Development of Specific Natural Bacterial Beta-Glucuronidase Inhibitors for Reducing Irinotecan-Associated Diarrhea

Wayne State University

Wayne State University Dissertations

January 2020

Development Of Specific Natural Bacterial Beta-Glucuronidase Inhibitors For Reducing Irinotecan-Associated Diarrhea

Fei Yang Wayne State University

Follow this and additional works at: https://digitalcommons.wayne.edu/oa_dissertations

Part of the Food Science Commons

Recommended Citation Yang, Fei, "Development Of Specific Natural Bacterial Beta-Glucuronidase Inhibitors For Reducing Irinotecan-Associated Diarrhea" (2020). Wayne State University Dissertations. 2515. https://digitalcommons.wayne.edu/oa_dissertations/2515

This Open Access Dissertation is brought to you for free and open access by DigitalCommons@WayneState. It has been accepted for inclusion in Wayne State University Dissertations by an authorized administrator of DigitalCommons@WayneState.

DEVELOPMENT OF SPECIFIC NATURAL BACTERIAL BETA-GLUCURONIDASE INHIBITORS FOR REDUCING IRINOTECAN-ASSOCIATED DIARRHEA

FEI YANG

DISSERTATION

Submitted to the Graduate School

of Wayne State University,

Detroit, Michigan

in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

2020

MAJOR: NUTRITION AND FOOD SCIENCE

Approved By:

Advisor Date

DEDICATION

To my beloved father and mother.

To my darling wife and daughter.

ACKNOWLEDGMENTS

My deepest and sincere gratitude goes first to my supervisor, Prof. Dr. Kequan Zhou. He opened a window for me to the amazing world of gut microbiota by bacterial β-glucuronidases and new probiotics, which gave me a chance to improve professional skills, and an insight into the state-of-art of my research fields. During these wonderful years with him, his competent guidance and instruction gave me the power to move forward, his constant trust and support provided me the confidence to overcome any difficulties faced. More importantly, his scientific spirit of suspicion, innovation and pursing truth have greatly enlightened me and will benefit my endeavors in the future. Never enough words to say thank you for his help and supports.

I am very grateful to Dr. K-L Catherine Jen, Dr. Smiti V. Gupta, and Dr. Weilong Hao, the honorable members of my dissertation committee. From the meetings we had, their valuable advices clarified my direction, and their brilliant ideas enabled me to settle specific problems. I am so lucky to have them on my committee.

I extend my gratitude to the department of Nutrition and Food Science and the Graduate

School of Wayne State University for offering me the scholarships and awards. My special thanks are addressed to all the faculty and staff of our department for their professionalism and considerate supports.

How to miss this unique chance to thank the present and past members in Dr. Zhou’s lab:

Wenjun Zhu, Qing Ai, Paba Edirisuriya, Dr. Shi Sun, Dr. Jiangqi Tang (Kiki), Dr. Kai Nie,

Lingjing Liu, Jiarun Cui, Jun Ma, Sampurna Guhathakurta, Maria Elena Hakim and Ninghui Zhou et al. We did research together, learn together, and play together. We share the joys when someone succeeds, we give a hand when someone is in difficulty. What a big and sweet family.

iii

I would also like to acknowledge Dr. Nicholas Peraino and Mr. Dennis Anderson in

Lumigen Instrument Center of Wayne State University for their expertise training and patient instruction in the HRESIMS and NMR, respectively.

My great gratitude is also devoted to my dearest family: dad Xiaoliu Yang, mom Qiaoxiang

Wang, wife Yueyuan Chen and daughter Erin Enxin Yang. I could not have made it without their warm company and selfless love.

Finally, I am infinitely thankful to all of you who have contributed to my achievement of this project by any mean.

TABLE OF CONTENTS

DEDICATION ...... ii

ACKNOWLEDGMENTS ...... iii

LIST OF TABLES ...... viii

LIST OF FIGURES ...... ix

LIST OF ABBREVIATIONS ...... xi

CHAPTER 1 Introduction ...... 1

1.1. Irinotecan and Delayed Diarrhea ...... 1

1.2. Bacterial β-Glucuronidases and Inhibitors ...... 4

1.3. Specific Aims ...... 7

CHAPTER 2 Isolation of Bioactive Compounds from Noni Fruits ...... 9

2.1. Introduction of Noni (Morinda citrifolia) ...... 10

2.1.1. Nutritional content and chemical constituents of noni ...... 11

2.1.2. Biological activities and safety of noni ...... 12

2.2. Isolation and Purification ...... 14

2.2.1. Equipment and materials ...... 15

2.2.2. Isolation procedures ...... 16

CHAPTER 3 Elucidation of Chemical Structures ...... 19

3.1. Spectroscopic Data Acquisition ...... 22

3.2. ECD Computation ...... 25

3.2.1. Computational chemistry and its application in CD calculation ...... 25

3.2.2. ECD calculation of compounds 1-4 ...... 27

3.3. Structural Elucidation ...... 31

3.3.1. (7S,8S,7'R,8'R)-Isoamericanol B ...... 32

3.3.2. Americanol B ...... 34

3.3.3. Moricitrin A ...... 37

3.3.4. Moricitrin B ...... 39

CHAPTER 4 Evaluation of Inhibitory Activities ...... 43

4.1. Inhibitory Assays on Enzymes ...... 43

4.1.1. Enzymes and Reagents ...... 43

4.1.2. Bacterial β-Glucuronidase Assay ...... 44

4.1.3. α-Amylase Assay ...... 44

4.1.4. α-Glucosidase Assay ...... 45

4.1.5. Pancreatic Lipase Assay ...... 45

4.1.6. Data Analysis ...... 45

4.2. Specific Inhibition Against Bacterial β-Glucuronidase ...... 45

4.3. Inhibition on Digestive Enzymes ...... 47

CHAPTER 5 Discussion ...... 49

5.1. Bioassay-guided Isolation ...... 49

5.2. Structure Elucidations ...... 50

5.3. In vitro Study ...... 51

5.4. In vivo Study ...... 53

APPENDIX ...... 55

REFERENCES ...... 86

ABSTRACT ...... 103

vii

LIST OF TABLES

Table 1. Common Terminology Criteria for Adverse Events for Diarrhea...... 4

Table 2. Classification of chromatography used for the isolation of natural products ...... 15

1 13 Table 3. H and C NMR Spectroscopic Data of Compounds 1-4 in CD3OD ...... 24

Table 4. Comparison between ECD and VCD ...... 28

Table 5. Possible enantiomers of compounds 1-4 ...... 28

Table 6. Criteria of Conformer Selection for DFT Optimization and TDDFT Calculation of Model Configurations of Compounds 1-4 ...... 30

Table 7. Inhibitory Activities against Bacterial β-Glucuronidase ...... 46

viii

LIST OF FIGURES

Figure 1. The trees, flowers and fruits of Camptotheca acuminata ...... 2

Figure 2. The chemical structures of camptothecin (CPT, left) and irinotecan (CPT-11, right) ...... 3

Figure 3. The mechanism of irinotecan-induced delayed diarrhea involving the production of intestinal SN-38 ...... 6

Figure 4. Fruits and flowers of Noni (Morinda citrifolia) ...... 10

Figure 5. Noni (Morinda citrifolia) fruit powder used in this study ...... 16

Figure 6. The isolation process of bioactive compounds from Noni fruit powder ...... 17

Figure 7. Semi-preparative HPLC chromatogram of fraction F-3-2-2 ...... 18

Figure 8. Description of absolute configurations for a pair of enantiomers, A, B, C, D represent four functional groups with priority from high to low...... 20

Figure 9. Determination of absolute configuration by ECD calculation ...... 29

Figure 10. Experimental (solid lines) and calculated (dashed lines) UV spectra of compounds 1-4 and their model structures ...... 30

Figure 11. Chemical structures of compounds 1-4. a, a', b and b' represent model configurations ...... 31

Figure 12. Key COSY (bold lines) and HMBC (blue arrows) correlations of compound 1 ...... 32

Figure 13. Distances between protons of steric centers for compounds 1 (energy minimized by MM2 force field) ...... 33

Figure 14. Experimental and calculated ECD spectra of compound 1 in MeOH, a and b represent the model configurations of each compound ...... 34

