performed to assess the difference in three cultivars (significance level P < 0.05). R: The R Project for Statistical

Computing 3.3.1 was used for statistical computing.

FRAP assay 7000

AB 5596.56 ± 266.49B 6000 5530.94 ± 235.55

M) 4479.37 ± 235.55A µ 5000

4000

3000 concentration(

2 2000 FeCl 1000

0 Ethiopian Costa Rican Hawaiian Cultivar

Figure 2.5. Reducing power of methanol extract (10mg/mL) from three SCG cultivars including Ethiopian

Yirgacheffe, Costa Rican Tarrazu, and Hawaiian Kona were compared. Concentration of FeCl2 was calculated from FRAP value (absorbance in 560nm) using y = 0.002x - 0.0054. Hawaiian Kona had highest antioxidant compared to other cultivars.

Means in the same column followed by different letters (A-B) represent statistical significance (P <0.05).

Sextuplicates were analyzed using one-way analysis of variance (ANOVA) followed by pairwise comparisons of means using Tukey’s range test was performed to assess the difference in three cultivars (n

= 6, significance level P < 0.05). R: The R Project for Statistical Computing 3.3.1 was used for statistical computing.

Frap Assay from Flash Chromatography

500 475.16 450 400 350

300 (uM)

2 250

FeCl 200 150 100 50 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 Fraction Number

Figure 2.6. Antioxidant activities of the fractions from flash chromatography (1:10 dilution). The fraction number 4 showed highest antioxidant activity as 475.16 µM FeCl2 condition among 43 fractions so F4 was injected into HPLC for further purification.

FRAP assay from HPLC 90 80 70 60 50 40 30 20 10

Vial3 Vial0 Vial1 Vial2 Vial4 Vial5 Vial6 Vial7 Vial8 Vial9

Vial26 Vial11 Vial12 Vial13 Vial14 Vial15 Vial16 Vial17 Vial18 Vial19 Vial20 Vial21 Vial22 Vial23 Vial24 Vial25 Vial10

Figure 2.7. 26 fractions was obtained from HPLC and tested through FRAP assay. Fraction 14 (vial 13) was revealed as bioactive fractions with high antioxidant activity.

4.1.2. High resolution metabolomics analysis of flash chromatography fractions

Purified bioactive fractions 13 was used to be analyzed for identification of compound in using global metabolomics platform, XCMS Online. The reason why fraction 14 was selected for high resolution metabolomics analysis is it showed highest anti-oxidant activity among 26 HPLC fractions.

Table 2.2 represents identified bioactive compounds in fraction 14 purified by HPLC.

The result was generated through XCMS Online high-resolution metabolomics analysis.

In fraction 14, several compounds which were used as fragrances in cosmetics. For example, linalyl cinnamate was detected at RT 12.75 and piperonal acetone was found at

RT 16.43. Additionally, coumarin derivative including hydroxycoumarin, and umbelliferone were detected. These bioactive compounds have anti-viral (Cingolani et al., 1969), anti-oxidant (Bailly et al., 2004), and photo protective (de Araujo Leite et al.,

2015) effects. Antioxidant (Ramesh and Pugalendi, 2006), anti-inflammatory (Lino et al.,

1997) and anti-mutagenic (Ohta et al., 1983) effects were reported from studies on umbelliferone. In fraction 14, another bioactive compound were found at RT 11.05 with m/z 151.0754. Especially, hydrocinnamic acid is widely used in food and perfume industry as fixative agent, or a preservative to maintain the original scents. It also acts as an emulsifier. In cosmetics, it is used frequently as softeners, and soaps providing floral scent (Grover and Patni, 2013). Another compound cafestol was identified as highly possibility among other candidate compounds such as fragrant compounds, coumarin derivatives, and hydrocinnamic acid. The peak intensity was the highest for cafestol compared to other compounds in same fraction. Cafestol is one of the coffee-specific diterpenes which has been reported as antioxidant agents. According to the Lee and

Jeong, cafestol is markedly effective in protection against H2O2- induced oxidative stress and DNA damage due to its free radical scavenging activity as strong antioxidants.

Table 2.2. Identified compounds in purified fraction 14 through XCMS Online HR metabolomics analysis.

Each row indicates feature, compound name, m/z, retention time, maximum intensity and its known bioactivity. Using XCMS Online platform, chemical profiling was compiled.

Max F14POS_Hconc_HR m/z RT bioactivity intensity

1329 Cafestol 679.9 11.83 618,624 Antioxidant

5 Linalyl cinnamate 285.1853 12.75 16,010 Fragrance

6 Piperonal acetone 191.0704 16.43 85,550 Fragrances

Anti-viral,

28 3-Hydroxycoumarin 163.0391 13.19 43,960 antioxidant,

photo protective

Antioxidant,

anti- 28 Umbelliferone 163.0391 13.19 43,960 inflammatory,

anti-mutagenic

29 Hydrocinnamic acid 151.0754 11.05 11,982 Anti-tyrosinase

Figure 2.8. Total ion chromatogram (TIC) of fraction 14 purified by HPLC. The TIC was generated by XCMS Online positive ion mode. The highest

peak intensity 618,624 was shown at RT 11.83 with m/z 679.9 and cafestol was the best possibility among other compound candidates.

Besides SCG extracts samples, we also evaluated the pure compounds that have been reported in the SCG or coffee using FRAP assay. Table 2.3 shows ferric reducing power of bioactive compounds which were reported as antioxidant in previous studies.

Interestingly, some compounds including quinic acid, resveratrol, chlorogenic acid, 4- hydroxybenzoic acid, caffein, chryin, apigenin, rutin hydrate, nargine P-coumaric acid, and hesperetin had no reducing ability through FRAP assay. Guaiacol showed highest reducing power followed by walnut sample, gallic acid, quercetin, and myricetin.

Table 2.3. FeCl2 concentration (µM) of bioactive compounds which were reported in previous study

(unpublished data by Nahom Taddese Ghile and Lin).

List of samples analized Ferrous Chloride Con.(µM)

Quinic acid 0

Resveratrol 0

Chlorogenic Acid 0

Gallic Acid 71.7

Quercetin 61.1

4-Hydroxybenzoic acid 0

Caffein 0

Chryin 0

Apigenin 0

Guaiacol 109.29

Rutin Hydrate 0

Naringin 0

Myricetin 18.62

P-coumaric acid 0

Hesperetin 0

walnut 74.81

Guaiacol showed highest Ferrous Chloride concentration which means highest antioxidant activity among samples. Walnut samples, Gallic acid, and Quercetin showed an intermediate antioxidant activity, followed by myricetin. quinic acid, resvertarol, chlorogenic acid, 4-hydroxbenzoic acid, caffein, chryin, apigenin, rutin hydrate, nargine, P-cormaric acid, and hesperetin did not have detectable value from the FRAP assay.

4.2 Antibacterial activity against Methicillin-resistant Staphylococcus aureus

(MRSA)

All the SCG extracts from the 3 varieties, have shown a strong or moderate antibacterial activity against Methicillin-resistant Staphylococcus aureus (MRSA) in our antimicrobial activity in zone of inhibition test (Figure 2.9). For this assay, the inhibition zone diameters around each of the disc (diameter of inhibition zone minus diameter of the disc) were measured. The inhibition zones were 8 mm or greater, suggesting strong antibacterial activity.

The SCG extracts from Hawaiian Kona and Yirgacheffe showed the highest antimicrobial activity. The mean values of the antibacterial activities of the SCG extracts of all the investigated cultivars is shown in Figure 2.10. The mean value of zone of inhibition was 8.85 mm in Hawaiian Kona sample followed by 8.19 and 5.76 in

Yirgacheffe and Tarrazu respectively. The difference between these Hawaiian Kona and

Yirgacheffe was not significant (P < 0.05).

Among the 43 collected fractionated Ethiopian Yirgacheffe SCG extracts generated by flash chromatography, fraction number 7 and 12 showed the highest antimicrobial activity (Figure 2.11). The rest of fractions beyond fraction number 20 that have no activity were not presented in the Figure 2.11. Unlike antioxidant result, Tarrazu showed the lowest antimicrobial activity. Although it is ideal to use multiple bacterial strains to examine anti-microbial activity of SCG, S. aureus causing dermatological issue such as acne was presented in this study due to its importance in cosmetic field.

Figure 2.9 SCG samples (Number from 47 to 55) had strong anti-microbial activity which is Ethiopian

Yirgacheffe, Costa Rican Tarrazu, Hawaiian Kona. The size of inhibition zone was measured three times for all samples. Among three cultivars, Kona showed the highest anti-microbial activity having 8.85 mm inhibition zone.

Zone of Inhibition 12

10 8.85 ± 1.57B 8.19± 0.97B 8

5.76 ± 0.66A 6

4 Zoneof Inhibition (mm)

0 0 Yirgacheffe Tarrazu Kona DMSO Cultivars

Figure 2.10. Antimicrobial activities of the E (Ethiopian Yirgacheffe), C (Costa Rican Tarrazu), H

(Hawaiian Kona). Using zone of inhibition assay for assessing anti-microbial activity against S. aureus, we found the following order of efficiency among coffee Arabica cultivars: Kona (8.85) > Yirgacheffe (8.19)

>> Tarrazu (5.76). DMSO did not show any inhibitory activity. Inhibition zone: 8 mm or greater — good antibacterial activity; 6–7 mm — moderate antibacterial activity; 4–5 mm — weak antibacterial activity; 2–

3 — very weak antibacterial activity.

Mean values followed by different letters (A and B) represent statistical different (P <0.05). Zone of inhibition is expressed as the diameter (mm). Values represent means ± standard deviations of inhibition zones from experiments, each using triplicate plates and was measure three times. Means with the same superscripts (A, and B) are not significantly different (P < 0.05) from each other. One-way analysis of variance (ANOVA) followed by pairwise comparisons of means was performed to assess the difference in three cultivars (n = 7, significance level P < 0.05). R: The R Project for Statistical Computing 3.3.1 was used.

zone of inhibition from flash chromatography 7 6.05 6.05 5.94 5.83 6 5.6 5.66 5.5 5.4 5 5

4 mm 3

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Fraction number

Figure 2.11. Antimicrobial activity of 43 fractions using flash chromatography from Ethiopian Yirgacheffe.

Fraction number 7 and 12 showed highest zone of inhibition value. The rest of fraction number over 20 that have no activity doesn’t represent in the chart.

5. Discussion

5.1. Antioxidant activity of SCG for cosmetic application

5.1.1. Comparison between three cultivars (Yirgacheffe, Tarrazu, and Kona) in antioxidant properties

Among three cultivars of SCG samples, Hawaiian Kona showed the highest antioxidant activities (Figure 2.5). Even though the activity doesn’t show much difference between

Tarazzu and Kona, their reducing power were significantly (significance level P < 0.05) higher than Yirgacheffe (Table 2.1). All three SCG extracts showed dose-dependent aspect in FRAP anti-oxidant assay. The reducing power was increased with the increasing concentrations. It means that at higher concentration (1:25), the reducing power was stronger than low concentration (1:50). The concentration of raw extracts was solvent ratio 100 (ml/g) and their FRAP value (mM Fe(II)/g SCG) was from 0.814 mM to

1.0 mM of FeCl2. Our results showed linear correlation to the other research. According to Mussatto, in methanol extracted SCG, sample solvent/solid ratio 10 (ml/g) showed

0.021 (mM Fe(II)/g SCG) FRAP value and sample solvent/solid ratio 30 (ml/g) showed

0.116 (mM Fe(II)/g SCG) FRAP value (Mussatto et al., 2011a). When the antioxidant activities of SCG were compared to other research on tea antioxidants which is widely known as strong antioxidant source containing large amounts of polyphenolic compounds, SCG have similar or much better reducing power. Their study showed that different types of tea had widely different antioxidant activities, ranging from 132 µmol/g for one type of black tea (“Premium Yun-Nan Puerh Tea”) to 1144 µmol/g for one of the green teas (“China Green Tea”) tested. Also they estimated that the a cup of tea with the highest antioxidant which has similar reducing power with SCG can be act as 100-200mg

of pure ascorbic acid (vitamin C). This highlights the enormous potential of SCG as a strong antioxidant material (Benzie and Szeto, 1999).

5.1.2. High-resolution analysis result of purified fractions 14; possible antioxidant effect

Cafestol was identified as the major antioxidant agent in the purified fraction 14, which had the strongest antioxidant activities among all the purified fractions. Previous research have demonstrated the strong antioxidant activities of cafestol (Lee and Jeong, 2007).

Oxidative stress plays an important role in skin aging. This process is deeply related to reactive oxygen species (ROS), induced by UV light. Antioxidants are major protector against the free radicals coming from oxidative stress. Especially, plant-derived polyphenols are considered as strong antioxidants and the use of them are increasing in cosmetic formulation (Potapovich et al., 2013). Our results are consistent with other studies conducted previously. According to Lee, and Jeong (Lee and Jeong, 2007), cafestol has shown to rapidly react with hydroxyl radicals to protect DNA damage and scavenge superoxide radicals.

5.1.3. FRAP antioxidant assay with the analytical standards

In table 2.3, guaiacol showed highest Ferrous Chloride concentration suggesting the highest antioxidant activity among samples. There are several assays that have been developed to measure antioxidant capacities of the natural products (Thaipong et al.,

2006). For example, Thaipong and Kriengsak (2006), compare and evaluate ABTS,

DPPH, FRAP, and ORAC assays in measuring the antioxidant ability in plant extract.

Among them FRAP assay was revealed as appropriate method due to high reproducibility, simplicity, rapidity in performance and highest correlation with both ascorbic acid and total phenolics (Thaipong et al., 2006). In other research using DPPH antioxidant assay which measures radical scavenging activity, guaiacol known as commercial antioxidants showed strong activity. Similar to their findings, our study also showed that the compound has the highest antioxidant activity when it was evaluated by

FRAP assay (Table 2.3). This strong antioxidant activity associated with the ortho substitution with the electron donor methoxy groups, which is an important factor to stabilize the phenoxy radical leading to antioxidant capacity. Walnut extracts, gallic acid, and quercetin showed an intermediate antioxidant activity, followed by myricetin. On the other hand, quinic acid, resvertarol, chlorogenic Acid, 4-hydroxbenzoic acid, caffein, chryin, apigenin, rutin hydrate, nargine, P-cormaric acid, and hesperetin did not show detectable value from the FRAP assay even though they are studied as antioxidant. This is because quinic acid is pro-metabolite for synthesis of tryptophan and nicotinamide in the gastrointestinal tract possessing DNA repair properties instead of showing direct antioxidant activity in FRAP assay (Pero et al., 2009). One study assessing antioxidant activity from FRAP assay validated quercetin and tannic acid showed the has high antioxidant activity, followed by gallic and caffeic acids. In addition, resveratrol showed the lowest reducing ability of all the studied antioxidants among quercetin, tannic acid, gallic acid, caffeic acid, BHA, rutin, trolox, catechin, ferulic acid, and ascorbic acid. It agrees with the results reported by this paper. Multiple studies demonstrated that resveratrol has less activity than quercetin or epicatechin toward reduction of hexanal

generation from oxidized LDL (Pulido et al., 2000). Furthermore in evaluation of the antioxidant activity of flavonoids using FRAP assay, -OH flavone, 7-OH flavone, chrysin, apigenin, naringenin, and hesperetin showed no or very low activity in the FRAP assay (Firuzi et al., 2005) which is similar aspect as this study. In other antioxidant property test, activity of even naringenin which was higher than its glycoside naringin

(Cavia‐Saiz et al., 2010). In most of comparison study for antioxidant activity, quercetin is located in high-level among flavonoids and phenolic acids. It acts as iron chelator, which maintains antioxidant activity in mechanism and usually included in sunscreen such as EltaMD UV Physical Broad-Spectrum, SPF41, and EltaMD by Swiss-American

(TM/®). Unlike quercetin, the effect of resveratrol in skin wrinkling and skin photo aging has been demonstrated in cellular level (Lee et al., 2016). Several cosmetic products contains resveratrol as plant extracts, or pure compound form in Moisture Reservoir

Hand Cream, Love Life Skin, 46 Graphic Place Moonachie, NJ Estee Lauder Re-Nutriv

Ultimate Youth Crème Reviews (Kusumawati and Indrayanto, 2013). The result of this study supports that SCG can be used in cosmetic formula as active ingredient due to its high concentration of antioxidant phenolic acid and flavonoids.