Figure 15. Key COSY (bold lines) and HMBC (blue arrows) correlations of compound 2 ...... 35

Figure 16. Distances between protons of steric centers for compounds 2 (energy minimized by MM2 force field) ...... 36

Figure 17. Experimental and calculated ECD spectra of compound 2 in MeOH, a and b represent the model configurations of each compound ...... 36

Figure 18. Key COSY (bold lines) and HMBC (blue arrows) correlations of compound 3 ...... 38

Figure 19. Distances between protons of steric centers for compounds 3 (energy minimized by MM2 force field) ...... 38

Figure 20. Experimental and calculated ECD spectra of compound 3 in MeOH, a and b represent the model configurations of each compound ...... 39

Figure 21. Key COSY (bold lines) and HMBC (blue arrows) correlations of compound 4 ...... 40

Figure 22. Distances between protons of steric centers for compounds 4 (energy minimized by MM2 force field) ...... 41

Figure 23. Experimental and calculated ECD spectra of compound 4 in MeOH, a and b represent the model configurations of each compound ...... 42

Figure 24. Inhibitory effects of compounds 1-4 against α-amylase (A), α-glucosidase (B) and pancreatic lipase (C)...... 48

Figure 25. Balb/cJ mouse ...... 53

LIST OF ABBREVIATIONS

CD Circular Dichroism

CD3OD Deuterated methanol for NMR measurement

CH2Cl2 Dichloromethane

CH3CN Acetonitrile

COSY Correlation Spectroscopy

DFT Density Functional Theory

ECD Electronic Circular Dichroism

HPLC High Performance Liquid Chromatography

H2O Water (ultrapure quality for chromatography)

HMBC Heteronuclear Multiple Bond Correlation

HSQC Heteronuclear Single Quantum Correlation

HRESIMS High Resolution Electrospray Ionization Mass Spectrometry

LC-MS Liquid Chromatography coupled to a Mass Spectrometer

MeOH Methanol m/z Mass-to-Charge ratio

NMR Nuclear Magnetic Resonance

NOESY Nuclear Overhauser Effect Spectroscopy

PCM Polarizable Continuum Model

ROESY Rotating Frame Nuclear Overhauser Effect Spectroscopy

TDDFT Time-dependent Density Functional Theory

UV Ultraviolet

VCD Vibrational Circular Dichroism

xii

CHAPTER 1 Introduction

Irinotecan was approved by the FDA as a second-line therapy for metastatic colon or rectal cancer in 1996. It is derived from the plant alkaloid camptothecin (CPT) and specifically inhibits eukaryotic DNA enzyme topoisomerase I. However, one of the leading side effects of irinotecan is severe diarrhea. After oral administration, irinotecan is metabolized into SN-38G in the liver, then SN-38G is excreted to the intestinal tract and deconjugated back to SN-38 by intestinal bacterial β-glucuronidase. The free SN-38 in the gut leads to the delayed diarrhea by damaging the intestinal mucosa. Therefore, selectively inhibiting the bacterial β-glucuronidase has become a promising strategy to alleviate irinotecan-induced delayed diarrhea. The purpose of this project is to explore more potent and higher selective inhibitors of bacterial β-glucuronidase from noni fruits, aiming to alleviate the irinotecan-induced delayed diarrhea with minimized side effects.

1.1. Irinotecan and Delayed Diarrhea

Camptothecin (CPT) was firstly discovered by Dr. Monroe E. Wall and Dr. Mansukh C.

Wani in the early 1960s (1). It is from Camptotheca acuminata Decne (family Nyssaceae), a tree native to China (Figure 1). Earlier to the late 1950s, the Eastern Regional Research Laboratory

(ERRL) of USDA screened approximate 1000 ethanolic plant extracts for antitumor activities. The

Camptotheca extract was shown high antitumor activity by the CA-755 assay. Then, Dr. Wall and

Dr. Wani conducted antitumor assay-guided isolation on the extracts of the tree’s wood and bark.

The obtained monoterpene indole alkaloid (MIA), CPT (Figure 2), showed remarkable activity in the life prolongation of mice treated with L1210 leukemia cells, with doses between 0.5 and 4.0 mg/kg. Based on this cytotoxic structure, a number of analogs were developed to fight off a variety

of solid tumors, such as water insoluble 9-nitro- and 9-amino-CPT, and water-soluble analogs of

10-hydroxy-CPT that includes CPT-11 (Figure 2) and topotecan.

Figure 1. The trees, flowers and fruits of Camptotheca acuminata

By 1985, it was discovered that CPT and its analogues can specifically inhibit the eukaryotic DNA enzyme topoisomerase I, which produces transient single-stranded DNA breaks but doesn’t require ATP (2, 3). The levels of topoisomerase I are dramatically elevated in some tumor cells, and also relatively high in quiescent and proliferating cells (4, 5). During the early S- phase of cell cycle, a cleavable complex that links topoisomerase and DNA is formed to unwind the DNA double helix prior to the replication fork. CPT and its analogues act to irreversibly arrest the replication fork by breaking the cleavable complex, which subsequently ceases DNA replication and results in cell death (6, 7). The findings of this unique anticancer mechanism greatly boosted the clinical trials of CPT and its analogues. Those compounds involved in a broad spectrum of cancers, including leukemia, lymphoma, small and non-small cell lung cancer,

colorectal cancer and cervical cancer (7). Among them, CPT-11, also called irinotecan, was successfully developed as an anticancer drug. It was originally developed by Japanese pharmaceutical companies Yakult Honsha and Daiichi, and firstly approved in Japan in 1994, then in France in 1995. It was approved by the FDA of United States in 1996, as a second-line therapy for metastatic colon or rectal cancer (1, 8).

Figure 2. The chemical structures of camptothecin (CPT, left) and irinotecan (CPT-11, right)

Up to date, irinotecan has been widely used to treat colorectal, lung and pancreatic cancers

(9). However, like other chemotherapeutic agents, such as fluorouracil and capecitabine of fluoropyrimidines, one of the leading side effects of irinotecan is diarrhea, which is also called chemotherapy-induced diarrhea (CID) (10). It has been reported that up to 88% of patients treated irinotecan suffer from diarrhea, and 31% of these cases are grade 3 or 4 (Table 1) (11, 12).

Table 1. Common Terminology Criteria for Adverse Events for Diarrhea

Grade Diarrhea Criteria 1 Increase of < 4 stools per day over baseline; mild increase in ostomy output compared with baseline 2 Increase of 4-6 stools per day over baseline; moderate increase in ostomy output compared with baseline 3 Increase of ≥7 stools per day over baseline; incontinence; hospitalization indicated; severe increase in ostomy output compared with baseline; limiting self-care activities of daily life 4 Life-threatening consequences; urgent intervention indicated 5 Death

Irinotecan may induce acute and delayed diarrhea (13). The acute diarrhea occurs immediately after administration due to acute cholinergic properties. It is does-dependent and could be rapidly and effectively treated by atropine (14). The delayed diarrhea happens more than

24 hours after drug administration. It is always a serious and dose-limiting adverse effect (14, 15).

In the liver, irinotecan is metabolized by carboxylesterases (CE) to produce SN-38, which exerts

100 to 1000-fold more potent cytotoxicity than irinotecan (16). Then, SN-38 is conjugated to the inactive and non-toxic SN-38-glucuronide (SN-38G) via uridine diphosphate- glucuronosyltransferase 1A1 (UGT1A1). SN-38G is further excreted to the intestinal tract through the bile duct. In the gut, SN-38G is deconjugated back to SN-38 by bacterial β-glucuronidase

(Figure 3). Even though the exact mechanisms are still in debate (13), this free SN-38 in the gut can damage the intestinal mucosa, which is responsible for the delayed diarrhea (17, 18).

1.2. Bacterial β-Glucuronidases and Inhibitors

The treatments of irinotecan-induced delayed diarrhea are still challenging. The National

Cancer Institute Common Toxicity Criteria are worldwide recognized guidelines for the

assessment and management of chemotherapy-induced diarrhea, which were published in 2004 and updated in 2014 (19, 20). According to the guidelines, the delayed diarrhea could be managed through dietary modification and using of antidiarrheal medicines based on specific mechanisms, such as somatostatin analog octreotide, loperamide and deodorized tincture of opium. However, these therapies may either worsen existing gastrointestinal symptoms, or cause other adverse effects like uneven heartbeat, respiratory depression, neurotoxicity and seizures (21). Therefore, more effective therapies for the delayed diarrhea are still urgently needed, and emerging new methods have been investigated, including herbal extracts, phytochemicals and probiotics (12, 21).