5.2. Antimicrobial activity of SCG for cosmetic application

Zone of inhibition assay indicates that Hawaiian Kona and Ethiopian Yirgacheffe SCG extract can be considered as great anti-bacterial source compared to Costa Rican Tarrazu cultivars (Figure 2.10). These days, anti-microbial activities become more important for safe use of the products prohibiting germ spread in using cosmetics. By formulating anti-

microbial compounds in cosmetics, the products can be safe from causing skin infections, or discoloration of the product itself during the usage, and after they are opened. Since coffee has been reported as antimicrobial material due to several compounds such as caffeine, chlorogenic acid, protocatechuic acid (Almeida et al., 2006), and melanoidins

(Rufian-Henares and de la Cueva, 2009), we expect that SCG still contains bioactive compounds for preventing spoilage or skin disease from microbial contamination.

Thus, zone of inhibition assay was conducted to evaluate the activity of raw SCG samples and flash chromatography fractions against S. aureus, one of the most common pathogen causing acne. Since disease from S. aureus occurs in hospital starting from skin infection, S. aureus are used prevalently these days for antimicrobial activity evaluation.

In this study, S. aureus strain which is called MRSA that resistant to antibiotics was used for testing anti-microbial activity of SCG. S. aureus major strain of skin infection causing a range of illnesses, from minor skin infections, such as pimples, impetigo, boils, cellulitis, folliculitis, carbuncles, and abscesses, to life-threatening diseases such as toxic shock syndrome, bacteremia, and sepsis. In cosmetics, acne is the most common skin issue in the United States, affecting up to 50 million Americans annually (Bickers et al.,

2006). Specially, S. aureus is known as one of the major bacteria causing acne in human skin. Since the acne-causing bacteria are usually gram positive (Daud et al., 2013), it is reasonable to believe that the SCG could be also used to treat acne pathogenesis and other skin conditions associated with similar infection by gram positive bacterial, such as

P. acnes, S. epidermidis, S. aureus, Klebsiella pneumoniae, Streptococcus, and

Enterobacter.

Similar to the findings reported by Monente (2015), SCG samples were shown antimicrobial activity against Gram-positive bacteria using the inhibition zones of the agar-well diffusion method (Figure 2.10). Another studies evaluate that spent coffee has more strong antibacterial activity against S. aureus than coffee brew in Arabica and

Robusta species (Monente et al., 2015) indicating SCG as a potential by-product with high value. Recent studies evaluated that caffeine, trigonelline and protocatechiuic acids can act as natural antimicrobial compounds (Almeida et al., 2006). Anti-fungal effect of caffeine is related to inhibition of aflatoxin production (Buchanan and Fletcher, 1978) and its antimicrobial effect is from DNA repair mechanism inhibition . Aromatic and phenolic compounds exert antimicrobial activity by altering the structure and function of the cytoplasmic membrane, inhibiting electron transport (Buchanan and Fletcher, 1978).

In addition, bioactivity can be influenced by the heat treatment during roasting process forming coffee melanoidins which have antibacterial activity. It is known that antimicrobial activity of melanoidins could be mediated by metal chelating mechanisms

(Rufian-Henares and de la Cueva, 2009). This process called Maillard reaction are responsible for antibacterial activity of roasted coffee which shows better antimicrobial activity than raw coffee extracts. One study on “antimicrobial activity of Psidium guajava and Juglans regia leaf extracts to acne-developing organisms” presented the in vitro inhibitory effect of the leaf extracts on acne lesions. They figured out that the acne legion constitutes Propionibacterium acnes, Staphylococcus aureus, and Staphylococcus epidermidis account for 47%, 13%, and 24% respectively. The zones of inhibition due to

15% of Juglans regia leaf extracts showed 10mm against S. aures, (Qa'dan et al., 2005) similar result as Hawaiian Kona SCG (maximum 12mm). Further studies such as high

resolution metabolomics are need to identify the compounds in fraction 7 and 12 obtained from flash chromatography. Besides caffeine, trigonelline and protocatechiuic, other compounds such as quercetin glycosides, guaiacol, and cinnamic acid can be candidates of antimicrobial compounds in fraction 7 and 12. Selected compounds which can be found in coffee for antimicrobial activity were based on prior study by (Aga et al., 1994) and (Arima and Danno, 2002). This translational research projects could help turn the low-value bioactive SCG into bioactive, and organic active ingredients in personal care product, as an excellent alternative, to the synthetic preservatives. As a result, it provides safe and long-lasting “green “cosmetics product for customers for skin conditioning or treatment, such as inhibition of the acne-causing bacteria or acne prevention.

Due to the strong inhibitory effects of the SCG extracts against gram positive bacteria, the active ingredients cane be formulated into other personal products, such as toothpaste, lotions, spray, sanitizers, shampoo, wound healing patches.

Other skin-care application

5.3. Importance of anti-inflammatory and anti-carcinogenesis activity in cosmetics

In table 2.2, positive modes, umbelliferone was found in SCG which have also anti- inflammatory effect (Rauf et al., 2014). It is a potential candidate compound that can be found in purified fraction 14 (Table 2.2). Skin is always subject to inflammation, and the immediate effects on skin is redness, cracked skin, dry patches, spots (post-inflammatory hyperpigmentation) and aging. The anti-inflammatory properties can be related to inhibiting of the cyclooxygenase and lipoxygenase activity and pro-inflammatory cytokines inhibition. The skin irritation is related to releasing of cyclooxygenase and

prostaglandin in the skin. This phenomenon increases in trans-epidermal water loss and skin pH value. The active compounds in SCG extracts specially flavonoids and phenolic acid showing anti-inflammatory activity, can attribute to the reduction of the erythema and also help to improve the skin barrier function (decrease in trans-epidermal water loss and pH value), which is associated with the fast skin return to homeostasis after exposure on irritants. Usually, exposure to UVB radiation induces oxidative and inflammatory skin damage, thereby increasing the risk of skin carcinogenesis (Choi et al., 2014). The study by Cooper Bowden, activation of two transcription factor such as NF-κB, and AP-1 by prolonged UV exposure have been identified to be involved in the processes of cell- proliferation, cell differentiation and cell survival and therefore play important roles in tumorigenesis (Cooper and Bowden, 2007).

5.4. High resolution metabolomics analysis followed by bioassay guided purification

Even though glycosylated form of flavonoids show weaker bioactivities compared to free form, this study demonstrates potential antioxidants which are from SCG flavonoids such as coumarin derivatives. By chemical profiling through XCMS Online high resolution analysis, we revealed that strong ferric reducing power might from flavonoids including umbelliferone which is contained in fraction 14. One study revealed that umbelliferone has a remarkable antioxidants from hydrogen donating or radical-scavenging ability, deoxyribose, chelating power, reducing power and lipid peroxidation assay (Singh et al.,

2010). It is therefore, imperative to study the plant material for its potential as a scavenger of free radicals. Flavonoids are the active ingredients of many herbal medicines which can exhibit anti-inflammatory, anti-oxidant, immuno-regulatory, anti-

100

tumor and anti-radiation effects (Grover and Patni, 2013; Oh et al., 1997). These days, studies on natural compounds with bioactivity and effort to apply to cosmetic approach are being conducted widely. Apparently, the aglycon forms (the molecular form without sugar residues) of flavonoids are the biologically most active compounds in mammal metabolism. Commonly the bioactive ingredients of plants are always exist in mixture of aglycones and hydrophilic glycosides (Ryu et al., 2013). Aglycones are hydrophobic and can be permeable through skin easier due to low molecular weight whereas glycosides have disadvantages to be a cosmetic material (Ryu et al., 2013). This phenomenon also can be explained that penetration enhancers have low molecular weight (Bos and

Meinardi, 2000). In addition to bioavailability difference between flavonoid aglycones and glycosides, difference in bioactivity is important as well. Antioxidant activity of quercetin and quercetin monoglucosides shows that it is apparent that the activity of quercetin is much higher than that of quercetin glucosides due to structural effect to radical scavenging activity (Ioku et al., 1995). Since bioactive fractions evaluated by zone of inhibition assay have not been purified through HPLC, further studies are need to identify compounds which has antimicrobial activity.

6. Conclusion

Bioassay-guided purification method coupled with global metabolomics was utilized to identify possible antioxidant candidates in SCG samples. We identified several antioxidant agents such as cafestol, and coumarin derivaties in purified fractions 13.

Also, comparison analysis of FRAP antioxidant assay was conducted to evaluate reducing power between three SCG cultivars including Ethiopian Yirgacheffe, Costa

101

Rican Tarrazu and Hawaiian Kona. According to antioxidant assay, Hawaiian Kona showed the highest antioxidant activities among SCG varieties. In antibacterial test,

Ethiopian Yirgacheffe and Hawaiian Kona showed higher activities compared to Costa

Rican Tarrazu. Besides antioxidant and antibacterial properties, multiple bioactive compounds were found in SCG related to skin care application such as anti- inflammatory, anti-carcinogenesis, and anti-tyrosinase. These findings from SCG suggesting the promising opportunity of SCG as a raw materials that can be formulated to cosmetic products. Future plan for this project is conducting target analysis to quantify the bioactive compounds for structural study (NMR) to confirm our results.

Since SCG has been considered as an abundant, and sustainable resources, and distributed freely throughout the world, identifying a new application of SCG for cosmetic industry not only benefit consumer and environment, it also help foster a new industry to stabilize the rural economy. It can be achieved by maximizing supply chain of coffee production from sustainability, and by generating economic values in both coffee production and consumption countries in global scale. Through utilizing the modern global metabolomics approach to identify bioactive compounds, this project demonstrate a novel approach to quickly identify the bioactive ingredients in the SCG.

102

References

Adi, AJ and ZM Noor 2009. Waste recycling: Utilization of coffee grounds and kitchen waste in vermicomposting. Bioresource Technology 100: 1027-1030.

Aga, Hajime, Takashi Shibuya, Toshiyuki Sugimoto, Masashi Kurimoto and Shuhei Nakajima 1994. Isolation and identification of antimicrobial compounds in Brazilian propolis. Bioscience, Biotechnology, and Biochemistry 58: 945-946.

Ahmad, Iqbal and Arina Z Beg 2001. Antimicrobial and phytochemical studies on 45 Indian medicinal plants against multi-drug resistant human pathogens. Journal of ethnopharmacology 74: 113-123.

Akiyama, Hisanori, Kazuyasu Fujii, Osamu Yamasaki, Takashi Oono and Keiji Iwatsuki 2001. Antibacterial action of several tannins against Staphylococcus aureus. Journal of antimicrobial chemotherapy 48: 487-491.

Alarcon De La Lastra, Catalina and Isabel Villegas 2005. Resveratrol as an anti‐inflammatory and anti‐aging agent: Mechanisms and clinical implications. Molecular nutrition & food research 49: 405-430.

Almeida, Ana Amélia P, Adriana Farah, Daniela AM Silva, Elzíria A Nunan and M Beatriz A Glória 2006. Antibacterial activity of coffee extracts and selected coffee chemical compounds against enterobacteria. Journal of agricultural and food chemistry 54: 8738-8743.

Amaral, Leonard, Miguel Viveiros and Jette E Kristiansen 2006. " Non-Antibiotics": alternative therapy for the management of MDRTB and MRSA in economically disadvantaged countries. Current drug targets 7: 887-891.

An, Sang Mi, Hyo Jung Kim, Jung‐Eun Kim and Yong Chool Boo 2008. Flavonoids, taxifolin and luteolin attenuate cellular melanogenesis despite increasing tyrosinase protein levels. Phytotherapy Research 22: 1200-1207.

An, Sang Mi, Jae‐Sook Koh and Yong Chool Boo 2010. p‐coumaric acid not only inhibits human tyrosinase activity in vitro but also melanogenesis in cells exposed to UVB. Phytotherapy Research 24: 1175-1180.

Arima, Hidetoshi and Gen-ichi Danno 2002. Isolation of antimicrobial compounds from guava (Psidium guajava L.) and their structural elucidation. Bioscience, Biotechnology, and Biochemistry 66: 1727-1730.

103

Arung, Enos Tangke, Shoko Furuta, Hiroya Ishikawa, Irawan Wijaya Kusuma, Kuniyoshi Shimizu and Ryuichiro Kondo 2011. Anti-melanogenesis properties of quercetin-and its derivative-rich extract from Allium cepa. Food Chemistry 124: 1024-1028.

Awad, Rosalie, Asim Muhammad, Tony Durst, Vance L Trudeau and John T Arnason 2009. Bioassay‐guided fractionation of lemon balm (Melissa officinalis L.) using an in vitro measure of GABA transaminase activity. Phytotherapy Research 23: 1075-1081.

Aziz, NH, SE Farag, LA Mousa and MA Abo-Zaid 1998. Comparative antibacterial and antifungal effects of some phenolic compounds. Microbios 93: 43-54.

Badjakov, I, M Nikolova, R Gevrenova, V Kondakova, E Todorovska and A Atanassov 2008. Bioactive compounds in small fruits and their influence on human health. Biotechnology & Biotechnological Equipment 22: 581-587.

Bailly, Fabrice, Cédric Maurin, Elisabeth Teissier, Hervé Vezin and Philippe Cotelle 2004. Antioxidant properties of 3-hydroxycoumarin derivatives. Bioorganic & medicinal chemistry 12: 5611-5618.

Balouiri, Mounyr, Moulay Sadiki and Saad Koraichi Ibnsouda 2016. Methods for in vitro evaluating antimicrobial activity: A review. Journal of Pharmaceutical Analysis 6: 71-79. Banister, Nigel and Peter Cheetham 2001. Skin lightening agents. In Skin lightening agents: Google Patents.

Bartel, Jörg, Jan Krumsiek and Fabian J Theis 2013. Statistical methods for the analysis of high- throughput metabolomics data. Computational and structural biotechnology journal 4: 1-9.

Belguidoum, K, H Amira-Guebailia, Y Boulmokh and O Houache 2014. HPLC coupled to UV– vis detection for quantitative determination of phenolic compounds and caffeine in different brands of coffee in the Algerian market. Journal of the Taiwan Institute of Chemical Engineers 45: 1314-1320.