β-Glucuronidases are members of glycosyl hydrolases that involve in the breakdown of complex carbohydrates (22). Human β-glucuronidases exist in the lysosomes of most tissues and body fluids (23). In the gut, brush border β-glucuronidases are mainly secreted by the bacteria of

Enterobacteriaceae family, including Salmonella, Klebsiella, Yersinia, Shigella, and especially

Escherichia coli. Gut bacterial β-glucuronidases assist the microbiota in harnessing available glycosyl units as carbon sources, they are also the key mediators of intestinal toxicities associated with anticancer drugs, including irinotecan (24).

Figure 3. The mechanism of irinotecan-induced delayed diarrhea involving the production of intestinal SN-38 As discussed above, the bacterial β-glucuronidase is essential for the generation of intestinal SN-38. Alternative strategies targeting this enzyme have been paid increasing attention along with the clinical application of irinotecan. They are aiming to alleviate the delayed diarrhea by reducing the generation or inhibiting the activity of bacterial β-glucuronidase (Figure 3). Broad- spectrum antibiotics have been used to reduce the production of bacterial β-glucuronidase by killing intestinal bacteria, such as penicillin and streptomycin (25, 26), neomycin (26), and the combination of cholestyramine and levofloxacin (27). However, these methods remove or modulate the gut bacteria, which may disorder the normal microbial metabolism of carbohydrate, vitamin and bile acids (28, 29), or increase the risks of pathogenic infection, such as Clostridium difficile and Escherichia coli (30-32). As a result, the using of antibiotics may deteriorate the poor health of patients who are suffering from the other side effects of chemotherapy (33).

It has initiated another active strategy to develop selective inhibitors against the enzyme bacterial β-glucuronidase. Saccharo-1,4-lactone was firstly reported in 1952 for its competitive

inhibition against β-glucuronidase from mouse liver (34, 35). In 2004, saccharic acid 1.4-lactone was proven in animal experiment to be able to reduce mucosal damage induced by irinotecan, however, exhibited a relatively low potency (36). The Kampo medicine (Hangeshashin-to) TJ-14 was also investigated for irinotecan-induced diarrhea (37, 38). Its protective effects were tentatively attributed to the glucuronides, such as baicalin, wogonoside, luteolin-3′-glucuronide and glycyrrhizin, due to their potential inhibition against β-glucuronidase (39, 40). But the exact mechanisms of the Kampo extracts remain unclear. Since 2010, a group from the University of

North Carolina has published a series of research papers in this area. For the first time, they reported inhibitors against β-glucuronidase from intestinal bacteria, instead of mammalian host.

Those compounds selectively inhibited bacterial β-glucuronidase with IC50 in the range of 0.28-

4.8 µM (33, 41, 42). More importantly, they could reduce the irinotecan-induced gastrointestinal damage, while without killing intestinal commensal bacteria and not toxic to the host intestinal cells (11). Cheng et al. also reported pyrazolo[4,3-c] quinoline derivatives as specific inhibitors of bacterial β-glucuronidase, without impacting the antitumor efficacy of irinotecan in mice (43, 44).

These inhibitors of bacterial β-glucuronidase have not achieved very potent and specific effects, however, they represent a very promising orientation to develop supplemental or therapeutic approaches to alleviate irinotecan-induced delayed diarrhea with minimized side effects.

1.3. Specific Aims

In a preliminary experiment, we screened about 50 extracts for their inhibitory activities against bacterial β-glucuronidase. Among them, an extract from noni fruits showed significant inhibition on the enzyme. The purpose of this study is to further investigate the chemical constituents of noni fruit extract and find out more potent and higher selective inhibitors against

intestinal bacterial β-glucuronidase, aiming to alleviate the delayed diarrhea induced by irinotecan administration, while with minimized side effects on the gastrointestinal tract. Therefore, the specific aims of this project are as follows:

Aim 1: Extraction and Isolation (Chapter 2). To identify the chemical constituents of noni fruit extract contributing to its inhibitory effect against bacterial β-glucuronidase, through phytochemical extraction and isolation guided by bioactive assay.

Aim 2: Elucidation of Chemical Structures (Chapter 3). To elucidate chemical structures of purified compounds using HRMS and 1D, 2D-NMR, and establish their absolute configurations by comparing experimental and calculated ECD spectra.

Aim 3: Evaluation of Inhibitory Activities (Chapter 4). To measure inhibitory activities of purified compounds against bacterial β-glucuronidase and digestive enzymes, evaluate their inhibitory selectivity and intestinal adverse effects.

CHAPTER 2 Isolation of Bioactive Compounds from Noni Fruits

Natural products are chemical substances or compounds isolated from living organism, including plants, animals, fungi and microorganisms (45, 46). They are produced by the primary or secondary metabolisms of organism. Primary metabolites are essential organic molecules for the growth, development and reproduction of organism, such as nucleic acids, amino acids, fatty acids and sugars. They are necessary building blocks to make life sustaining macromolecules including DNA, proteins, lipids and carbohydrates. Secondary metabolites are not directly involved in these essential processes, but they have ecological function that increase the competitiveness of the organism to deal with very complicated and varied environment. They are much more organism-dependent, and present great diversities in chemical structures and biological functions. Such as flavonoids, phenolics, tannins and alkaloids, they have lots of structural subtypes and are among the most bioactive compounds (47). Therefore, natural products of secondary metabolites have been widely studied and developed for the application in food, spices, traditional and modern medicines (48, 49). Like antibacterial drug penicillins from Penicillium chrysogenum, opioid analgesic drug morphine from Papaver somniferum, and artemisinin

(qinghaosu) against Plasmodium falciparum malaria, natural products contribute greatly to the history and landscape of new molecular entities (NMEs) (50). By the end of 2013, over one-third

(38%) of NMEs approved by U.S. Food and Drug Administration (FDA) are natural products or their derivatives (51).

In our preliminary experiment, the bioactive extract used was also a mixture of secondary metabolites from noni fruits. Therefore, the first priority of this project was to isolate the mixture and obtain the purified compounds that contribute to the inhibitory activity against bacterial β- glucuronidase.

2.1. Introduction of Noni (Morinda citrifolia)

Morinda citrifolia L. (Rubiaceae) (Figure 4) is a small evergreen tree that is originated in

South Asia, but now is widely distributed across the tropical countries of the world, including

Polynesia, Northeastern Australia and the Caribbean (52). The most popular name “noni” is from

Hawaiian and Tahitian islands (53), whereas it is called Indian mulberry in India, mengkudu in

Malaysians and painkiller bush in the Caribbean (54), and the fruits are also known as cheese fruit in Australia (55).

Figure 4. Fruits and flowers of Noni (Morinda citrifolia)

All parts of noni have been traditionally used for medicinal purposes, including the roots, bark, leaves, seeds and fruits (52). Modern research on noni might be incentivized by the report from the Pacific Tropical Garden Bulletin in 1985, which suggested a number of potential health benefits for noni juice, such as treating high blood pressure, menstrual cramps, arthritis and many others (55, 56). Since the 1990s, noni has become a popular dietary supplement in the USA (57).

The leaves and fruits could be sold as capsules, teas and juice. In particular, the noni fruit juice is

the predominant form in the market, and it could be pasteurized, fermented, or flavored by other fruit juice to make the product more palatable (58). The noni fruit juice was approved by the

European commission in 2003 as well, and the use of noni has been broadly available in stores and on the Internet (52).

2.1.1. Nutritional content and chemical constituents of noni

Nutritional content of M. citrifolia have been evaluated for health utilization. The leaves were considered as an abundant source of carotenoids by Aalbersberg et al. (59). Peerzada et al. found that noni fruits from Australia contain up to 2012 mg of potassium and 158 mg of vitamin

C per 100 g of dry weight (60). The noni fruit juice was also shown to possess relatively high levels of potassium (30-150 ppm) and vitamin C (30-155 mg/kg) (61). Bui et al. analyzed the saccharide content of noni fruits collected in Vietnam, it was found that the monosaccharides were mostly galactose, arabinose, rhamnose and galacturonic acid, while polysaccharides were mainly pectic polysaccharides, such as arabinan, pectins homogalacturonan, and type Ⅰ and Ⅱ arabinogalactans (62).