Bendre, Ratnamala S, Jamatsing D Rajput, Suresh D Bagul and PS Karandikar 2016. Outlooks on medicinal properties of eugenol and its synthetic derivatives. Nat Prod Chem Res 4: 2.

Benzie, Iris FF and YT Szeto 1999. Total antioxidant capacity of teas by the ferric reducing/antioxidant power assay. Journal of agricultural and food chemistry 47: 633-636.

Bernard, Francois-Xavier, Serge Sable, Beatrice Cameron, Jean Provost, Jean-Francois Desnottes, Joel Crouzet and Francis Blanche 1997. Glycosylated flavones as selective inhibitors of topoisomerase IV. Antimicrobial agents and chemotherapy 41: 992-998.

104

Bickers, David R, Henry W Lim, David Margolis, Martin A Weinstock, Clifford Goodman, Eric Faulkner, Ciara Gould, Eric Gemmen and Tim Dall 2006. The burden of skin diseases: 2004: A joint project of the American Academy of Dermatology Association and the Society for Investigative Dermatology. Journal of the American Academy of Dermatology 55: 490-500.

Bin Hafeez, Bilal, Mohammad Asim, Imtiaz Ahmad Siddiqui, Vaqar Mustafa Adhami, Imtiyaz Murtaza and Hasan Mukhtar 2008. Delphinidin, a dietary anthocyanidin in pigmented fruits and vegetables: a new weapon to blunt prostate cancer growth. Cell cycle 7: 3320-3326.

Blumberg, Simone, Oliver Frank and Thomas Hofmann 2010. Quantitative studies on the influence of the bean roasting parameters and hot water percolation on the concentrations of bitter compounds in coffee brew. Journal of agricultural and food chemistry 58: 3720-3728.

Bognar, Eszter, Zsolt Sarszegi, Aliz Szabo, Balazs Debreceni, Nikoletta Kalman, Zsuzsanna Tucsek, Balazs Sumegi and Ferenc Gallyas Jr 2013. Antioxidant and anti-inflammatory effects in RAW264. 7 macrophages of malvidin, a major red wine polyphenol. PLoS One 8: e65355.

Bonina, Francesco, Maria Lanza, Lucia Montenegro, Claudio Puglisi, Antonio Tomaino, Domenico Trombetta, Francesco Castelli and Antonella Saija 1996. Flavonoids as potential protective agents against photo-oxidative skin damage. International journal of pharmaceutics 145: 87-94.

Boots, Agnes W, Guido RMM Haenen and Aalt Bast 2008. Health effects of quercetin: from antioxidant to nutraceutical. European journal of pharmacology 585: 325-337.

Bos, Jan D and Marcus MHM Meinardi 2000. The 500 Dalton rule for the skin penetration of chemical compounds and drugs. Experimental dermatology 9: 165-169.

Bravo, Jimena, Isabel Juániz, Carmen Monente, Bettina Caemmerer, Lothar W Kroh, M Paz De Peña and Concepción Cid 2012. Evaluation of spent coffee obtained from the most common coffeemakers as a source of hydrophilic bioactive compounds. Journal of agricultural and food chemistry 60: 12565-12573.

Buchanan, RL and AM Fletcher 1978. Methylxanthine inhibition of aflatoxin production. Journal of Food Science 43: 654-655.

Burda, Stanislaw and Wieslaw Oleszek 2001. Antioxidant and antiradical activities of flavonoids. Journal of agricultural and food chemistry 49: 2774-2779.

Campos-Vega, Rocio, Guadalupe Loarca-Pina, Haydé A Vergara-Castañeda and B Dave Oomah 2015. Spent coffee grounds: A review on current research and future prospects. Trends in Food Science & Technology 45: 24-36.

105

Castro, Myrella Lessio, Walquíria Ribeiro Vilela, Rogéria Cristina Zauli, Masaharu Ikegaki, Vera Lúcia Garcia Rehder, Mary Ann Foglio, Severino Matias de Alencar and Pedro Luiz Rosalen 2009. Bioassay guided purification of the antimicrobial fraction of a Brazilian propolis from Bahia state. BMC complementary and alternative medicine 9: 25.

Cavia‐Saiz, Monica, Maria D Busto, Maria Concepción Pilar‐Izquierdo, Natividad Ortega, Manuel Perez‐Mateos and Pilar Muñiz 2010. Antioxidant properties, radical scavenging activity and biomolecule protection capacity of flavonoid naringenin and its glycoside naringin: a comparative study. Journal of the Science of Food and Agriculture 90: 1238-1244.

Chamcheu, Jean Christopher, Harish C Pal, Imtiaz A Siddiqui, Vaqar M Adhami, Seyoum Ayehunie, Brendan T Boylan, Felicite K Noubissi, Naghma Khan, Deeba N Syed and Craig A Elmets 2015. Prodifferentiation, anti-inflammatory and antiproliferative effects of delphinidin, a dietary anthocyanidin, in a full-thickness three-dimensional reconstituted human skin model of psoriasis. Skin pharmacology and physiology 28: 177-188.

CHEN, CHUNG‐WEN and CHI‐TANG HO 1995. Antioxidant properties of polyphenols extracted from green and black teas. Journal of food lipids 2: 35-46.

Chen, Jing, Kao-Wei Chua, Chu C Chua, Hailong Yu, Aijie Pei, Balvin HL Chua, Ronald C Hamdy, Xingshun Xu and Chun-Feng Liu 2011. Antioxidant activity of 7, 8-dihydroxyflavone provides neuroprotection against glutamate-induced toxicity. Neuroscience letters 499: 181-185.

Choi, Hyeon-Son, Eu Ddeum Park, Yooheon Park and Hyung Joo Suh 2015. Spent coffee ground extract suppresses ultraviolet B-induced photoaging in hairless mice. Journal of Photochemistry and Photobiology B: Biology 153: 164-172.

Choi, Ki-Seok, Joydeb Kumar Kundu, Kyung-Soo Chun, Hye-Kyung Na and Young-Joon Surh 2014. Rutin inhibits UVB radiation-induced expression of COX-2 and iNOS in hairless mouse skin: p38 MAP kinase and JNK as potential targets. Archives of biochemistry and biophysics 559: 38-45.

Choi, Mi Jin, Boh Kyung Kim, Kun Young Park, Takako Yokozawa, Young Ok Song and Eun Ju Cho 2010. Anti-aging effects of cyanidin under a stress-induced premature senescence cellular system. Biological and pharmaceutical bulletin 33: 421-426.

Choquenet, Benjamin, Céline Couteau, Eva Paparis and Laurence JM Coiffard 2008. Quercetin and rutin as potential sunscreen agents: determination of efficacy by an in vitro method. Journal of natural products 71: 1117-1118.

Chou, Tzung‐Han, Hsiou‐Yu Ding, Wei Jing Hung and Chia‐Hua Liang 2010. Antioxidative characteristics and inhibition of α‐melanocyte‐stimulating hormone‐stimulated melanogenesis of vanillin and vanillic acid from Origanum vulgare. Experimental dermatology 19: 742-750.

106

Christin, Christin, Age K Smilde, Huub CJ Hoefsloot, Frank Suits, Rainer Bischoff and Peter L Horvatovich 2008. Optimized time alignment algorithm for LC− MS data: correlation optimized warping using component detection algorithm-selected mass chromatograms. Analytical chemistry 80: 7012-7021.

Cingolani, Gian, Fulvio Gualtieri and Maria Pigini 1969. Researches in the field of antiviral compounds. Mannich bases of 3-hydroxycoumarin. Journal of medicinal chemistry 12: 531-532.

Cloughley, John B 1980. The effect of temperature on enzyme activity during the fermentation phase of black tea manufacture. Journal of the Science of Food and Agriculture 31: 920-923.

Coble, Jamie B and Carlos G Fraga 2014. Comparative evaluation of preprocessing freeware on chromatography/mass spectrometry data for signature discovery. Journal of chromatography A 1358: 155-164.

Cooper, SJ and GT Bowden 2007. Ultraviolet B regulation of transcription factor families: Roles of nuclear factor-kappa B (NF-κB) and activator protein-1 (AP-1) in UVB-induced skin carcinogenesis. Current cancer drug targets 7: 325-334.

Cruz, Rebeca, Paula Baptista, Sara Cunha, José Alberto Pereira and Susana Casal 2012. Carotenoids of lettuce (Lactuca sativa L.) grown on soil enriched with spent coffee grounds. Molecules 17: 1535-1547.

Daglia, Maria, Maria Teresa Cuzzoni and Cesare Dacarro 1994a. Antibacterial activity of coffee. Journal of Agricultural and Food Chemistry 42: 2270-2272.

Daglia, Maria, Maria Teresa Cuzzoni and Cesare Dacarro 1994b. Antibacterial activity of coffee: relationship between biological activity and chemical markers. Journal of Agricultural and Food Chemistry 42: 2273-2277.

Daud, Farhat S, Gauri Pande, Mamta Joshi, Ruchita Pathak and Shubhangi Wankhede 2013. A study of antibacterial effect of some selected essential oils and medicinal herbs against acne causing bacteria. J. Pharm. Sci. Invent 2: 27-34. de Araujo Leite, Jocelmo Cássio, Tainá Myra Xavier de Castro, José Maria Barbosa-Filho, José Pinto de Siqueira-Junior and Luis Fernando Marques-Santos 2015. Photoprotective effect of coumarin and 3-hydroxycoumarin in sea urchin gametes and embryonic cells. Journal of Photochemistry and Photobiology B: Biology 146: 44-51.

Devasagayam, TPA, JP Kamat, Hari Mohan and PC Kesavan 1996. Caffeine as an antioxidant: inhibition of lipid peroxidation induced by reactive oxygen species. Biochimica et Biophysica Acta (BBA)-Biomembranes 1282: 63-70.

107

Do, Youn-Kyung, Jong-Min Kim, Sang-Mok Chang, Ji-Hwan Hwang and Woo-Sik Kim 2009. Enhancement of polyphenol bio-activities by enzyme reaction. Journal of Molecular Catalysis B: Enzymatic 56: 173-178.

El Gharras, Hasna 2009. Polyphenols: food sources, properties and applications–a review. International journal of food science & technology 44: 2512-2518.

Esquivel, Patricia and Víctor M Jiménez 2012. Functional properties of coffee and coffee by- products. Food Research International 46: 488-495.

Fardiaz, Srikandi 1995. Antimicrobial activity of coffee (Coffea robusta) extract. ASEAN Food Journal 10: 103-106.

Ferrali, M, C Signorini, B Caciotti, L Sugherini, L Ciccoli, D Giachetti and M Comporti 1997. Protection against oxidative damage of erythrocyte membrane by the flavonoid quercetin and its relation to iron chelating activity. FEBS letters 416: 123-129.

Firuzi, Omidreza, Antonio Lacanna, Rita Petrucci, Giancarlo Marrosu and Luciano Saso 2005. Evaluation of the antioxidant activity of flavonoids by “ferric reducing antioxidant power” assay and cyclic voltammetry. Biochimica et Biophysica Acta (BBA)-General Subjects 1721: 174-184.

Fukui, Yuko and Takashi Iwashita 2013. Epigallocatechin gallate trimer and α-glucosidase inhibitor containing epigallocatechin gallate polymer. In Epigallocatechin gallate trimer and α- glucosidase inhibitor containing epigallocatechin gallate polymer: Google Patents.

Gani, Adil, SM Wani, FA Masoodi and G Hameed 2012. Whole-grain cereal bioactive compounds and their health benefits: a review. J Food Process Technol 3: 146-156. Gao, Hong, Jun Nishida, Shizuka Saito and Jun Kawabata 2007. Inhibitory effects of 5, 6, 7- trihydroxyflavones on tyrosinase. Molecules 12: 86-97.

García-Bores, AM and J GUILLERMO Avila 2008. Natural products: Molecular mechanisms in the photochemoprevention of skin cancer. Rev Latinoamer Quím 36: 83-102. García-Lafuente, Ana, Eva Guillamón, Ana Villares, Mauricio A Rostagno and José Alfredo Martínez 2009. Flavonoids as anti-inflammatory agents: implications in cancer and cardiovascular disease. Inflammation Research 58: 537-552.

García-Reyes, Juan F, M Dolores Hernando, Antonio Molina-Díaz and Amadeo R Fernández- Alba 2007. Comprehensive screening of target, non-target and unknown pesticides in food by LC-TOF-MS. TrAC Trends in Analytical Chemistry 26: 828-841.

García-Sosa, Karlina, Ninibe Villarreal-Alvarez, Petra Lübben and Luis M Peña-Rodríguez 2006. Chrysophanol, an antimicrobial anthraquinone from the root extract of Colubrina greggii. Journal of the Mexican Chemical Society 50: 76-78.

108

Gowda, Harsha, Julijana Ivanisevic, Caroline H Johnson, Michael E Kurczy, H Paul Benton, Duane Rinehart, Thomas Nguyen, Jayashree Ray, Jennifer Kuehl and Bernardo Arevalo 2014. Interactive XCMS Online: simplifying advanced metabolomic data processing and subsequent statistical analyses. Analytical chemistry 86: 6931.

Graf, J 2000. Herbal anti-inflammatory agents for skin disease. Skin Therapy Lett 5: 3-5.

Grover, Neha and Vidya Patni 2013. Phytochemical characterization using various solvent extracts and GC-MS analysis of methanolic extract of Woodfordia fruticosa (L) Kurz. Leaves. International Journal of Pharmacy and Pharmaceutical Sciences 5: 291-295.

Guyot, Bernard, Dominique Gueule, Jean-Claude Manez, Jean-Jacques Perriot, J Giron and Luc Villain 1996. Influence de l'altitude et de l'ombrage sur la qualité des cafés Arabica. Plantations, recherche, développement 3: 272-283.

Hämäläinen, Mari, Riina Nieminen, Pia Vuorela, Marina Heinonen and Eeva Moilanen 2007. Anti-inflammatory effects of flavonoids: genistein, kaempferol, quercetin, and daidzein inhibit STAT-1 and NF-κB activations, whereas flavone, isorhamnetin, naringenin, and pelargonidin inhibit only NF-κB activation along with their inhibitory effect on iNOS expression and NO production in activated macrophages. Mediators of inflammation 2007.

Han, Byung Hoon, Myung Hwan Park and Yong Nam Han 1981. Studies on the antioxidant components of Korean ginseng (III). Archives of pharmacal research 4: 53-58.

Hanasaki, Yukiko, Shunjiro Ogawa and Shozo Fukui 1994. The correlation between active oxygens scavenging and antioxidative effects of flavonoids. Free Radical Biology and Medicine 16: 845-850.

Haraguchi, Hiroyuki, Katsuhito Tanimoto, Yukiyoshi Tamura, Kenji Mizutani and Takeshi Kinoshita 1998. Mode of antibacterial action of retrochalcones from Glycyrrhiza inflata. Phytochemistry 48: 125-129.

Heuberger, Adam L, Matthew R Lewis, Ming-Hsuan Chen, Mark A Brick, Jan E Leach and Elizabeth P Ryan 2010. Metabolomic and functional genomic analyses reveal varietal differences in bioactive compounds of cooked rice. PLoS One 5: e12915.

Hodges, D Mark, John M DeLong, Charles F Forney and Robert K Prange 1999. Improving the thiobarbituric acid-reactive-substances assay for estimating lipid peroxidation in plant tissues containing anthocyanin and other interfering compounds. Planta 207: 604-611.

Huang, Mou-Tuan, Jian-Guo Xie, Zhi Yuan Wang, Chi-Tang Ho, You-Rong Lou, Chung-Xiou Wang, Gordon C Hard and Allan H Conney 1997. Effects of tea, decaffeinated tea, and caffeine on uvb light-induced complete carcinogenesis in SKH-1 mice: demonstration of caffeine as a biologically important constitutent of tea. Cancer Research 57: 2623-2629.

109

Huang, Wu-Yang, Jian Wang, Ya-Mei Liu, Qi-Sheng Zheng and Chun-Yang Li 2014. Inhibitory effect of Malvidin on TNF-α-induced inflammatory response in endothelial cells. European journal of pharmacology 723: 67-72.

Ikigai, Hajime, Taiji Nakae, Yukihiko Hara and Tadakatsu Shimamura 1993. Bactericidal catechins damage the lipid bilayer. Biochimica et Biophysica Acta (BBA)-Biomembranes 1147: 132-136.

Ioku, Kana, Tojiro Tsushida, Yoko Takei, Nobuji Nakatani and Junji Terao 1995. Antioxidative activity of quercetin and quercetin monoglucosides in solution and phospholipid bilayers. Biochimica et Biophysica Acta (BBA)-Biomembranes 1234: 99-104.

Iwai, Kazuya, Noriaki Kishimoto, Yukari Kakino, Kyo Mochida and Tokio Fujita 2004. In vitro antioxidative effects and tyrosinase inhibitory activities of seven hydroxycinnamoyl derivatives in green coffee beans. Journal of agricultural and food chemistry 52: 4893-4898.

Jin, Seong Eun, You Kyung Son, Byung-Sun Min, Hyun Ah Jung and Jae Sue Choi 2012. Anti- inflammatory and antioxidant activities of constituents isolated from Pueraria lobata roots. Archives of pharmacal research 35: 823-837.

Joët, Thierry, Andréina Laffargue, Frédéric Descroix, Sylvie Doulbeau, Benoît Bertrand and Stéphane Dussert 2010. Influence of environmental factors, wet processing and their interactions on the biochemical composition of green Arabica coffee beans. Food Chemistry 118: 693-701.

Ju, Huangxian 2013. Grand challenges in analytical chemistry: towards more bright eyes for scientific research, social events and human health. Frontiers in chemistry 1.

Jung, Sung Keun, Ki Won Lee, Ho Young Kim, Mi Hyun Oh, Sanguine Byun, Sung Hwan Lim, Yong-Seok Heo, Nam Joo Kang, Ann M Bode and Zigang Dong 2010. Myricetin suppresses UVB-induced wrinkle formation and MMP-9 expression by inhibiting Raf. Biochemical pharmacology 79: 1455-1461.

Kanlayavattanakul, Mayuree and Nattaya Lourith 2013. Spent Arabica Coffee as a Rich Source of Antioxidant Appraisal for Cosmetic Applications. Advanced Science, Engineering and Medicine 5: 173-176.

Katalinic, V, M Milos, T Kulisic and M Jukic 2006. Screening of 70 medicinal plant extracts for antioxidant capacity and total phenols. Food Chemistry 94: 550-557.

Kato, Ryuichi, Teruo Nakadate, Satoshi Yamamoto and Takashi Sugimura 1983. Inhibition of 12-O-tetradecanoylphorbol-13-acetate-induced tumor promotion and ornithine decarboxylase activity by quercetin: possible involvement of lipoxygenase inhibition. Carcinogenesis 4: 1301- 1305.

110

Kaufmann, A 2014. Combining UHPLC and high-resolution MS: a viable approach for the analysis of complex samples? TrAC Trends in Analytical Chemistry 63: 113-128.

Kenner, Dan and Yves Requena 2001. Botanical medicine: a European professional perspective: Paradigm publications.

Khaiat, Alain and Claude Saliou 2015. Botanical Extracts. Cosmeceuticals and Active Cosmetics: 385.

Kiattisin, Kanokwan, Thananya Nantarat and Pimporn Leelapornpisid 2016. Evaluation of antioxidant and anti-tyrosinase activities as well as stability of green and roasted coffee bean extracts from Coffea arabica and Coffea canephora grown in Thailand. Journal of Pharmacognosy and Phytotherapy 8: 182-192.

Kilmartin, Paul A and Chyong F Hsu 2003. Characterisation of polyphenols in green, oolong, and black teas, and in coffee, using cyclic voltammetry. Food Chemistry 82: 501-512.

Kim, Seong-Jin, Mi-Hye Lim, Inn-Ki Chun and Young-Ho Won 1997. Effects of flavonoids of Ginkgo biloba on proliferation of human skin fibroblast. Skin pharmacology and physiology 10: 200-205.

Kim, You-Jung 2007. Antimelanogenic and antioxidant properties of gallic acid. Biological and pharmaceutical bulletin 30: 1052-1055.

Kim, Young Mee, Jung Sik Huh, Yoongho Lim and Moonjae Cho 2015. Soy Isoflavone Glycitin (4'‐Hydroxy‐6‐Methoxyisoflavone‐7‐D‐Glucoside) Promotes Human Dermal Fibroblast Cell Proliferation and Migration via TGF‐β Signaling. Phytotherapy Research 29: 757-769.

Kitagawa, Shuji, Kenta Yoshii, Shin-ya Morita and Reiko Teraoka 2011. Efficient topical delivery of chlorogenic acid by an oil-in-water microemulsion to protect skin against UV- induced damage. Chemical and Pharmaceutical Bulletin 59: 793-796.

Kong, Yeon Hee, Youn Ock Jo, Chang-Won Cho, Dongwook Son, Soojin Park, Jeonghae Rho and Sang Yoon Choi 2008. Inhibitory effects of cinnamic acid on melanin biosynthesis in skin. Biological and pharmaceutical bulletin 31: 946-948.

Korgenski, E Kent and Judy A Daly 1998. Evaluation of the BIOMIC video reader system for determining interpretive categories of isolates on the basis of disk diffusion susceptibility results. Journal of clinical microbiology 36: 302-304.

Kris-Etherton, PM, M Lefevre, GR Beecher, MD Gross, CL Keen and TD Etherton 2004. Bioactive compounds in nutrition and health-research methodologies for establishing biological function: the antioxidant and anti-inflammatory effects of flavonoids on atherosclerosis. Annu. Rev. Nutr. 24: 511-538.

111

Kuhl, Carsten, Ralf Tautenhahn, Christoph Bottcher, Tony R Larson and Steffen Neumann 2011. CAMERA: an integrated strategy for compound spectra extraction and annotation of liquid chromatography/mass spectrometry data sets. Analytical chemistry 84: 283-289.

Kusumawati, Idha and Gunawan Indrayanto 2013. Natural antioxidants in cosmetics. Stud. Nat. Prod. Chem 40: 485-505.

Lanni, C and EL Becker 1985. Inhibition of neutrophil phospholipase A2 by p-bromophenylacyl bromide, nordihydroguaiaretic acid, 5, 8, 11, 14-eicosatetrayenoic acid and quercetin. International Archives of Allergy and Immunology 76: 214-217.

Lee-Hilz, Yee Y, Anne-Marie JF Boerboom, Adrie H Westphal, Willem JH van Berkel, Jac MMJG Aarts and Ivonne MCM Rietjens 2006. Pro-oxidant activity of flavonoids induces EpRE- mediated gene expression. Chemical research in toxicology 19: 1499-1505.

Lee, Jin Young, Yeon Kyung Kim, Jin Young Seo, Chong Won Choi, Jae Sung Hwang, Byeong Gon Lee, Ih Seop Chang and Jin Ho Chung 2008. Loss of elastic fibers causes skin wrinkles in sun-damaged human skin. Journal of dermatological science 50: 99-107.

Lee, Kuo-Hsiung, Hideji Itokawa and Mutsuo Kozuka 2003a. Oriental herbal products: the basis for development of dietary supplements and new medicines in the 21st century. ACS Publications.

Lee, Kyung Jin and Hye Gwang Jeong 2007. Protective effects of kahweol and cafestol against hydrogen peroxide-induced oxidative stress and DNA damage. Toxicology letters 173: 80-87.

Lee, Mei-Hsien, Yi-Pei Lin, Feng-Lin Hsu, Gui-Rong Zhan and Kun-Ying Yen 2006. Bioactive constituents of Spatholobus suberectus in regulating tyrosinase-related proteins and mRNA in HEMn cells. Phytochemistry 67: 1262-1270.

Lee, Sanghyun, Kyoung Soon Kim, Youmie Park, Kuk Hyun Shin and Bak-Kwang Kim 2003b. In vivo anti-oxidant activities of tectochrysin. Archives of pharmacal research 26: 43-46.

Lee, Taek Hwan, Lalita Subedi, Hussain Mustatab Wahedi, Yong Un Park and Sun Yeou Kim 2016. Resveratrol-enriched rice protects human skin against UVB-induced photoaging. The FASEB Journal 30: lb557-lb557.

Li, Xican, Xiaozhen Wang, Dongfeng Chen and Shuzhi Chen 2011. Antioxidant activity and mechanism of protocatechuic acid in vitro. Functional Foods in Health and Disease 1: 232-244.

Libiseller, Gunnar, Michaela Dvorzak, Ulrike Kleb, Edgar Gander, Tobias Eisenberg, Frank Madeo, Steffen Neumann, Gert Trausinger, Frank Sinner and Thomas Pieber 2015. IPO: a tool for automated optimization of XCMS parameters. BMC bioinformatics 16: 118.

112

Lin, Chih-Chien, Chao-Hsun Yang, Pey-Shiuan Wu, Chang-Chin Kwan and Yi-Shyan Chen 2011. Antimicrobial, anti-tyrosinase and antioxidant activities of aqueous aromatic extracts from forty-eight selected herbs. Journal of Medicinal Plants Research 5: 6203-6209.

Lino, CS, ML Taveira, GSB Viana and FJA Matos 1997. Analgesic and antiinflammatory activities of Justicia pectoralis Jacq and its main constituents: coumarin and umbelliferone. Phytotherapy Research 11: 211-215.

Lu, Yao-Ping, You-Rong Lou, Jian-Guo Xie, Qing-Yun Peng, Jie Liao, Chung S Yang, Mou- Tuan Huang and Allan H Conney 2002. Topical applications of caffeine or (−)-epigallocatechin gallate (EGCG) inhibit carcinogenesis and selectively increase apoptosis in UVB-induced skin tumors in mice. Proceedings of the National Academy of Sciences 99: 12455-12460.

Ludwig, Iziar A, Lidia Sanchez, Bettina Caemmerer, Lothar W Kroh, M Paz De Peña and Concepción Cid 2012. Extraction of coffee antioxidants: impact of brewing time and method. Food Research International 48: 57-64.

Lv, Hongming, Zhenxiang Yu, Yuwei Zheng, Lidong Wang, Xiaofeng Qin, Genhong Cheng and Xinxin Ci 2016. Isovitexin exerts anti-inflammatory and anti-oxidant activities on lipopolysaccharide-induced acute lung injury by inhibiting MAPK and NF-κB and activating HO-1/Nrf2 pathways. International journal of biological sciences 12: 72.

MacDonald‐Wicks, Lesley K, Lisa G Wood and Manohar L Garg 2006. Methodology for the determination of biological antioxidant capacity in vitro: a review. Journal of the Science of Food and Agriculture 86: 2046-2056.

Mahal, HS and T Mukherjee 2006. Scavenging of reactive oxygen radicals by resveratrol: antioxidant effect. Research on chemical intermediates 32: 59-71.

Makni, Mohamed, Yassine Chtourou, Hamadi Fetoui, El Mouldi Garoui, Tahia Boudawara and Najiba Zeghal 2011. Evaluation of the antioxidant, anti-inflammatory and hepatoprotective properties of vanillin in carbon tetrachloride-treated rats. European journal of pharmacology 668: 133-139.

Manach, Claudine, Augustin Scalbert, Christine Morand, Christian Rémésy and Liliana Jiménez 2004. Polyphenols: food sources and bioavailability. The American journal of clinical nutrition 79: 727-747.

Momtaz, S, BM Mapunya, PJ Houghton, C Edgerly, A Hussein, S Naidoo and N Lall 2008. Tyrosinase inhibition by extracts and constituents of Sideroxylon inerme L. stem bark, used in South Africa for skin lightening. Journal of ethnopharmacology 119: 507-512.

113

Monente, Carmen, Jimena Bravo, Ana Isabel Vitas, Leire Arbillaga, M Paz De Peña and Concepción Cid 2015. Coffee and spent coffee extracts protect against cell mutagens and inhibit growth of food-borne pathogen microorganisms. Journal of Functional Foods 12: 365-374. Moon, Joon-Kwan and Takayuki Shibamoto 2009. Antioxidant assays for plant and food components. Journal of agricultural and food chemistry 57: 1655-1666.

Moore, Kevin and L Jackson Roberts 1998. Measurement of lipid peroxidation. Free radical research 28: 659-671.

Mori, Akihisa, Chikao Nishino, Nobuyasu Enoki and Shinkichi Tawata 1987. Antibacterial activity and mode of action of plant flavonoids against Proteus vulgaris and Staphylococcus aureus. Phytochemistry 26: 2231-2234.

Moure, Andrés, Manuel Pazos, Isabel Medina, Herminia Domínguez and Juan Carlos Parajó 2007. Antioxidant activity of extracts produced by solvent extraction of almond shells acid hydrolysates. Food Chemistry 101: 193-201.

Mu, H, Y-H Bai, S-T Wang, Z-M Zhu and Y-W Zhang 2009. Research on antioxidant effects and estrogenic effect of formononetin from Trifolium pratense (red clover). Phytomedicine 16: 314-319.

Mussatto, Solange I, Lina F Ballesteros, Silvia Martins and José A Teixeira 2011a. Extraction of antioxidant phenolic compounds from spent coffee grounds. Separation and Purification Technology 83: 173-179.

Mussatto, Solange I, Ercília MS Machado, Silvia Martins and José A Teixeira 2011b. Production, composition, and application of coffee and its industrial residues. Food and Bioprocess Technology 4: 661.

Nichols, Joi A and Santosh K Katiyar 2010. Skin photoprotection by natural polyphenols: anti- inflammatory, antioxidant and DNA repair mechanisms. Archives of dermatological research 302: 71-83.

Nishijima, Setsuko, Ichiro Kurokawa, Naoki Katoh and Kunitomo Watanabe 2000. The bacteriology of acne vulgaris and antimicrobial susceptibility of Propionibacterium acnes and Staphylococcus epidermidis isolated from acne lesions. The Journal of dermatology 27: 318-323.

Noda, Yasuko, Takao Kaneyuki, Akitane Mori and Lester Packer 2002. Antioxidant activities of pomegranate fruit extract and its anthocyanidins: delphinidin, cyanidin, and pelargonidin. Journal of agricultural and food chemistry 50: 166-171.