The earliest research on chemical constituents of M. citrifolia could be traced back to the investigation on noni roots, heartwood and cell culture, which were reported by Simonsen et al. in

1918 and 1920 (63, 64). After 50 years of silence, the phytochemical studies of these parts of noni, and also seeds and flowers, emerged again during 1970s and 1980s (52). Since 1990s, the phytochemical investigation on noni have been focused on the leaves and fruits (52). Generally, the roots were found predominated by anthraquinones and anthraquinone glycosides (65), while some iridoids, triterpenes and flavonoids were isolated from leaves (66). The plant cell cultures were studied for their abilities to synthesize anthraquinoid pigments, and the focus on fruits came

along with the commercial application of fruit juice as nutritional supplement (55). Up to date, more than 200 chemical constituents have been reported from different parts of M. citrifolia, including fatty acid glycosides, alcohol glycosides, iridoids and iridoid glycosides, lignans, neolignans, anthraquinones and anthraquinone glycosides, flavonoids, phenylpropanoids, triterpenoids and others (52, 55, 67). Their chemical structures were primarily identified by nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry (52).

2.1.2. Biological activities and safety of noni

The biological activities of both extracts and pure compounds of M. citrifolia have been comprehensively evaluated in vitro and in vivo. The investigated parts include the leaves, roots, seeds, and particularly the fruits (52, 55). Their pharmacological activities mainly involve in antimicrobial (68), anti-fungus (69), anti-virus (70), anti-oxidation (71), anti-inflammation (72), anti- obesity (73), anti-diabetes (74), anti-arthritic (75), anti-tubercular (76), anti-pigmentation (77), analgesic (78), and so on. Especially, the anticancer effects of M. citrifolia have attracted relative more attention (79). The juice or extracts from noni fruits have been shown antiproliferative effects against various tumor cells (80), reducing mammary tumor growth in mice (81), inducing apoptosis in human cervical cancer cells and on the Ehrlich ascites tumor in Balb‐c mice (82, 83).

Pure compounds isolated from M. citrifolia proved to contribute to the overall anti-tumor activities of noni extracts. Masuda et al. revealed that the two lignans, 3,3'-bisdemethylpinoresinol and americanin A, were active constituents of noni seeds to inhibit melanogenesis in murine B16 melanoma cells (84). Two glycosides from noni fruit juice could inhibit AP-1 transactivation and cell transformation in the mouse epidermal JB6 cell line (85). The anthraquinones from noni roots showed significant inhibition against human lung and colon cancer cells (86, 87). An

polysaccharide‐rich substance from noni fruit juice was also found having antitumor activity (88,

89).

It is worth noting that several clinical trials have been conducted for M. citrifolia (79). In a randomized, double blinded and placebo-controlled trial, Prapaitrakool et al. showed that an extract from M. citrifolia Linn. fruits effectively reduced the incidence of early postoperative nausea (0-6 h) from 80% to 48% (P = 0.04), at the dose level of 600 mg and taken orally 1 hour before surgery (90). In a Phase Ⅰ clinical trial aiming to determine noni dose for patients with advanced cancer, Issell et al. suggested that a dose of four capsules four times per day (8 g) could be appropriate for Phase Ⅱ trial in terms of quality of life measures (91). Noni has been used to treat women with primary dysmenorrhea, but show no reduction effects on menstrual pain or bleeding comparing to placebo (72). A combination of noni powder and anti-rheumatism therapy showed effect to treat a case of psoriasis (92), another noni supplements combined with calorie restriction and exercise interventions resulted in body weight loss and fat mass decrease for overweight participants (73). However, the biological effects and contribution of noni were not clarified in these combination studies, which need to be verified by further experiment.

A number of animal studies have indicated that noni administration are nontoxic to the rats

(93, 94). However, several cases of plausible toxicity have been reported for human who consumed noni fruit juice (52). The elevated potassium levels were reported for a patient with chronic renal insufficiency, but lacking detailed information on how much noni juice the patient took (95). Since recommended dose of 1-3 oz per day would not cause hyperkalaemia, it is necessary to confirmed whether the patient consumed a does more than recommendation (52). Another concern of noni juice consumption is hepatotoxicity. Several patients from Austria and Germany demonstrated

significant increasing in liver enzymes when taking noni juice, such as transaminases and lactate dehydrogenase, and the levels of these liver enzymes returned back to normal levels after interruption of noni juice intake (96-98). However, most of these patients were simultaneously taking other herbs or medicines, the observed hepatotoxicity could not be directly linked with noni juice without further evidences. A possible correlation between anthraquinones and hepatotoxicity were suggested, however, the content of this constitutes was relatively high in noni roots, while extremely low in noni fruits (55). Thus, it is not reasonable to deduce the toxicity of noni fruit juice by anthraquinones. On the contrast, West et al. listed dietary utilization of noni and summarized previous animal and human studies on noni, which gave a general conclusion that noni is safe for dietary supplementation (93).

2.2. Isolation and Purification

For the isolation of natural products, the normally used methods include extraction, precipitation, adsorption, and especially chromatography (46). Each chromatographic method consists of a stationary phase that doesn’t move and a mobile phase that does move (99, 100). The mobile phase can be organic solvents, water or gas, which carries the mixture of sample to go through the stationary phase. During the process, different compounds interact differently with the two phases, and move forward with different velocities, which result in different retention properties and chromatographic separation. Chromatographic methods could be divided into many subtypes in terms of separation mechanism, phase condition and shape of chromatographic bed

(Table 2). Of which, the thin layer chromatography (TLC) is often used to predict or evaluate separation effects, while column chromatography is primarily utilized to implement the practical isolation of natural products due to its multiple functions and flexible application (101). Normally

used methods of column chromatography include normal phase (silica gel), reverse phase (ODS,

C18), size exclusion (Sephadex LH-20) and so on. In order to obtain finer separation, the specific stationary materials, mobile phase solvents and column size should be carefully optimized based on the physical and chemical characteristics of target samples. It is worth noting that the semi- preparative HPLC system combines high-efficient separation, on-line detection and analysis, and even automatic sample collector together, which make it a widely used powerful technique for the separation of samples that are difficult to be isolated by normal methods (analogues) and compounds with small amounts (about 5 mg or less) (102).

Table 2. Classification of chromatography used for the isolation of natural products

Technique Format Stationary Phase Mobile Phase Separation Mechanism paper chromatography (PC) planar paper (cellulose) liquid partition (adsorption, ion-exchange, size- exclusion) thin-lay chromatography planar silica, cellulose, ion- liquid adsorption (partition, (TLC) exchange resin, ion-exchange, size- controlled porosity solid exclusion) liquid chromatography (LC): column solid or bonded phase, liquid modified partition open column, media- controlled porosity (adsorption), size- pressure, high-performance solid, ion-exchange exclusion, ion-change, liquid chromatography resin selective adsorption (HPLC) gas chromatography (GC): gas-liquid column liquid gas adsorption gas-solid column solid gas partition

2.2.1. Equipment and materials

In this study, semi-preparative HPLC was carried out on a Hitachi LaChrom Elite liquid chromatograph system (Tokyo, Japan) equipped with L-2130 pump and L-2455 detector, and a

Waters XBridge BEH300 Prep C18 column (250 mm × 19 mm i.d., 10 μm; Milford, MA, USA).

Diaion HP20 adsorbent resin (highly porous type, ≥ 250 μm; Alfa Aesar, Ward Hill, MA, USA),

silica gel (40-60 μm, 60 A; Acros Organics, Morris Plains, NJ, USA), and Sephadex LH-20 (GE

Healthcare, Uppsala, Sweden) were used for column chromatography fractionation.

Noni fruit powder was purchased from Starwest Botanicals, Inc. (Sacramento, CA, USA) with lot numbers 58868 and 63285 (Figure 5). It was collected from India and identified as

Morinda citrifolia L. (Rubiaceae). The packages were stored in dark at 4 °C before use. A voucher specimen (BS2015-029) was deposited in our laboratory, Department of Nutrition and Food

Science, Wayne State University.