Núñez-Sellés, Alberto J 2005. Antioxidant therapy: myth or reality? Journal of the brazilian Chemical society 16: 699-710.

114

Oh, Kyung-Tack, Jae-Won Ka, Hyon-Pyo Kim and Chang-Hee Lee 1997. Syntheses of Porphyrin-Isoflavonoid Conjugates and Octa-Substituted Porphyrins as a Building Blocks for 3- Dimensional Array of Porphyrins. BULLETIN-KOREAN CHEMICAL SOCIETY 18: 907-910.

Ohemeng, KA, CF Schwender, KP Fu and JF Barrett 1993. DNA gyrase inhibitory and antibacterial activity of some flavones (1). Bioorganic & Medicinal Chemistry Letters 3: 225- 230.

Ohta, Toshihiro, Kazuko Watanabe, Masaaki Moriya, Yasuhiko Shirasu and Tsuneo Kada 1983. Anti-mutagenic effects of coumarin and umbelliferone on mutagenesis induced by 4- nitroquinoline 1-oxide or UV irradiation in E. coli. Mutation Research/Genetic Toxicology 117: 135-138.

Oi, Naomi, Hanyong Chen, Myoung Ok Kim, Ronald A Lubet, Ann M Bode and Zigang Dong 2012. Taxifolin suppresses UV-induced skin carcinogenesis by targeting EGFR and PI3K. Cancer Prevention Research 5: 1103-1114.

Pandey, Kanti B and Syed Ibrahim Rizvi 2009. Current understanding of dietary polyphenols and their role in health and disease. Current Nutrition & Food Science 5: 249-263.

Park, Hye Young, Gi-Young Kim, Jin Won Hyun, Hye Jin Hwang, Nam Deuk Kim, Byung-Woo Kim and Yung Hyun Choi 2012. 7, 8-Dihydroxyflavone exhibits anti-inflammatory properties by downregulating the NF-κB and MAPK signaling pathways in lipopolysaccharide-treated RAW264. 7 cells. International journal of molecular medicine 29: 1146-1152.

Park, Kwan Hee, Min Park, Sun Eun Choi, Mi Sook Jeong, Joo Hee Kwon, Myung Hwan Oh, Hyung Kyun Choi, Seong Jun Seo and Min Won Lee 2009. The anti-oxidative and anti- inflammatory effects of caffeoyl derivatives from the roots of Aconitum koreanum R. RAYMOND. Biological and pharmaceutical bulletin 32: 2029-2033.

Perchellet, Jean-Pierre, Hala U Gali, Elisabeth M Perchellet, Darren S Klish and Andrew D Armbrust 1992. Antitumor-promoting activities of tannic acid, ellagic acid, and several gallic acid derivatives in mouse skin. In Plant Polyphenols, 783-801: Springer.

Pereira, José Alberto, Ivo Oliveira, Anabela Sousa, Patrícia Valentão, Paula B Andrade, Isabel CFR Ferreira, Federico Ferreres, Albino Bento, Rosa Seabra and Letícia Estevinho 2007. Walnut (Juglans regia L.) leaves: phenolic compounds, antibacterial activity and antioxidant potential of different cultivars. Food and Chemical Toxicology 45: 2287-2295.

Pero, Ronald W, Harald Lund and Tomas Leanderson 2009. Antioxidant metabolism induced by quinic acid. Increased urinary excretion of tryptophan and nicotinamide. Phytotherapy Research 23: 335-346.

115

Petrovic, Mira, Meritxell Gros and Damia Barcelo 2006. Multi-residue analysis of pharmaceuticals in wastewater by ultra-performance liquid chromatography–quadrupole–time- of-flight mass spectrometry. Journal of chromatography A 1124: 68-81.

Potapovich, Alla I, Vladimir A Kostyuk, Tatyana V Kostyuk, Chiara de Luca and Liudmila G Korkina 2013. Effects of pre-and post-treatment with plant polyphenols on human keratinocyte responses to solar UV. Inflammation Research 62: 773-780.

Procházková, Dagmar, I Boušová and N Wilhelmová 2011. Antioxidant and prooxidant properties of flavonoids. Fitoterapia 82: 513-523.

Pulido, Raquel, Laura Bravo and Fulgencio Saura-Calixto 2000. Antioxidant activity of dietary polyphenols as determined by a modified ferric reducing/antioxidant power assay. Journal of Agricultural and Food Chemistry 48: 3396-3402.

Qa'dan, Fadi, Abdul-Jalil Thewaini, Dalia A Ali, Rana Afifi, Abdalla Elkhawad and Khalid Z Matalka 2005. The antimicrobial activities of Psidium guajava and Juglans regia leaf extracts to acne-developing organisms. The American Journal of Chinese Medicine 33: 197-204.

Ramalakshmi, K, L Jagan Mohan Rao, Yuko Takano-Ishikawa and Masao Goto 2009. Bioactivities of low-grade green coffee and spent coffee in different in vitro model systems. Food Chemistry 115: 79-85.

Raman, A, U Weir and SF Bloomfield 1995. Antimicrobial effects of tea‐tree oil and its major components on Staphylococcus aureus, Staph. epidermidis and Propionibacterium acnes. Letters in Applied Microbiology 21: 242-245.

Ramani, Rama and Vishnu Chaturvedi 2000. Flow cytometry antifungal susceptibility testing of pathogenic yeasts other than Candida albicans and comparison with the NCCLS broth microdilution test. Antimicrobial agents and chemotherapy 44: 2752-2758.

Ramesh, B and KV Pugalendi 2006. Antioxidant role of umbelliferone in STZ-diabetic rats. Life sciences 79: 306-310.

Ramos, Freddy A, Yoshihisa Takaishi, Miki Shirotori, Yousuke Kawaguchi, Koichiro Tsuchiya, Hirofumi Shibata, Tomihiko Higuti, Tetsuo Tadokoro and Minoru Takeuchi 2006. Antibacterial and antioxidant activities of quercetin oxidation products from yellow onion (Allium cepa) skin. Journal of agricultural and food chemistry 54: 3551-3557.

Rauf, Abdur, Rehan Khan, Haroon Khan, Samreen Pervez and Abdul Saboor Pirzada 2014. In vivo antinociceptive and anti-inflammatory activities of umbelliferone isolated from Potentilla evestita. Natural product research 28: 1371-1374.

116

Razzaghi-Abyaneh, Mehdi and Mahendra Rai 2013. Antifungal metabolites from plants: Springer.

Re, Roberta, Nicoletta Pellegrini, Anna Proteggente, Ananth Pannala, Min Yang and Catherine Rice-Evans 1999. Antioxidant activity applying an improved ABTS radical cation decolorization assay. Free Radical Biology and Medicine 26: 1231-1237.

Ren, Xiaolin, Yuling Shi, Di Zhao, Mengyu Xu, Xiaolong Li, Yongyan Dang and Xiyun Ye 2016. Naringin protects ultraviolet B-induced skin damage by regulating p38 MAPK signal pathway. Journal of dermatological science 82: 106-114.

Ribeiro, Helena, Joana Marto, Sara Raposo, Maria Agapito, Vera Isaac, Bruna G Chiari, Pedro F Lisboa, Alexandre Paiva, Susana Barreiros and Pedro Simões 2013. From coffee industry waste materials to skin‐friendly products with improved skin fat levels. European journal of lipid science and technology 115: 330-336.

Rice-Evans, Catherine A, Nicholas J Miller and George Paganga 1996. Structure-antioxidant activity relationships of flavonoids and phenolic acids. Free Radical Biology and Medicine 20: 933-956.

Ruff, Matthias, Miriam S Mueller, Martin Loos and Heinz P Singer 2015. Quantitative target and systematic non-target analysis of polar organic micro-pollutants along the river Rhine using high-resolution mass-spectrometry–identification of unknown sources and compounds. Water research 87: 145-154.

Rufian-Henares, Jose A and Silvia P de la Cueva 2009. Antimicrobial Activity of Coffee Melanoidins A Study of Their Metal-Chelating Properties. Journal of agricultural and food chemistry 57: 432-438.

Runyoro, Deborah KB, Mecky IN Matee, Olipa D Ngassapa, Cosam C Joseph and Zakaria H Mbwambo 2006. Screening of Tanzanian medicinal plants for anti-Candida activity. BMC complementary and alternative medicine 6: 11.

Russo, Alessandra, Venera Cardile, Laura Lombardo, Luca Vanella and Rosaria Acquaviva 2006. Genistin inhibits UV light-induced plasmid DNA damage and cell growth in human melanoma cells. The Journal of nutritional biochemistry 17: 103-108.

Ryu, Jae Sik, Hyun Jung Lee, Song Hwan Bae, Sun Young Kim, Yooheon Park, Hyung Joo Suh and Yoon Hwa Jeong 2013. The bioavailability of red ginseng extract fermented by Phellinus linteus. Journal of ginseng research 37: 108.

117

Saavedra, Maria J, Anabela Borges, Carla Dias, Alfredo Aires, Richard N Bennett, Eduardo S Rosa and Manuel Simões 2010. Antimicrobial activity of phenolics and glucosinolate hydrolysis products and their synergy with streptomycin against pathogenic bacteria. Medicinal Chemistry 6: 174-183.

Saghaie, Lotfollah, Morteza Pourfarzam, A Fassihi and B Sartippour 2013. Synthesis and tyrosinase inhibitory properties of some novel derivatives of kojic acid. Research in pharmaceutical sciences 8: 233.

Saija, Antonella, Mario Scalese, Maria Lanza, Daniela Marzullo, Francesco Bonina and Francesco Castelli 1995. Flavonoids as antioxidant agents: importance of their interaction with biomembranes. Free Radical Biology and Medicine 19: 481-486.

Sakano, Katsuhisa, Mika Mizutani, Mariko Murata, Shinji Oikawa, Yusuke Hiraku and Shosuke Kawanishi 2005. Procyanidin B2 has anti-and pro-oxidant effects on metal-mediated DNA damage. Free Radical Biology and Medicine 39: 1041-1049.

Sanchez-Moreno, C 2002. Methods used to evaluate the free radical scavenging activity in foods and biological systems. Revista de Agaroquimica y Tecnologia de Alimentos 8: 121-137.

Sansanelli, Sabrina, Dario Zanichelli, Alessandro Filippini, Maura Ferri and Annalisa Tassoni 2014. Production of free and glycosylated isoflavones in in vitro soybean (Glycine max L.) hypocotyl cell suspensions and comparison with industrial seed extracts. Plant Cell, Tissue and Organ Culture (PCTOC) 119: 301-311.

Saraf, S and CD Kaur 2010. Phytoconstituents as photoprotective novel cosmetic formulations. Pharmacognosy reviews 4: 1.

Sato, Yuki, Shirou Itagaki, Toshimitsu Kurokawa, Jiro Ogura, Masaki Kobayashi, Takeshi Hirano, Mitsuru Sugawara and Ken Iseki 2011. In vitro and in vivo antioxidant properties of chlorogenic acid and caffeic acid. International journal of pharmaceutics 403: 136-138. Schmid, Daniel, F Zulli, Hans-Peter Nissen and Hanspeter Prieur 2003. Penetration and metabolism of isoflavones in human skin. Cosmetics and toiletries 118: 71-76.

Schmid, Daniel and Fred ZÜLLI 2002. Topically applied soy isoflavones increase skin thickness. Cosmetics and toiletries 117: 45-50.

Seeram, N, Rupo Lee, M Hardy and D Heber 2005. Rapid large scale purification of ellagitannins from pomegranate husk, a by-product of the commercial juice industry. Separation and Purification Technology 41: 49-55.

Seo, Ga Young, Sanggyu Park, Jung-Sik Huh and Moonjae Cho 2014. The protective effect of glycitin on UV-induced skin photoaging in human primary dermal fibroblast. Journal of the Korean Society for Applied Biological Chemistry 57: 463-468.

118

Seo, YK, SJ Kim, YC Boo, JH Baek, SH Lee and JS Koh 2011. Effects of p‐coumaric acid on erythema and pigmentation of human skin exposed to ultraviolet radiation. Clinical and experimental dermatology 36: 260-266.

SHIBASAKI, ISAO 1982. Food preservation with nontraditional antimicrobial agents. Journal of Food Safety 4: 35-58.

Si, Weiduo, Joshua Gong, Rong Tsao, Milosh Kalab, Raymond Yang and Yulong Yin 2006. Bioassay-guided purification and identification of antimicrobial components in Chinese green tea extract. Journal of chromatography A 1125: 204-210.

Silveira, Joana Léa Meira and Tania Mari Bellé Bresolin 2011. Pharmaceutical use of galactomannans. Química Nova 34: 292-299.

Singh, Rajbir, Bikram Singh, Sukhpreet Singh, Neeraj Kumar, Subodh Kumar and Saroj Arora 2010. Umbelliferone–An antioxidant isolated from Acacia nilotica (L.) Willd. ex. Del. Food Chemistry 120: 825-830.

Slavtcheff, Craig, Michael Brennan, Sandra Meraviglia, Jessica Goldberg and Michael Massaro 2005. Self-supporting aerosol cleansing composition. In Self-supporting aerosol cleansing composition: Google Patents.

Smith, Colin A 2013. LC/MS Preprocessing and Analysis with xcms. In LC/MS Preprocessing and Analysis with xcms.

Smith, Colin A, Elizabeth J Want, Grace O'Maille, Ruben Abagyan and Gary Siuzdak 2006. XCMS: processing mass spectrometry data for metabolite profiling using nonlinear peak alignment, matching, and identification. Analytical chemistry 78: 779-787.

Sroka, Z and W Cisowski 2003. Hydrogen peroxide scavenging, antioxidant and anti-radical activity of some phenolic acids. Food and Chemical Toxicology 41: 753-758.

Svobodová, Alena, Jitka Psotová and Daniela Walterová 2003. Natural phenolics in the prevention of UV-induced skin damage. A review. Biomed Pap Med Fac Univ Palacky Olomouc Czech Repub 147: 137-145.

Tabassum, Nadia, Ji-Hyung Lee, Soon-Ho Yim, Galzad Javzan Batkhuu, Da-Woon Jung and Darren R Williams 2016. Isolation of 4, 5-O-dicaffeoylquinic acid as a pigmentation inhibitor occurring in artemisia capillaris thunberg and its validation in vivo. Evidence-Based Complementary and Alternative Medicine 2016.

119

Takagi, Mio, Yosuke Tachi, Jie Zhang, Takuro Shinozaki, Kenta Ishii, Takashi Kikuchi, Motohiko Ukiya, Norihiro Banno, Harukuni Tokuda and Toshihiro Akihisa 2014. Cytotoxic and Melanogenesis‐Inhibitory Activities of Limonoids from the Leaves of Azadirachta indica (Neem). Chemistry & biodiversity 11: 451-468.

Tautenhahn, Ralf, Christoph Boettcher and Steffen Neumann 2008. Highly sensitive feature detection for high resolution LC/MS. BMC bioinformatics 9: 504.