Figure 5. Noni (Morinda citrifolia) fruit powder used in this study

2.2.2. Isolation procedures

The isolation of active compounds from noni fruit extract was guided by inhibitory assay of bacterial β-glucuronidase (Figure 6). After each step of separation, only the bioactive fractions were selected for further purification.

Figure 6. The isolation process of bioactive compounds from Noni fruit powder

Specifically, noni fruit powder (13.2 kg) was extracted with 50% acetone/water twice (at ratio of 10 L/kg and 8 L/kg, respectively) at room temperature with stirring. The combined extracts were concentrated until acetone was completely removed, then the water phase was subjected to separation over a HP-20 resin column (11 × 70 cm), and eluted with 0%, 30%, 50%, 70% and 100% ethanol/water, sequentially. After being concentrated, 29 g of 70% ethanol/water extract (active fraction) was further fractionated via silica gel column chromatography (7 × 46 cm), and eluted

with CH2Cl2-MeOH (1:0, 20:1, 10:1, 6:1, 4:1, 1:1 and 0:1, v/v), to yield five sub-fractions (F-1 to

5). Active fraction F-3 (10 g) was then separated on a silica gel containing column with CH2Cl2-

MeOH (20:1, 10:1, 5:1, and 2:1, v/v) to obtain four sub-fractions. Active subfraction F-3-2 (2.7 g) was further purified by a Sephadex LH-20 column chromatography (CH2Cl2-MeOH, 1:1, v/v) to afford two main sub-fractions. Active fraction F-3-2-2 (800 mg) was purified by semi-preparative

HPLC (CH3CN-H2O, 1:20, v/v; 4 mL/min; detector wavelength at 230 nm), to obtain four purified active compounds, 1 (31.93 min, 12 mg), 2 (48.19 min, 11 mg), 3 (69.50 min, 33 mg), and 4 (83.42 min, 87 mg) (Figure 7).

Figure 7. Semi-preparative HPLC chromatogram of fraction F-3-2-2

CHAPTER 3 Elucidation of Chemical Structures

Chemical structure refers to the spatial arrangement of atoms in a molecule and the chemical bonds that connect atoms together. Carbon and hydrogen are the elementary atoms of organic compounds or natural products. Most of natural products also contain oxygen, and some may include nitrogen, phosphorous, sulfur or the halogens. These atoms are mainly linked by covalence bonds that share electron pairs in the forms of single, double or triple bonds. In the history, the chemical structures of natural products could not be described until the advent of structural theory based on the frameworks of Kekulé (1857), Couper (1858) and Butlerov (1859), and the foundation of stereochemistry by Le Bel (1874) and van’t Hoff (1875) (103).

Early procedure of structural elucidation is very complicated (46). It was often started with the measurement of sharp melting point to confirm the purity, and elemental analysis to decide molecular formula. It was followed by a combination of derivatization and degradation reactions to deduce functional groups and fragments. Eventually, the proposed structure needs to be validated by known chemical synthesis. Consequently, a large amount of sample (grams) was required for destructive reactions, and it always took several or even tens of years to figure out a specific structure. For example, the structures of morphine and strychnine were determined 118 and 127 years after they were firstly isolated, respectively (104, 105). Until 1960s, structural elucidation was developed into a new era that is dominated by spectroscopic techniques and X-ray crystallography (103). Up to date, frequently used spectroscopic methods include nuclear magnetic resonance (NMR), which contains 1H-, 13C-, and 2D-NMR, mass spectrometry (MS), ultraviolet

(UV) and infrared (IR). In addition, a lot of new methods have been developed and applied in this area, such as 3D-NMR, solid-state NMR and computer-assisted structure elucidation (106, 107).

Combination of these techniques greatly simplified elucidation process. The time used could be

reduced to several days or even minutes, sample amount required could be several milligrams, and it is much easier to solve the structures of complicated natural products (108).

Absolute configuration is a concept of stereochemistry, it refers to the spatial arrangement of the atoms of a chiral molecular entity, and could be described in R (Rectus) or S (Sinister)

(Figure 8) (109). Theoretically, if a chiral molecule contains number n (≥ 1) of chiral center atom(s)

(usually carbon), it can exist in any of the 2n of absolute configurations. In other words, it has 2n-1 pair(s) of possible enantiomers, and each pair of enantiomers has mirror-image structures.

However, as natural products of living organisms, all discovered primary metabolites (L-amino acids, nucleotides and monosaccharides) and about 80% of known secondary metabolites are chiral non-racemic or asymmetric enantiopure molecules (110). Besides, it is found that two enantiomers of a chiral molecule always possess totally different biological activities and pharmacological behaviors, which may due to their biological receptors of organisms that are already chiral in structures (111). Among the new molecular entities approved by the FDA, chiral drugs have increased from 30-40% in the 1990s to over 60% of in the 2000s (112). Therefore, determining the absolute configuration of a chiral molecule has become an important, but also challenging area in the chemistry of natural products (110, 113).

Figure 8. Description of absolute configurations for a pair of enantiomers, A, B, C, D represent four functional groups with priority from high to low.

Single crystal X-ray diffraction analysis has been considered the most accurate and important method to determine absolute configuration. However, the measurement requires high quality of crystal, which limits its application for many natural products that are difficult to culture crystal. The modified Mosher’s NMR method is another powerful tool to decide absolute configuration. But it is not easy to find suitable functional groups (normally OH and NH2) in many molecules to conduct chiral derivatization. Chiroptical methods are based on the optical interactions between chiral non-racemic compounds and the right and left circularly polarized light.

Currently, the primary chiroptical spectroscopy includes optical rotation (OR), optical rotatory dispersion (ORD), electronic circular dichroism (ECD) and vibrational circular dichroism (VCD).

Especially, ECD and VCD have been more frequently used for determining absolute configurations of natural products, due to their less demand on structural properties, sample amount and status (114). In practical use, absolute configuration of a specific compound is established by comparing its experimental circular dichroism (CD) spectrum to the CD spectra of appropriate analogues with known absolute configuration. However, it is hard to find comparable analogues for most of novel natural products. In this case, calculated CD spectra have become a credible alternative along with the development of computational chemistry (115).

In this study, the structures of four purified compounds were elucidated through the combination of HRESIMS, MNR and UV. Their absolute configurations were established with the assistance of ECD calculation, because of limited amounts, difficulty of culturing single crystal and lack of comparable known CD spectra.

3.1. Spectroscopic Data Acquisition

The optical rotations were measured in MeOH on an Autopol Ⅲ automatic polarimeter

(Rudolph Research Analytical, Hackettstown, NJ, USA) with a sodium lamp operating at 589 nm.

ECD spectra in solvent methanol were recorded on a JASCO J-815 spectrometer (Tokyo, Japan).

NMR spectra were obtained from Varian VNMRS-500 MHz (Palo Alto, CA, USA) and Agilent

DD2-600 MHz (for optimized HMBC spectra; Santa Clara, CA, USA) spectrometers, with δ in ppm related to TMS, and J in Hz. HRESIMS analyses were performed on Thermo Scientific LTQ

Orbitrap XL Mass Spectrometer (San Jose, CA, USA).

22 (7S,8S,7'R,8'R)-Isoamericanol B (1). Brown amorphous powder; [α] D - 8.0 (c 0.05, MeOH);

UV (MeOH) λmax (log ε) 206 (5.01), 230 (sh) (4.52), 270 (4.12), 287 (sh) (4.05), 305 (sh) (3.57),

315 (sh) (3.36) nm; ECD (MeOH) λmax (Δε) 205 (+ 2.58), 211 (- 7.49), 218 (+ 0.36), 226 (- 1.31),

1 13 241 (+ 0.14), 288 (- 1.50) nm; H (500 MHz) and C (125 MHz) NMR (CD3OD) data, see Table

- 3; HRESIMS m/z 493.1491 [M - H] (calcd for C27H25O9, 493.1493, Figure A1).

22 Americanol B (2). Brown amorphous powder; [α]D + 2.0 (c 0.05, MeOH); UV (MeOH)

λmax (log ε) 206 (5.00), 230 (sh) (4.54), 270 (4.16), 287 (sh) (4.10), 305 (sh) (3.65), 315 (sh) (3.43) nm; ECD (MeOH) λmax (Δε) 198 (+ 1.53), 206 (- 14.7), 219 (- 0.29), 225 (- 1.51), 245 (- 0.09), 288

1 13 (- 1.96) nm; H (500 MHz) and C (125 MHz) NMR (CD3OD) data, see Table 3; HRESIMS m/z

- 493.1486 [M - H] (calcd for C27H25O9, 493.1493, Figure A2).

22 Moricitrin A (3). Brown amorphous powder; [α] D + 2.0 (c 0.05, MeOH); UV (MeOH) λmax

(log ε) 206 (5.06), 230 (sh) (4.40), 283 (4.00), 310 (sh) (2.97) nm; ECD (MeOH) λmax (Δε) 195 (+

8.69), 200 (- 11.57), 206 (- 5.06), 209 (- 5.80), 232 (+ 0.34), 247 (- 2.07), 257 (- 0.67), 280 (- 2.58)

1 13 nm; H (500 MHz) and C (125 MHz) NMR (CD3OD) data, see Table 3; HRESIMS m/z 657.1952

- [M - H] (calcd for C36H33O12, 657.1967, Figure A3).

22 Moricitrin B (4). Brown amorphous powder; [α] D - 8.0 (c 0.05, MeOH); UV (MeOH) λmax

(log ε) 206 (5.06), 230 (sh) (4.42), 283 (4.03), 310 (sh) (3.08) nm; ECD (MeOH) λmax (Δε) 198 (+

10.64), 210 (- 6.92), 221 (- 0.61), 229 (- 1.30), 233 (- 0.17), 239 (- 1.42), 245 (- 0.53), 283 (- 2.47)

1 13 nm; H (500 MHz) and C (125 MHz) NMR (CD3OD) data, see Table 3; HRESIMS m/z 657.1954

- [M - H] (calcd for C36H33O12, 657.1967, Figure A4).

1 13 Table 3. H and C NMR Spectroscopic Data of Compounds 1-4 in CD3OD

1 2 3 4 position δC, type δH (J in Hz) δC, type δH (J in Hz) δC, type δH (J in Hz) δC, type δH (J in Hz) 1 129.4, C 129.4, C 135.5, C 135.6, C 2 115.5, CH 6.87, d (2.3) 115.5, CH 6.87, d (2.1) 115.8, CH 6.89, m 115.8, CH 6.96, m 3 146.6, C 146.6, C 145.2, C 144.4, C 4 147.1, C 147.2, C 144.4, C 144.8, C 5 116.3, CH 6.81, m 116.4, CH 6.82, m 117.9, CH 6.91, d (8.3) 118.0, CH 6.92, d (8.3) 6 120.4, CH 6.78, m 120.4, CH 6.77, m 120.3, CH 6.85, m 120.1, CH 6.83, m 7 77.6, CH 4.84, d (8.2) 77.6, CH 4.83, d (8.1) 87.0, CH 4.66, brs 87.0, CH 4.67, brs 8 79.9, CH 4.03, m 79.8, CH 4.01, m 55.3, CH 3.04, brs 55.4, CH 3.06, brs 9 62.0, CH2 3.49, m 62.1, CH2 3.50, m 72.7, CH2 3.79, m 72.7, CH2 3.81, m 3.70, m 3.70, m 4.18, m 4.19, m

1' 131.3, C 131.3, C 135.5, C 135.6, C 2' 117.2, CH 7.01, m 117.2, CH 7.06, m 115.8, CH 6.89, m 115.8, CH 6.96, m 3' 145.4, C 145.8, C 145.2, C 144.4, C 4' 145.4, C 145.0, C 144.4, C 144.8, C 5' 118.1, CH 7.02, m 118.1, CH 6.93, m 117.9, CH 6.91, d (8.3) 118.0, CH 6.92, d (8.3) 6' 121.8, CH 6.97, m 121.6, CH 6.93, m 120.3, CH 6.85, m 120.1, CH 6.83, m 7' 77.3, CH 4.90, m 77.3, CH 4.91, m 87.0, CH 4.66, brs 87.0, CH 4.67, brs 8' 79.9, CH 4.03, m 79.8, CH 4.01, m 55.3, CH 3.04, brs 55.4, CH 3.06, brs 9' 62.0, CH2 3.50, m 62.1, CH2 3.50, m 72.7, CH2 3.79, m 72.7, CH2 3.81, m 3.70, m 3.70, m 4.18, m 4.19, m 1" 132.0, C 132.2, C 129.5, C 129.5, C 2" 115.5, CH 6.97, d (2.3) 115.5, CH 7.03, d (1.8) 115.6, CH 6.85, m 115.5, CH 6.84. m 3" 144.5, C 144.8, C 146.6, C 146.6, C 4" 145.1, C 144.8, C 147.1, C 147.1, C 5" 117.9, CH 6.90, m 118.0, CH 6.87, m 116.4, CH 6.80, d (8.0) 116.3, CH 6.79, m 6" 120.9, CH 6.92, m 120.8, CH 6.91, m 120.4, CH 6.74, dd 120.4, CH 6.74, m (8.0, 2.0) 7" 131.3, CH 6.49, d 131.3, CH 6.51, d 77.6, CH 4.77, d (8.0) 77.6, CH 4.79, d (7.9) (15.3) (16.0) 8" 128.2, CH 6.21, m 128.2, CH 6.22, dt 79.9, CH 3.97, m 79.9, CH 3.98, m (16.0, 5.8) 9" 63.7, CH2 4.19, m 63.8, CH2 4.20, m 62.1, CH2 3.45, dd 62.1, CH2 3.45, dd (12.2, 4.5) (12.2, 4.5) 3.65, d 3.65, d (12.2) (12.2) 1'" 129.5, C 129.5, C 2'" 115.6, CH 6.85, m 115.5, CH 6.84. m 3'" 146.6, C 146.6, C 4'" 147.1, C 147.1, C 5'" 116.4, CH 6.80, d (8.0) 116.3, CH 6.79, m 6'" 120.4, CH 6.74, dd 120.4, CH 6.74, m (8.0, 2.0) 7'" 77.6, CH 4.77, d (8.0) 77.6, CH 4.79, d (7.9) 8'" 79.9, CH 3.97, m 79.9, CH 3.98, m 9'" 62.1, CH2 3.45, dd 62.1, CH2 3.45, dd (12.2, 4.5) (12.2, 4.5) 3.65, d 3.65, d (12.2) (12.2)

3.2. ECD Computation

3.2.1. Computational chemistry and its application in CD calculation

Computational chemistry is founded on the development of theoretical chemistry and efficient computer programs (116). It can be used to calculate structures and properties of molecules, such as relative energy, dipole moment, vibrational frequency, charge distribution, reactivity and spectroscopic quantity. The aims of computational chemistry are to analyze complicated experimental data, predict new reactions or molecules, or predict spectroscopies.

There are two types of methods to model molecular systems according to starting point theory.

One type is classical method based on the laws of classical physics, which treat atoms as spheres and bonds as springs, while ignoring electronic distribution in molecules. Classical methods primarily contain molecular mechanics (MM) and molecular dynamics (MD). They are implemented through force fields that refer to functional forms and parameter sets of calculation.

Many force fields have been designed for different simulation purposes, such as AMBER and

CHSRMM for molecular dynamics of macromolecules, MM2 for conformational analysis, and

MMFF (Merck Molecular Force Field) for a broad range of molecules (117). Another type is quantum chemistry methods based on quantum mechanics (QM), which describe electronic distribution of molecules using Schrödinger’s equation. Complete solution of Schrödinger’s equation can provide an exact description of electronic structure of a molecule. However, up to data, the equation can only be exactly solved for hydrogen atom, but not for any other atomic or molecular systems involving motions of three or more particles. Alternatively, approximate solutions have been developed for practical application, which include ab initio methods (Hartree-

Fock, MP2), semi-empirical methods (AM1, MNDO, PM3), and density-functional theory (DFT) and its extension time-dependent DFT (TDDFT). A basis set is a set of functions used to carry out

the calculation of Hartree-Fock (HF) or DFT. Commonly used basis sets include B3LYP, 6-31G,

6-311G, cc-pVDZ and so on (117).