Tautenhahn, Ralf, Gary J Patti, Duane Rinehart and Gary Siuzdak 2012. XCMS Online: a web- based platform to process untargeted metabolomic data. Analytical chemistry 84: 5035. Thaipong, Kriengsak, Unaroj Boonprakob, Kevin Crosby, Luis Cisneros-Zevallos and David Hawkins Byrne 2006. Comparison of ABTS, DPPH, FRAP, and ORAC assays for estimating antioxidant activity from guava fruit extracts. Journal of food composition and analysis 19: 669- 675.

Tomás‐Barberán, Francisco A and Michael N Clifford 2000. Flavanones, chalcones and dihydrochalcones–nature, occurrence and dietary burden. Journal of the Science of Food and Agriculture 80: 1073-1080.

Torel, Joseph, Josiane Cillard and Pierre Cillard 1986. Antioxidant activity of flavonoids and reactivity with peroxy radical. Phytochemistry 25: 383-385.

Tseng, Tsui-Hwa, Chau-Jong Wang and Erl-Shyh Kao 1996. Hibiscus protocatechuic acid protects against oxidative damage induced by tert-butylhydroperoxide in rat primary hepatocytes. Chemico-Biological Interactions 101: 137-148.

Tsuchihashi, Hideyasu, Makoto Kigoshi, Misato Iwatsuki and Etsuo Niki 1995. Action of β- carotene as an antioxidant against lipid peroxidation. Archives of biochemistry and biophysics 323: 137-147.

Ullrich, SE 1995. Potential for immunotoxicity due to environmental exposure to ultraviolet radiation. Human & experimental toxicology 14: 89-91.

Vignoli, JA, DG Bassoli and MT Benassi 2011. Antioxidant activity, polyphenols, caffeine and melanoidins in soluble coffee: The influence of processing conditions and raw material. Food Chemistry 124: 863-868.

Vittar, Natalia Belen Rumie, Laura Comini, Ivana Maria Fernadez, Elizabeth Agostini, Susana Nuñez-Montoya, Jose Luis Cabrera and Viviana Alicia Rivarola 2014. Photochemotherapy using natural anthraquinones: Rubiadin and Soranjidiol sensitize human cancer cell to die by apoptosis. Photodiagnosis and photodynamic therapy 11: 182-192.

120

Wang, Haibo, Muraleedharan G Nair, Gale M Strasburg, Yu-Chen Chang, Alden M Booren, J Ian Gray and David L DeWitt 1999. Antioxidant and antiinflammatory activities of anthocyanins and their aglycon, cyanidin, from tart cherries. Journal of natural products 62: 294-296.

Wang, Jun, Lu-Lu Zhao, Guo-Xia Sun, Yao Liang, Fu-An Wu, ZI Chen and SM Cui 2011. A comparison of acidic and enzymatic hydrolysis of rutin. African Journal of Biotechnology 10: 1460-1466.

Wang, Shiow Y and Hsin-Shan Lin 2000. Antioxidant activity in fruits and leaves of blackberry, raspberry, and strawberry varies with cultivar and developmental stage. Journal of agricultural and food chemistry 48: 140-146.

Wen, Bo, Zhanlong Mei, Chunwei Zeng and Siqi Liu 2017. metaX: a flexible and comprehensive software for processing metabolomics data. BMC bioinformatics 18: 183.

Williams, Christine A, Jeffrey B Harborne, Hans Geiger and J Robin S Hoult 1999. The flavonoids of Tanacetum parthenium and T. vulgare and their anti-inflammatory properties. Phytochemistry 51: 417-423.

Wölfle, Ute, Philipp R Esser, Birgit Simon-Haarhaus, Stefan F Martin, Jürgen Lademann and Christoph M Schempp 2011. UVB-induced DNA damage, generation of reactive oxygen species, and inflammation are effectively attenuated by the flavonoid luteolin in vitro and in vivo. Free Radical Biology and Medicine 50: 1081-1093.

Wu, Nan-Lin, Jia-You Fang, Marcelo Chen, Chia-Jung Wu, Chieh-Chen Huang and Chi-Feng Hung 2011. Chrysin protects epidermal keratinocytes from UVA-and UVB-induced damage. Journal of agricultural and food chemistry 59: 8391-8400.

Wu, Tian-Shung, Zhih-Jing Tsang, Pei-Lin Wu, Fu-Wen Lin, Chia-Yin Li, Che-Ming Teng and Kuo-Hsiung Lee 2001. New constituents and antiplatelet aggregation and anti-HIV principles of Artemisia capillaris. Bioorganic & medicinal chemistry 9: 77-83.

Xiao, Cheng, Jian Li, Xinxin Dong, Xiaojuan He, Xuyan Niu, Cha Liu, Guoyue Zhong, Rudolf Bauer, Dajian Yang and Aiping Lu 2011. Anti-oxidative and TNF-α suppressive activities of puerarin derivative (4AC) in RAW264. 7 cells and collagen-induced arthritic rats. European journal of pharmacology 666: 242-250.

Xing, Jing, Rui Li, Nan Li, Jian Zhang, Yueqing Li, Ping Gong, Dongna Gao, Hui Liu and Yu Zhang 2015. Anti-inflammatory effect of procyanidin B1 on LPS-treated THP1 cells via interaction with the TLR4–MD-2 heterodimer and p38 MAPK and NF-κB signaling. Molecular and cellular biochemistry 407: 89-95.

121

Yamakoshi, Jun, Fujio Otsuka, Atsushi Sano, Shoichi Tokutake, Makoto Saito, Mamoru Kikuchi and Yoshiro Kubota 2003. Lightening effect on ultraviolet‐induced pigmentation of guinea pig skin by oral administration of a proanthocyanidin‐rich extract from grape seeds. Pigment Cell & Melanoma Research 16: 629-638.

Yanagimoto, Kenichi, Hirotomo Ochi, Kwang-Geun Lee and Takayuki Shibamoto 2004. Antioxidative activities of fractions obtained from brewed coffee. Journal of agricultural and food chemistry 52: 592-596.

Yoon, Ji Hye, Tae-Gyu Lim, Kyung Mi Lee, Ae Ji Jeon, Su Yeon Kim and Ki Won Lee 2010. Tangeretin reduces ultraviolet B (UVB)-induced cyclooxygenase-2 expression in mouse epidermal cells by blocking mitogen-activated protein kinase (MAPK) activation and reactive oxygen species (ROS) generation. Journal of agricultural and food chemistry 59: 222-228.

Zhao, Mouming, Haiyan Wang, Bao Yang and Hong Tao 2010. Identification of cyclodextrin inclusion complex of chlorogenic acid and its antimicrobial activity. Food Chemistry 120: 1138- 1142.

Zhu, Qin Yan, Yu Huang and Zhen-Yu Chen 2000. Interaction between flavonoids and α- tocopherol in human low density lipoprotein. The Journal of nutritional biochemistry 11: 14-21. Zuorro, Antonio and Roberto Lavecchia 2012. Spent coffee grounds as a valuable source of phenolic compounds and bioenergy. Journal of Cleaner Production 34: 49-56.

122

Appendix A.

Table 3.1. Phenolic acids, flavonoids, and other bioactive compounds including coffee alkaloids, phenylpropanoids, stilbenoids, and anthraquinones

extracted from three different coffee wastes (Ethiopian Yirgacheffe, Costa Rican Tarrazu, and Hawaiian Kona). Maximum intensity of peaks detected by

XCMS online comparing the relative intensity between the three cultivars of SCG are shown.

Costa Compound name m/z RT Ethiopian Hawaiian Bioactivity Rican 16.76 1,351,108 1,099,826 1,535,870 Anti-tyrosinase, 19.35 41,746 x 40,944 E-Cinnamic acid 149.0604 antioxidant, 19.61 34,294 x 34,294 antimicrobial 4.25 37,560 38,420 x

123 3,5-Di-tert-butyl-4- 299.1617 21.59 346,552 329,470 334,008 hydroxycinnamic acid, (E)- 3,5-Di-tert-butyl-4- 299.1617 23.77 418,452 419,996 hydroxycinnamic acid, (E)- 3,4-Difluorohydrocinnamic acid 187.0569 0.74 927,880 927,880 x 2-ISOPROPYL-3- 221.1168 21.59 120,608 120,608 116,090 METHOXYCINNAMIC ACID Hydrocinnamic acid 151.0753 14.21 49,678 x 52,386 Flavoring agent 3,5-Di-tert-butyl-4- 299.1617 33.5 20,938 20,790 23,652 hydroxycinnamic acid, (E)- Hydrocinnamic acid 151.0755 23.79 40,730 42,226 42,226 Flavoring agent Hydrocinnamic acid, p-(3-(N- ethylacetamido)-2,4,6- 749.8794 18.84 2,894 2,894 3,226 triiodobenzamido)- 4-(Trifluoromethyl)hydrocinnamic 236.0895 2.63 36,746 acid

2-ISOPROPYL-3- 221.1168 20.38 36,354 36,386 36,354 METHOXYCINNAMIC ACID 2-ISOPROPYL-3- 221.1173 17.57 19,396 19,396 20,822 METHOXYCINNAMIC ACID 3,5-Di-tert-butyl-4- 277.1786 22.45 34,334 34,334 33,802 hydroxycinnamic acid, (E)- o-Methoxyhydrocinnamic acid 181.0848 6.03 29,768 34,924 27,494 3,5-Di-tert-butyl-4- 277.1786 21.6 19,322 19,322 19,322 hydroxycinnamic acid, (E)- 3,5-Di-tert-butyl-4- 279.1944 16.73 28,956 hydroxyhydocinnamic acid 4-Methoxy-(2E)-cinnamic acid 179.0705 9.73 17,846 14,142 Vanillic acid, 5-(4-carboxy-2- Antioxidant, 357.0585 3.47 68,432 14,468 20,036 methoxyphenoxy)- anti-tyrosinase P-coumaroylquinic acid 339.106 1.94 8,388

124 Anti-

Coumaric acid 349.0898 4.92 42,442 pigmentation, antioxidant Anti- pigmentation, trans-p-Coumaric acid 4-glucoside 349.0898 4.92 42,442 antioxidant, antimicrobial 1-Feruloyl-5-caffeoylquinic acid 531.1492 7.72 25,668 22,098 1,4-Dicaffeoylquinic acid 539.1146 5.83 71,610 78,570 1,4-Dicaffeoylquinic acid 539.1149 6.67 78,602 85,892 74,576 1-Feruloyl-5-caffeoylquinic acid 553.1315 7.71 142,308 119,366 1,4-Dicaffeoylquinic acid 517.1335 5.81 16,768 1,3-Dicaffeoylquinic acid 539.1149 6.67 78,602 85,892 74,576 1,5-Dicaffeoylquinic acid 539.1149 6.67 78,602 85,892 74,576 Antioxidant, 3,5-Di-O-caffeoylquinic acid 539.1149 6.67 78,602 85,892 74,576 anti-tyrosinase

Antioxidant, 4,5-Di-O-caffeoylquinic acid 539.1149 6.67 78,602 85,892 74,576 antipigmentation 1,4-Di-O-caffeoylquinic acid 539.1149 6.67 78,602 85,892 74,576 1,4-Dicaffeoylquinic acid 539.1146 5.83 71,610 78,570 61,832 Isochlorogenic acid A (3,5- Antioxidant, 539.1146 5.83 71,610 78,570 61,832 Dicaffeoylquinic acid) anti-tyrosinase 1,5-Dicaffeoylquinic acid 539.1146 5.83 71,610 78,570 61,832 1,3-Dicaffeoylquinic acid 539.1146 5.83 71,610 78,570 61,832 Antioxidant, Isochlorogenic acid A 539.1149 6.67 78,602 85,892 74,576 anti-tyrosinase Isochlorogenic acid b (3,4- 539.1149 6.67 78,602 85,892 74,576 Dicaffeoylquinic acid) 1.51 x 85,878 x Anti- inflammatory, Chlorogenic acid (5-Caffeoylquinic 1.82 x 245,560 200,058 antioxidant,

125 355.1025 acid) 2.29 x x x anti-aging, anti-

skin cancer, 3.07 x 134,046 94,340 antimicrobial 2.29 x 245,560 x 355.1025 1.51 x x 200,058 1-O-Caffeoylquinic acid 3.07 x 245,560 x 1.82 x 134,046 94,340

anti- inflammatory, Apigenin 8-C-(6''- antioxidant, 497.1064 6.69 20,560 x 20,746 acetylgalactoside) anticancer, UVinduced skin carcinogenesis anti- Apigenin 8-C-(6''- 497.108 5.83 27,832 24,098 22,526 inflammatory, acetylgalactoside) antioxidant,

anticancer, UVinduced skin carcinogenesis anti- inflammatory, Apigenin 4'-[feruloyl-(->2)- antioxidant, glucuronyl-(1->2)-glucuronide] 7- 992.2384 22.72 2,538 2,208 3,108 anticancer, glucuronide UVinduced skin carcinogenesis anti- inflammatory, Apigenin 7-[feruloyl-(->2)- antioxidant, glucuronyl-(1->2)-glucuronide] 4'- 992.2384 22.72 2,538 2,208 3,108 anticancer, glucuronide UVinduced skin carcinogenesis anti-

126 inflammatory,

Apigenin 7-[feruloyl-(->2)- antioxidant, [glucuronyl-(1->3)]-glucuronyl-(1- 992.2384 22.72 2,538 2,208 3,108 anticancer, >2)-glucuronide] UVinduced skin carcinogenesis anti- inflammatory, Apigenin 7-rutinoside-4'-trans- antioxidant, 779.1633 27.69 2,454 3,200 3,194 caffeate anticancer, UVinduced skin carcinogenesis anti- inflammatory, antioxidant, Apigenin 7-rhamnoside 439.0999 10.98 21,104 x 19,424 anticancer, UVinduced skin carcinogenesis

anti- inflammatory, Apigenin 7-(2'',6''-di-p- antioxidant, 763.1482 20.24 2,718 4,206 3,402 coumarylglucoside) anticancer, UVinduced skin carcinogenesis anti- inflammatory, Apigenin 7-(3'',6''-di-p- antioxidant, 763.1482 20.24 2,718 4,206 3,402 coumarylglucoside) anticancer, UVinduced skin carcinogenesis anti- inflammatory, Apigenin 7-(4'',6''-di-p- antioxidant, 763.1482 20.24 2,718 4,206 3,402 coumarylglucoside) anticancer,

127 UVinduced skin carcinogenesis anti- inflammatory, Apigenin 7-(3'',6''-Di-E-p- antioxidant, 763.1482 20.24 2,718 4,206 3,402 coumaroylgalactoside) anticancer, UVinduced skin carcinogenesis anti- inflammatory, Apigenin 7-(2'',3''- antioxidant, 539.1149 6.67 78,602 85,892 74,576 diacetylglucoside) anticancer, UVinduced skin carcinogenesis anti- Apigenin 7-(3'',4''- 539.1149 6.67 78,602 85,892 74,576 inflammatory, diacetylglucoside) antioxidant,

anticancer, UVinduced skin carcinogenesis anti- inflammatory, Apigenin 7-O-(6-O- antioxidant, 558.0948 34.86 242,958 x x malonylglucoside) anticancer, UVinduced skin carcinogenesis anti- inflammatory, Apigenin 7-(2'',3''- antioxidant, 539.1146 5.83 71,610 78,570 61,832 diacetylglucoside) anticancer, UVinduced skin carcinogenesis anti-