Molecular mechanics (MM) is restricted to describe equilibrium structures and conformations. It could be used to minimize energies, study molecular motions, and search stable conformations. While quantum mechanics additionally provides information of non-equilibrium forms, it is suitable to calculate molecular orbital energies, heat formation of a conformation, dipole moment and transition-state geometries and energies (116, 118). Quantum mechanics has more computing power than molecular mechanics, but it is more computational expensive (time) and only limited to hundreds of atoms. Selection of specific methods is dependent on calculation goals, and the trade-off between accuracy and computational expense.

The procedure of CD computation primarily consists of conformational analysis and DFT or TDDFT calculation (117). Because of rotation about single bonds, a compound presents in multiple conformers that are rapidly equilibrating. The population distributions of conformers are determined by their relative free energies, a conformer with lower free energy is more stable and has higher population. The interconversion rate between isomers are decided by energy barriers, the higher barrier means lower interconversion rate. CD spectrum is sensitive to conformation, it is a spectral combination of all conformers following Boltzmann distribution. Therefore, the first step is searching for stable conformers with lower free energies that contribute the most to the dynamic equilibrium, by molecular mechanics force fields. Obtained stable conformers could be optimized using less expensive basis sets of DFT calculation. Then, each stable conformer is calculated for CD spectrum. DFT methods is often used for VCD computation that involves ground electronic state. However, UV and ECD are properties of excited electronic state, which need

TDDFT method to calculate oscillator strength that simulate UV curve, and rotatory strength that predict ECD curve (115). Since CD spectrum of a compound is experimentally measured in the form of solution, the effects of solvent need to be taken into account for computation. Either explicit or implicit solvation models could be utilized to simulate solvent effects. The implicit solvation model treats solvent as a continuum medium, particularly the polarizable continuum model (PCM) has been most commonly used for CD calculation (117).

3.2.2. ECD calculation of compounds 1-4

ECD and VCD are different in measuring mechanisms, application requirements and calculation methods (Table 4) (119, 120). VCD spectrum doesn’t rely on the existence of chromophore groups in a molecule and has a relatively broader spectrum range, which shows a prospect of wide application. However, VCD is a relative young technique that the first commercial instrument was introduced in 1997 (121), and usable theories of spectrum interpretation have not been established. In addition, VCD measurement demand a relatively high sample concentration (20-50 mg/mL) (110). These are still restrictions for VCD application currently. While ECD has been used for more than five decades, there are plenty of practical accumulation, and many semi-empirical rules and exciton chirality methods have been developed to interpret ECD spectra (110). With the advantage of less requirement on sample amount, ECD is still a frequently used tool for determining absolute configurations.

Table 4. Comparison between ECD and VCD

ECD VCD Radiation ultraviolet and visible light infrared (IR) region Induced Transition electronic vibrational Sample Amount 1-10 µg/mL 20-50 mg/mL Chromophore Group needed not necessary Calculation excitation state TDDFT ground state DFT

Compounds 1-4 have relatively small amounts, but possess chromophore groups, which makes ECD computation a feasible choice to determine their absolute configurations. Since each compound has two pairs of possible enantiomers, and ECD spectra of paired enantiomers are theoretically symmetric, only one enantiomer of each pair was calculated using quantum- mechanical methods (Table 5).

Table 5. Possible enantiomers of compounds 1-4

Calculated enantiomer Paired enantiomer 1a (7S, 8S, 7'R, 8'R) 1a' (7R, 8R, 7'S, 8'S) 1b (7R, 8R, 7'R, 8'R) 1b' (7S, 8S, 7'S, 8'S) 2a (7S, 8S, 7'R, 8'R) 2a' (7R, 8R, 7'S, 8'S) 2b (7R, 8R, 7'R, 8'R) 2b' (7S, 8S, 7'S, 8'S) 3a (7S, 8R, 7'S, 8'R, 7"S, 8"S, 7'"S, 8'"S) 3a' (7R, 8S, 7'R, 8'S, 7"R, 8"R, 7'"R, 8'"R) 3b (7S, 8R, 7'S, 8'R, 7"R, 8"R, 7'"R, 8'"R) 3b' (7R, 8S, 7'R, 8'S, 7"S, 8"S, 7'"S, 8'"S) 4a (7S, 8R, 7'S, 8'R, 7"S, 8"S, 7'"S, 8'"S) 4a' (7R, 8S, 7'R, 8'S, 7"R, 8"R, 7'"R, 8'"R) 4b (7S, 8R, 7'S, 8'R, 7"R, 8"R, 7'"R, 8'"R) 4b' (7R, 8S, 7'R, 8'S, 7"S, 8"S, 7'"S, 8'"S)

In this investigation, conformational analysis was performed by CONFLEX 8 (Revision A,

Tokyo, Japan), DFT (density functional theory) geometry optimization and TDDFT (time- dependent density functional theory) calculations were carried out with Gaussian 09 (Wallingford,

CT, USA), and the calculated UV and ECD spectra were plotted using SpecDis (Version 1.71,

Berlin, Germany) (115).

Figure 9. Determination of absolute configuration by ECD calculation

Specifically shown in Figure 9, force field MMFF94S was used for conformer searching of each configuration, and energy window was set at 3 kcal/mol. The obtained stable conformers with populations above 0.5% or 1.0% (Table 6) were optimized further by a DFT method at the

B3LYP/6-31G(d) level, with solvent effects of methanol included using a polarizable continuum model (PCM) (122). Thereafter, the dominant conformers possessing relative energies less than

0.8 or 2.0 kcal/mol were selected (Table 6), and their TDDFT calculations were performed at the

APFD/6-311+g (2d, p) level with PCM (methanol) (123). The spectra obtained of the conformers were averaged by SpecDis according to their Boltzmann distributions, which generate theoretical

ECD (Figure 14, 17, 20 and 23) and UV (Figure 10) spectra of each configuration. UV spectra could be used to validate the method that also used to calculate ECD spectra.

Table 6. Criteria of Conformer Selection for DFT Optimization and TDDFT Calculation of Model Configurations of Compounds 1-4

configuration conformer for DFT optimization conformer for TDDFT calculation (population, number) (relative energy in kcal/mol, number) 1a > 0.5%, 38 < 0.8, 17 1b > 0.5%, 28 < 0.8, 13 2a > 0.5%, 30 < 0.8, 16 2b > 0.5%, 32 < 0.8, 21 3a > 1.0%, 9 < 2.0, 2 3b > 1.0%, 5 < 2.0, 3 4a > 1.0%, 19 < 2.0, 18 4b > 1.0%, 2 < 2.0, 1

Figure 10. Experimental (solid lines) and calculated (dashed lines) UV spectra of compounds 1- 4 and their model structures

3.3. Structural Elucidation

The four compounds were elucidated as two sesquineolignans, (7S,8S,7'R,8'R)- isoamericanol B (1) and americanol B (2), and two dineolignans, moricitrin A (3) and moricitrin

B (4) (Figure 11). Of which, compounds 2-4 are new structures, and the absolute configuration of

1 was assigned for the first time.

1 (= 1a, 7S, 8S, 7'R, 8'R) 2 (= 2a', 7R, 8R, 7'S, 8'S)

3 (= 3a', 7R, 8S, 7'R, 8'S, 7"R, 8"R, 7'"R, 8'"R) 4 (= 4b', 7R, 8S, 7'R, 8'S, 7"S, 8"S, 7'"S, 8'"S)

Figure 11. Chemical structures of compounds 1-4. a, a', b and b' represent model configurations

3.3.1. (7S,8S,7'R,8'R)-Isoamericanol B

22 Compound 1 was obtained as brown amorphous powder with an optical rotation of [α] D -

8.0 (c 0.05, MeOH). Its molecular formula was determined as C27H26O9 by analyzing the

HRESIMS peak at m/z 493.1491 [M - H]- (calcd for 493.1493) and 1H and 13C NMR spectroscopic data (Table 1). Compound 1 was identified as a known sesquineolignan, isoamericanol B1 or B2, by further comparing its NMR data with the previously reported structures (124). The paired signals at δ 144.5 (C-3") and 145.1(C-4") in the 13C NMR spectrum suggested an isoamericanol- type structure with an additional caffeoyl alcohol substitution. The isoamericanol-type moiety was confirmed by the observation of a diagnostic correlation between H-9'b (δ 3.50) and C-4" after optimizing the proton-carbon coupling constant of HMBC to 3 Hz (Figure 12). Two signals at δ

145.4 (C-3' and C-4') suggested that the additional hydroxymethyl group is located on the same side of the molecule as C-4', and the dihydroxy-phenyl group is close to C-3'. The coupling constant J7",8" = 15.3 Hz suggested an E-configuration at the double bond of C-7" and C-8". The large coupling constant between H-7 and H-8 (J = 8.2 Hz) indicated a trans substitution at C-7 and

C-8 (125). All the signals of 1 were assigned with the assistance of HSQC, COSY and HMBC

NMR spectra.