128 inflammatory, Apigenin 7-(3'',4''- antioxidant, 539.1146 5.83 71,610 78,570 61,832 diacetylglucoside) anticancer, UVinduced skin carcinogenesis anti- inflammatory, Apigenin 7-neohesperidoside-4'- antioxidant, 925.254 35.68 40,232 36,984 63,418 sophoroside anticancer, UVinduced skin carcinogenesis anti- inflammatory, Apigenin 7-(6'''-acetylallosyl-(1- antioxidant, 659.1566 32.85 75,804 27,948 55,172 >2)glucoside) anticancer, UVinduced skin carcinogenesis

anti- inflammatory, Apigenin 7-(6''-acetylalloside)-4'- antioxidant, 659.1566 32.85 75,804 27,948 55,172 alloside anticancer, UVinduced skin carcinogenesis anti- inflammatory, 8-C-beta-D- antioxidant, Glucofuranosylapigenin 2''-O- 497.108 5.83 27,832 24,098 22,526 anticancer, acetate UVinduced skin carcinogenesis Fujikinetin 7-O-glucoside 497.108 5.83 27,832 24,098 22,526 anti- Genistin 6''-O-acetate 497.108 5.83 27,832 24,098 22,526 carcinogeninc

129 Antioxidative, Vitexin 2''-acetate 497.108 5.83 27,832 24,098 22,526 anti- inflammatory Isovitexin 6''-O-acetyl 497.108 5.83 27,832 24,098 22,526 Antioxidative, Vitexin 3''-O-acetate 497.108 5.83 27,832 24,098 22,526 anti- inflammatory Antioxidative, Vitexin 6''-O-acetate 497.108 5.83 27,832 24,098 22,526 anti- inflammatory Antioxidant, Isovitexin 2''-O-(6'''-(E)-p- 779.1633 27.69 2,454 3,200 3,194 anti- coumaroyl)glucoside inflammatory Antioxidative, Vitexin 7-O-glucoside 2''-p- 779.1633 27.69 2,454 3,200 3,194 anti- coumarate inflammatory

Antioxidant, Isovitexin 7-O-(6'''-O-E-p- 779.1633 27.69 2,454 3,200 3,194 anti- coumaroyl)glucoside inflammatory Anti- Daidzein 4'-O-glucoside 439.0999 10.98 21,104 x 19,424 inflammatory, anti-carcinogenic Anti- inflammatory, anti- 3',4'-Dihydroxyflavone 4'-glucoside 439.0999 21,104 x 19,424 carcinogenic, anti- inflammatory anti- Chrysin 8-C-beta-D- inflammation, 439.0999 10.98 21,104 x 19,424 glucopyranoside anti-cancer, and anti-oxidation

130 anti- Chrysin 6-C-beta-D- inflammation, 439.0999 10.98 21,104 x 19,424 glucopyranoside anti-cancer, and anti-oxidation anti- inflammation, Chrysin 7-galactoside 439.0999 10.98 21,104 x 19,424 anti-cancer, and anti-oxidation anti- inflammation, Chrysin-7beta-monoglucoside 439.0992 11 21,104 18,130 19,424 anti-cancer, and anti-oxidation Anti- Daidzein-4,7-diglucoside 601.1539 15.2 5,336 4,514 4,628 inflammatory, anti-carcinogenic

Antioxidative, Vitexin-2''-rhamnoside 601.1539 15.2 5,336 4,514 4,628 anti- inflammatory Antioxidative, Vitexin rhamnoside 601.1539 15.2 5,336 4,514 4,628 anti- inflammatory Antioxidant, Puerarin 4'-O-glucoside 601.1539 15.2 5,336 4,514 4,628 anti- inflammatory Anti- Daidzein 7,4'-di-O-glucoside 601.1539 15.2 5,336 4,514 4,628 inflammatory, anti-carcinogenic Anti- inflammatory, Genistein 7-O-rutinoside 601.1539 15.2 5,336 4,514 4,628 antioxidant and

131 anticancer Anti- inflammatory, Genistein 4'-O-neohesperidoside 601.1539 15.2 5,336 4,514 4,628 antioxidant and anticancer Skin protection, anti- Chrysin 7-gentiobioside 601.1539 15.2 5,336 4,514 4,628 inflammation and anti- oxidation Antioxidant, Isovitexin 2''-O-rhamnoside 601.1539 15.2 5,336 4,514 4,628 anti- inflammatory Anti- Coumestrol 3-O-glucoside 469.0529 29.75 2,420 2,684 2,380 inflammatory Skin protection, Chrysin 7-glucuronide 469.0529 29.75 2,420 2,684 2,380 anti- inflammation

and anti- oxidation Anti- Daidzein 4'-O-glucuronide 469.0529 29.75 2,420 2,684 2,380 inflammatory, anti-carcinogenic Anti- Daidzein 7-O-glucuronide 469.0529 29.75 2,420 2,684 2,380 inflammatory, anti-carcinogenic Anti- 7,8-DIHYDROXYFLAVONE 255.0655 20.28 8,512 6,930 7,272 inflammatory, antioxidant 3,7-DIHYDROXYFLAVONE 255.0655 20.28 8,512 6,930 7,272 anticancer Anti- Eriodictyol (5,7,3′,4′- inflammatory,

tetrahydroxylflavanone) Anticancer,

132 antioxidant

6-Methoxyeriodictyol 319.0811 3.78 30,916 28,356 Anti-tyrosinase Homoeriodictyolchalcone 2'- 465.1413 24.3 15,284 19,588 glucoside Anti- Daidzein 255.0655 20.28 8,512 6,930 7,272 inflammatory, anti-carcinogenic Skin protection, anti- Chrysin 255.0655 20.28 8,512 6,930 7,272 inflammation and anti- oxidation 3,2'-Dihydroxyflavone 255.0655 20.28 8,512 6,930 7,272 2',3'-Dihydroxyflavone 255.0655 20.28 8,512 6,930 7,272 3,3'-DIHYDROXYFLAVONE 255.0655 20.28 8,512 6,930 7,272 6,2'-Dihydroxyflavone 255.0655 20.28 8,512 6,930 7,272

Antioxidant, Tectochrysin 291.0627 8.42 22,418 x x anticancer Anti-photoaging, Anti- Glycitin 464.1547 23.9 3,750 x x inflammatory, Anti-aging Retusin 7-O-glucoside 464.1547 23.9 3,750 x x 8-Hydroxyluteolin 4'-methyl ether 7-(6'''-acetylallosyl)(1->2)(6''- 763.1482 20.24 2,718 4,206 3,402 acetylglucoside) Isorhamnetin 3-(4''',6'''- Anti- diacetylglucosyl) (1->3)- 763.1482 20.24 2,718 4,206 3,402 inflammatory, galactoside Anticancer 2',2'-BISEPIGALLOCATECHIN 763.1482 20.24 2,718 4,206 3,402 MONOGALLATE

133 8-Hydroxyluteolin 4'-methyl ether 7-(6'''-acetylallosyl)(1->2)(6''- 747.1806 6.48 5,678 5,356 4,544 acetylglucoside) Isorhamnetin 3-(4''',6'''- Anti- diacetylglucosyl) (1->3)- 747.1806 6.48 5,678 5,356 4,544 inflammatory, galactoside Anticancer Rhamnocitrin 3-(5'''-p- anti- 779.1633 27.69 2,454 3,200 3,194 coumarylapiosyl)-(1->2)-glucoside inflammatory Formononetin 7-O-glucoside-6''- 539.1149 6.67 78,602 85,892 74,576 antioxidants malonate Formononetin 7-(6''- 539.1149 6.67 78,602 85,892 74,576 antioxidants malonylglucoside) Formononetin 7-O-glucoside-6''-O- 539.1149 6.67 78,602 85,892 74,576 antioxidants malonate Formononetin 7-O-glucoside-6''- 539.1149 6.67 78,602 85,892 74,576 antioxidants malonate 2'',6''-Di-O-Acetyl isovitexin 539.1149 6.67 78,602 85,892 74,576

Formononetin 7-O-glucoside-6''- 539.1146 5.83 71,610 78,570 61,832 antioxidants malonate 2'',6''-Di-O-Acetyl isovitexin 539.1146 5.83 71,610 78,570 61,832 Formononetin 7-O-glucoside-6''-O- 539.1146 5.83 71,610 78,570 61,832 antioxidants malonate Formononetin 7-(6''- 539.1146 5.83 71,610 78,570 61,832 antioxidants malonylglucoside) Fujikinetin 7-O-laminaribioside 659.1566 32.85 75,804 27,948 55,172 Anti-bacterial, anti-oxidant, Quercetin 3-sophoroside-7- 820.2114 4.57 3,178 2,780 2,312 skin protection, glucuronide anti- melanogenesis Anti-bacterial, anti-oxidant, Quercetin 3-gentiobioside-7- 134 820.2114 4.57 3,178 2,780 2,312 skin protection, glucuronide anti-

melanogenesis Anti-bacterial, anti-oxidant, Quercetin 7-glucuronoside 3- 820.2114 4.57 3,178 2,780 2,312 skin protection, sophoroside anti- melanogenesis Anti-bacterial, anti-oxidant, Quercetin 3-(3'',6''-di-p- 779.1633 27.69 2,454 3,200 3,194 skin protection, coumarylglucoside) anti- melanogenesis Anti-bacterial, anti-oxidant, Quercetin 3-(4''-acetylrhamnoside)- 659.1566 32.85 75,804 27,948 55,172 skin protection, 7-rhamnoside anti- melanogenesis

Anti- Kaempferol 3-(4''-(Z)-p- 779.1633 27.69 2,454 3,200 3,194 inflammatory coumarylrobinobioside) ,antioxidant Anti- Kaempferol 3-(4''-(E)-p- 779.1633 27.69 2,454 3,200 3,194 inflammatory coumarylrobinobioside) ,antioxidant Anti- Kaempferol 3-rhamnosyl-(1- 779.1633 27.69 2,454 3,200 3,194 inflammatory >6)(2''-p-coumarylglucoside) ,antioxidant Kaempferol 3- [ 6'''-p- Anti- coumarylglucosyl- (1->2) - 779.1633 27.69 2,454 3,200 3,194 inflammatory rhamnoside ] ,antioxidant Anti- Kaempferol 3-(2G-(E)-p- 779.1633 27.69 2,454 3,200 3,194 inflammatory coumaroylrutinoside) ,antioxidant

135 Anti- Kaempferol 2G- 779.1633 27.69 2,454 3,200 3,194 inflammatory

coumaroylrutinoside ,antioxidant Anti- Kaempferol 3-[2''-(p- 779.1633 27.69 2,454 3,200 3,194 inflammatory coumaroylglucosyl)rhamnoside] ,antioxidant Anti- Kaempferol 3-[4''-(p- 779.1633 27.69 2,454 3,200 3,194 inflammatory coumaroylglucosyl)rhamnoside] ,antioxidant Kaempferol 3-[6'''-p- Anti- coumarylglucosyl-(1->2)- 779.1633 27.69 2,454 3,200 3,194 inflammatory rhamnoside] ,antioxidant Anti- Kaempferol 3-(2''-galloyl-alpha-L- 588.139 12.4 4,606 3,658 3,146 inflammatory arabinopyranoside) ,antioxidant Anti-tyrosinase, 6-Hydroxykaempferol 3,6- 820.2114 4.57 3,178 2,780 2,312 anti- diglucoside 7-glucuronide inflammatory

Kaempferol 3-[2''',3''',5'''-triacetyl- Anti- alpha-L-arabinofuranosyl-(1->6)- 707.1755 19.09 4,230 3,716 4,370 inflammatory glucoside ,antioxidant Kaempferol 3-(2''-(E)-p-coumaroyl- Anti- alpha-L-arabinofuranoside)-7- 711.1949 7.54 4,010 4,650 4,184 inflammatory rhamnoside ,antioxidant Anti- Kaempferol 3-(2'',4''-di-(E)-p- 763.1482 20.24 2,718 4,206 3,402 inflammatory coumarylrhamnoside) ,antioxidant Anti- Kaempferol 3-(2'',4''- 539.1149 6.67 78,602 85,892 74,576 inflammatory diacetylrhamnoside) ,antioxidant Kaempferol 3-(2Gal- Anti- glucosylrobinobioside)-7- 925.254 35.68 40,232 36,984 63,418 inflammatory rhamnoside ,antioxidant

136 Kaempferol 3-rhamnosyl-(1- Anti- >2)[rhamnosyl-(1->6)-galactoside]- 925.254 35.68 40,232 36,984 63,418 inflammatory 7-glucoside ,antioxidant Kaempferol 3-galactosyl-(1->6)- Anti- glucoside-7-rhamnosyl-(1->3)- 925.254 35.68 40,232 36,984 63,418 inflammatory rhamnoside ,antioxidant Kaempferol 3-O-alpha-L- rhamnopyranosyl(1->6)-beta-D- Anti- glucopyranosyl(1->2)-beta-D- 925.254 35.68 40,232 36,984 63,418 inflammatory glucopyranoside-7-O-alpha-L- ,antioxidant rhamnopyranoside Kaempferol 3-glucosyl-(1->3)- Anti- rhamnosyl-(1->2)-[rhamnosyl-(1- 925.254 35.68 40,232 36,984 63,418 inflammatory >6)-galactoside] ,antioxidant Anti- Kaempferol 3-(6''- 659.1566 32.85 75,804 27,948 55,172 inflammatory acetylgalactoside)-7-rhamnoside ,antioxidant

Anti- Kaempferol 3-(6''-acetylglucoside)- 659.1566 32.85 75,804 27,948 55,172 inflammatory 7-rhamnoside ,antioxidant Anti- Rhamnazin 3-[6''-(3-hydroxy-3- 659.1566 32.85 75,804 27,948 55,172 inflammatory methylglutaryl)glucoside] ,antioxidant Anti- Kaempferol 3-(6''- 659.1566 32.85 75,804 27,948 55,172 inflammatory acetylgalactoside) 7-rhamnoside ,antioxidant Rhamnazin 3-[6''-(3-hydroxy-3- 659.1566 32.85 75,804 27,948 55,172 methylglutaryl)glucoside] Anti- Kaempferol 3-(2'',4''- 539.1146 5.83 71,610 78,570 61,832 inflammatory diacetylrhamnoside) ,antioxidant Anti- Kaempferol 3-(3'',4''-

137 539.1146 5.83 71,610 78,570 61,832 inflammatory diacetylrhamnoside) ,antioxidant Kaempferol 3-[glucosyl-(1->3)- Anti- rhamnosyl-(1->2)-[rhamnosyl-(1- 925.254 35.68 40,232 36,984 63,428 inflammatory >6)-galactoside]] ,antioxidant Antioxidant, Anti- Cyanidin 3-O-(2''-O-galloyl-6''-O- inflammatory, alpha-rhamnopyranosyl-beta- 747.1806 6.48 5,678 5,356 4,544 Anti-cancer, galactopyranoside Skin protection, anti-aging Antioxidant, Anti- inflammatory, Cyanidin 3-(2G-galloylrutinoside) 747.1806 6.48 5,678 5,356 4,544 Anti-cancer, Skin protection, anti-aging

Antioxidant, Anti- Cyanidin 3-O-(2''-O-galloyl-6''-O- inflammatory, alpha-rhamnopyranosyl-beta- 764.203 15.53 4,478 5,480 4,478 Anti-cancer, galactopyranoside Skin protection, anti-aging Antioxidant, Anti- inflammatory, Cyanidin 3-(2G-galloylrutinoside) 764.203 15.53 4,478 5,480 4,478 Anti-cancer, Skin protection, anti-aging 6-Glucopyranosylprocyanidin B2 779.1633 27.69 2,454 3,200 3,194 Antioxidant Anti- 6-Glucopyranosylprocyanidin B1 779.1633 27.69 2,454 3,200 3,194 inflammatory

138 8-Glucopyranosylprocyanidin B2 779.1633 27.69 2,454 3,200 3,194 Antioxidant Anti- 8-Glucopyranosylprocyanidin B1 779.1633 27.69 2,454 3,200 3,194 inflammatory Procyanidin B3 7-glucoside 779.1633 27.69 2,454 3,200 3,194 antioxidant Antioxidant, Anti- inflammatory, Cyanidin-3-(6'-malonylglucoside) 558.0948 34.86 242,958 x x Anti-cancer, Skin protection, anti-aging Antioxidant, Anti- inflammatory, Cyanidin 3-(3''-malonyl-glucoside) 558.0948 34.86 242,958 x x Anti-cancer, Skin protection, anti-aging

Antioxidant, Anti- inflammatory, Cyanidin 3-(4'''-acetylrutinoside) 659.1566 32.85 75,804 27,948 55,172 Anti-cancer, Skin protection, anti-aging Antioxidant, Anti- inflammatory, Cyanidin-3-(2'-acetylrutinoside) 659.1566 32.85 75,804 27,948 55,172 Anti-cancer, Skin protection, anti-aging Anti- Pelargonidin 3-rhamnoside 439.0999 10.98 21,104 x 19,424 inflammatory, antioxidant Anti-

139 Pelargonidin 3-robinobioside 601.1539 15.2 5,336 4,514 4,628 inflammatory, antioxidant Anti- Pelargonidin 3-neohesperidoside 601.1539 15.2 5,336 4,514 4,628 inflammatory, antioxidant Anti- Pelargonidin 3-rutinoside 601.1539 15.2 5,336 4,514 4,628 inflammatory, antioxidant Anti- Pelargonidin 3-rhamnoside-5- 601.1539 15.2 5,336 4,514 4,628 inflammatory, glucoside antioxidant Anti- Pelargonidin 3-[6-((Z)-p- 779.1633 27.69 2,454 3,200 3,194 inflammatory, coumaroyl)glucoside]-5-glucoside antioxidant Anti- Pelargonidin 3-(6''-acetylglucoside) 497.1064 6.69 20,560 x 20,746 inflammatory, antioxidant

Anti- Pelargonidin 3-(6''-acetylglucoside) 497.108 5.83 27,832 24,098 22,526 inflammatory, antioxidant Anti- Pelargonidin 3-O-β-D-glucoside 5- 779.1633 27.69 2,454 3,200 3,194 inflammatory, O-(6-coumaroyl-β-D-glucoside) antioxidant Anti- Pelargonidin 3-(6''- 659.1566 32.85 75,804 27,948 55,172 inflammatory, acetylglucoside)-5-glucoside antioxidant Anti- Pelargonidin 3-glucoside-5-(6''- 659.1566 32.85 75,804 27,948 55,172 inflammatory, acetylglucoside) antioxidant Malvidin 3-O-(6-O-(4-O-malonyl- Anti- alpha-rhamnopyranosyl)-beta- 763.1482 20.24 2,718 4,206 3,402 inflammatory, glucopyranoside antioxidant

140 Malvidin 3-O-(6-O-(4-O-malonyl- Anti- alpha-rhamnopyranosyl)-beta- 747.1806 6.48 5,678 5,356 4,544 inflammatory, glucopyranoside antioxidant Anti- malvidin-3-O-glucoside- 539.1149 6.67 78,602 85,892 74,576 inflammatory, acetaldehyde antioxidant Anti- Malvidin 3-glucoside-4- 677.1273 14.37 3,758 3,148 3,960 inflammatory, vinylguaiacol antioxidant Anti- Malvidin 3-glucoside-4- 639.1679 31.59 2,922 3,126 3,246 inflammatory, vinylguaiacol antioxidant Anti- Malvidin 3-glucoside-4- 661.1548 29.01 3,202 3,348 3,202 inflammatory, vinylguaiacol antioxidant Dihydroisorhamnetin 319.081 9.7 62,062

Anti- Isorhamnetin 5-glucoside 499.1222 5.85 58,816 inflammatory, Anticancer Anti- Isorhamnetin 3-(4''- 705.1318 33.64 356,714 34,344 685,630 inflammatory, sulfatorutinoside) Anticancer Anti- Isorhamnetin 3-(3'''- 839.1756 36.5 1,629,892 199,560 inflammatory, ferulylrobinobioside) Anticancer Anti- Isorhamnetin 3-glucosyl-(1->2)- 841.1747 36.49 1,093,388 inflammatory, galactoside-7-glucoside Anticancer Anti- Isorhamnetin 3-(6''-(E)- 864.2498 36.48 254,224 inflammatory, sinapoylsophoroside) Anticancer

141 Anti- Isorhamnetin 3-rhamnosyl-(1->2)- 949.2854 36.65 372,300 397,468 inflammatory,

gentiobiosyl-(1->6)-glucoside Anticancer Hesperetin 8-Hydroxyhesperetin 319.081 9.7 62,062 x x antioxidant Photo protective, Anti- Hesperetin 5-O-glucoside 465.1395 24.23 15,284 19,588 inflammatory, antioxidant Antioxidant; Emollient; Skin- Hesperetin 5,7-laurate 667.4204 36.05 192,366 237,416 257,946 Conditioning Agent - Miscellaneous Photo protective, Naringenin 273.0741 3.7 19,376 15,444 Anti- inflammatory,

anti-oxidant and anti-tumor 3'-Geranylchalconaringenin 409.2032 28.44 35,230 Chalconaringenin 2',4'-di-O- 619.1579 2.63 30,088 glucoside Photo protective, Anti- Naringenin 295.0583 5.12 11,234 inflammatory, anti-oxidant and anti-tumor Photo protective, Anti- Naringenin-7-O-Glucoside 457.1098 7.95 45,070 inflammatory, anti-oxidant and anti-tumor

142 Photo protective, Anti- Naringenin 7-(2-p- 619.1167 33.84 354,694 inflammatory, Coumaroylglucoside) anti-oxidant and anti-tumor Chalconaringenin 2'-O-glucoside 797.1941 36.2 59,314 4'-O-gentobioside Daidzein Dihydrodaidzein diacetate 379.058 4.19 23,852 Anti- inflammatory, anti- Daidzein 4'-O-glucoside 439.0994 8.27 21,104 carcinogenic, anti- inflammatory Anti- Daidzein 6''-O-acetate 459.1266 24.31 27,846 inflammatory,

anti- carcinogenic, anti- inflammatory Anti- inflammatory, anti- Daidzein 7,4'-di-O-glucoside 579.1673 5.76 12,738 carcinogenic, anti- inflammatory Anti- Genistein 7,4'-bis(O- inflammatory, 859.2501 36.63 64,594 glucosylapioside) antioxidant and anticancer 6,8-Diprenylgenistein 407.1835 24.31 141,212

143 Prevent UV- Genistin 433.1105 5.57 59,032 induced skin damage Prevent UV- Genistein 7-O-rutinoside 579.1673 5.76 12,738 induced skin damage Taxifolin 6-Methoxytaxifolin 357.0585 3.47 68,432 20,036 6-Methoxytaxifolin 357.0583 4.7 14,468 Skin photo protection, Naringin 4'-glucoside 743.2366 4.19 3,344 antioxidant, anti- carcinogenic Chrysin Skin protection, Chrysin-7beta-monoglucoside 439.0994 8.27 21,104 18,130 anti- inflammation

and anti- oxidation 2-ETHOXYCARBONYL-2- ETHOXYOXALOYLOXYDIHYD 495.1057 9.02 4,408 ROCHRYSIN DIMETHYL ETHER Strychochrysine 601.3394 12.22 28,694 Helichrysin 449.1474 2.63 30,618 Skin protection, anti- Chrysin 7-(4''-acetylglucoside) 459.1266 24.31 27,846 inflammation and anti- oxidation Skin protection, anti-

144 Chrysin 7-gentiobioside 579.1673 5.76 12,738 inflammation and anti-

oxidation Epigallocatechin 3-O-vanillate 457.1109 5.92 39,556 31,318 8-C-Ascorbylepigallocatechin 3- 633.1122 35.86 799,104 246,502 691,586 gallate Epigallocatechin-(4beta->8)- 747.1512 34.53 34.53 27,012 257,604 epicatechin-3-O-gallate ester Epigallocatechin-(4beta->8)- 764.1797 33.57 74,912 20,244 epicatechin-3-O-gallate ester Epigallocatechin 3-O-gallate- 921.1464 33.77 74,102 (4beta->6)-epicatechin 3-O-gallate 2',2'-BISEPIGALLOCATECHIN 932.185 36.66 74,360 DIGALLATE Epigallocatechin-(4beta->8)- 764.18 34.03 400,350 epicatechin-3-O-gallate ester

Epigallocatechin-(4beta->8)- 764.1809 36.2 323,184 epicatechin-3-O-gallate ester Epigallocatechin 3-O-gallate- 921.1564 35.45 65,912 (4beta->6)-epicatechin 3-O-gallate Epigallocatechin 3-O-gallate- 921.1551 36.23 49,962 (4beta->6)-epicatechin 3-O-gallate Theaflavin Theaflavin monogallate A 717.1419 3.7 13,048 Antioxidant THEAFLAVIN DIGALLATE 891.1442 34.98 50,776 antioxidant Antioxidant, 195.088 Anti-skin cancer, Caffeine (alkaloid) 2.63 36,703,416 38,792,408 1,861,824 2 antimicrobial, emollient

145 195.088 Isocaffeine 2.63 36,703,416 38,792,408 1,861,824 2

Anti-microbial, anti- 179.105 inflammatory, Eugenol (phenylpropanoid) 14.22 329,800 238,938 343,726 7 analgesic, antioxidant and anticancer 229.086 Anti- 7.71 x x 106,592 2 inflammatory, Resveratrol (stilbenoid) antioxidant, anti- 251.068 aging, photo 36.7 x 6,864 10,756 1 protection

229.086 Guaiacol benzoate 7.71 x x 106,592 Anti-microbial 2

Anti-cancer, anti- Chrysophanol (anthraquinone) inflammatory antimicrobial Anticancer, anti- Chrysophanol 1-tetraglucoside 925.254 35.68 40,232 36,984 63,428 inflammatory, anti-microbial 255.065 Soranjidiol 20.28 8,512 6,930 7,272 Anti-cancer 5 Antibacterial, 464.154 anti- Physcion 8-glucoside 23.90 3,750 x x 7 inflammatory, anti-skin cancer

146

Appendix B.

Whitening activities of the spent coffee grounds

1. Introduction

In addition to antioxidant and anti-microbial activity, skin whitening effect is emerging as a major interest in cosmetic market. The pigmentation (melanogenesis) on the skin causes multiple issues including lentigo, and freckles. Melanin is produced by oxidation reaction of tyrosine by tyrosinase, oxidase controlling the melanin production (Saghaie et al.,

2013). Multiple natural whitening chemicals extracted from plants such as arbutin, bisaborol, ricorice has been used in real product by inhibiting tyrosinase (Khaiat and

Saliou, 2015). Besides, tyrosinase inhibition, there are other methods in blocking skin- pigmentation process. For example, vitamin C derivatives such as ascorbyl glucoside, and ascorbyl tetraisopalmitate act as antioxidant inhibiting oxidation of tyrosine.

147

AppendixB1. Mode of action of anti-pigmentation

148

AppendixB.2. Melanin synthesis pathway

149

2. Objectives

In Appendix B, the tyrosinase inhibition assay was conducted to evaluate skin whitening effect of SCG extracts from three cultivars.

3. Materials and Methods

Tyrosinase inhibition activities (skin-whitening property)

Tyrosinase inhibition activity will be determined as described by (Momtaz et al., 2008) with tyrosine as substrates. Samples were dissolved in dimethyl sulfoxide (DMSO) to a concentration of 20 mg/ml, and further diluted in potassium phosphate buffer (50 mM, pH

6.5) to 600 μg/ml. Assays were carried out in a Falcon 96 flat bottom clear plate and

SYNERGY HT microplate reader (BIO-TEK) was used. All the steps in the assay were conducted at room temperature. In triplicate, each prepared sample (70 μl) was mixed with

30 μl of tyrosinase (333 Units/ml in phosphate buffer, pH 6.5). After 5 min incubation, 110

μl of substrate (2 mM L -tyrosine) was added to the reaction mixtures and incubated further for 30 min. Arbutin (0-1100 µg/ml) was used as a positive control while a blank test was used as each sample that had all the components except L-tyrosine. Results were compared with a control consisting of DMSO instead of the test sample. Absorbance values of the wells were then determined at 492 nm. The percentage tyrosinase inhibition was calculated as follows:

%inhibition=[(Acontrol–Asample)/Acontrol]x100 where A control is the absorbance of

DMSO and A sample is the absorbance of the test reaction mixture containing extract or arbutin.

150

4. Results and discussion

All SCG extracts including three cultivars had neglectable effect on tyrosinase inhibition test. However, SCG had neglectable. From previous studies, several compounds that can be found in SCG such as E-cinnamic acid, vanillic acid, 3,5-Di-O-caffeoylquinic acid are reported as anti-tyrosinase agents (Kong et al., 2008) (Chou et al., 2010) (Iwai et al.,

2004). However, from out study, SCG extracts from Ethiopian Yirgacheffe, Costa Rian

Tarrazu and Hawaiian Kona did not show any anti-tyrosinase activities. Our findings are not consistent with other studies which have reported that SCG and coffee bean can act as a skin-whitening agents related to anti-tyrosinase properties (Kanlayavattanakul and

Lourith, 2013) (Kiattisin et al., 2016). In latter study they conducted anti-tyrosinase and anti-oxidant assay using coffee bean including green and roasted. Even though they found ethanolic extracts of green coffee bean which has the highest antioxidant activities showed that highest anti-tyrosinase activity (44.27 inhibition %), all the ethanolic extracts including green bean and roasted bean presented a lower anti-tyrosinase activity than kojic acid and arbutin. From this result, they concluded that antioxidant compounds might promote the tyrosinase inhibition activity due to their antioxidative synergistic

(Chang, 2009). Our result indicates that anti-tyrosinase effect is not a major mode of action in SCG. Therefore, the skin-whitening effect of coffee extracts and SCG may more related to mode of action in anti-pigmentation as antioxidant agents.

151

120

100

Series1

40 %inhibition

0 -200 0 200 400 600 800 1000 1200 -20 µg/ml

Appendix B3. Calibration curve of the positive control (arbutin). Each dot in X axis represents around 0,

1.1, 11, 110, 1100 µg/ml concentration of arbutin (STD; n = 5).

152