Figure 12. Key COSY (bold lines) and HMBC (blue arrows) correlations of compound 1

ECD spectra have been used extensively to establish the absolute configuration of natural product structures based on relative configurations. However, ROESY correlations between H-

7/H-8 and H-7'/H-8' were not detectable in this investigation due to the long distances between these two steric centers (6.7-8.0 Å, Figure 13), making it impossible to determine the relative configurations of H-7 and H-8 based on the ROESY spectrum. This is consistent with the previous report on elucidating the same type of compounds with this technique (124).

proton distance H-7 to H-7' 7.7 Å H-7 to H-8' 6.7 Å H-8 to H-7' 8.0 Å H-8 to H-8' 7.8 Å

Figure 13. Distances between protons of steric centers for compounds 1 (energy minimized by MM2 force field)

Theoretically, possible trans substitutions of H-7 and H-8 generate two pairs of enantiomers, namely, 1a (7S, 8S, 7'R, 8'R) and 1a' (7R, 8R, 7'S, 8'S), 1b (7R, 8R, 7'R, 8'R) and 1b'

(7S, 8S, 7'S, 8'S). Within this small scope, it was hypothesized that each configuration possesses a unique ECD spectrum. Hence, ECD spectra were measured to determine their relative and absolute configurations simultaneously. The ECD spectra of 1a and 1b were calculated. As shown in Figure

14, the experimental ECD spectrum matched best with 1a, suggesting that 1 has the same absolute configuration as 1a (7S, 8S, 7'R, 8'R). Therefore, the structure of 1 was determined as (7"E)-

(7S,8S,7'R,8'R)-3,4,9,9',9"-pentahydroxy-7,3':7',3"-diepoxy-8,4':8',4"-bisoxysesquineolign-7"- ene, namely, (7S,8S,7'R,8'R)-isoamericanol B. This represents the first time that the absolute configuration of this compound has been established.

Figure 14. Experimental and calculated ECD spectra of compound 1 in MeOH, a and b represent the model configurations of each compound

3.3.2. Americanol B

22 Compound 2 was purified as brown amorphous powder with an optical rotation of [α] D +

2.0 (c 0.05, MeOH). Its molecular formula was determined as C27H26O9 by analyzing the

HRESIMS peak at m/z 493.1486 [M - H]- (calcd for 493.1493) and its 1H and 13C NMR data (Table

1). By comparing its NMR data to those of similar structures isoamericanol C1 or C2 (124), the consistent signals at δC 145.8 (C-3') and 145.0 (C-4') supported a hydroxymethyl group being

located on the same side of the molecule as C-3', and a dihydroxy-phenyl group being on the same side as C-4'. However, different chemical shifts at δC 144.8 (C-3" and C-4") suggested a different substitution for the hydroxypropenyl group in compound 2. Weak correlations from H-9' (δ 3.50) and H-7' to the signals of δC 144.8 were observed in the HMBC spectrum with the proton-carbon coupling constant optimized as 3 Hz (Figure 15), however, the signals at δC 144.8 were assigned to both C-3" and C-4", thus, these correlations could not be used to distinguish the different substitutions. Nevertheless, the characteristics of C-3" and C-4" were identical to those of reported

C-3' and C-4' in americanol A (53, 124, 126). Therefore, the hydroxypropenyl group was proposed as being located on the same side of the molecule as C-8' in compound 2. The coupling constant

J7",8" = 16.0 Hz revealed an E-configuration at C-7" and C-8", J7,8 = 8.1 Hz suggested the trans orientations of H-7 and H-8. All signals were assigned with the assistance of HSQC, COSY and

HMBC NMR spectra.

Figure 15. Key COSY (bold lines) and HMBC (blue arrows) correlations of compound 2

Compound 2 also has two pairs of trans enantiomers, 2a (7S, 8S, 7'R, 8'R) and 2a' (7R, 8R,

7'S, 8'S), and 2b (7R, 8R, 7'R, 8'R) and 2b' (7S, 8S, 7'S, 8'S), and the long distances between H-

7/H-8 and H-7'/H-8' (6.4-7.6 Å, Figure 16) restricted the use of ROESY spectrum for determining its relative configuration.

proton distance H-7 to H-7' 6.4 Å H-7 to H-8' 7.2 Å H-8 to H-7' 6.5 Å H-8 to H-8' 7.6 Å

Figure 16. Distances between protons of steric centers for compounds 2 (energy minimized by MM2 force field)

The ECD spectra of 2a and 2b were calculated, and the experimental spectrum showed an inversed pattern of Cotton effects to 2a (Figure 17). Hence, compound 2 was assigned with the same configuration as 2a', and it was determined as a new sesquineolignan, (7"E)-(7R,8R,7'S,8'S)-

3,4,9,9',9"-pentahydroxy-7,4':7',4"-diepoxy-8,3':8',3"-bisoxysesquineolign-7"-ene, and was named americanol B.

Figure 17. Experimental and calculated ECD spectra of compound 2 in MeOH, a and b represent the model configurations of each compound

3.3.3. Moricitrin A

22 Compound 3 was purified as brown amorphous powder with an optical rotation of [α] D +

2.0 (c 0.05, MeOH). Its molecular formula was determined as C36H34O12 based on the HRESIMS peak at m/z 657.1952 [M - H]- (calcd for 657.1967) and 1H and 13C NMR data (Table 1). It should be noted that the 1H and 13C NMR spectra only showed half of the total numbers of proton and carbon atoms present, suggesting that the structure consists of two identical moieties. The 13C

NMR spectrum presented typical signals of four C6C3 units, with each including six aromatic carbons (δ 115-150) and three alkoxy carbons (δ 62-87), suggesting that compound 3 is a dineolignan. The 13C NMR signals at δ 87.0 (C-7, 7'), 55.3 (C-8, 8') and 72.7 (C-9, 9') and their

1 corresponding H NMR signals at δ 4.66 (H-7, 7'), 3.04 (H-8, 8'), and 4.18 and 3.79 (H-9a, 9'a and

13 9b, 9'b) suggested a symmetric furofuran-type moiety (124). In turn, the C NMR signals at δ 77.6

(C-7", 7'"), 79.9 (C-8", 8'") and 62.1 (C-9", 9'"), and the 1H NMR signals at δ 4.77 (H-7", 7'"), 3.97

(H-8", 8'"), and 3.65 and 3.45 (H-9"a, 9'"a and 9"b, 9'"b) suggested a 1,4-benzodioxane-type unit to be present. Collectively, the structure of compound 3 was established as having a 3, 3'- bisdemethylpinoresinol skeleton with symmetrical additions of two phenylpropanoid groups to C-

3 and C-4, and C-3' and C-4', respectively. The chemical shifts at δC 145.2 (C-3, 3') and 144.4 (C-

4, 4') were consistent with the signals of C-3' and C-4' in isoprincepin (53, 124), suggesting that the two dihydroxy-phenyl groups are located on the same side of the molecule as C-3 and C-3', respectively. The locations of the dihydroxy-phenyl groups were confirmed based on the detected

HMBC correlations between H-7" and C-3, H-7'" and C-3', H-9"b (δ 3.45) and C-4, as well as H-

9'"b (δ 3.45) and C-4', after optimizing the proton-carbon coupling constant to 3 Hz (Figure 18).

All the signals were assigned with the assistance of the HSQC, COSY and HMBC NMR spectra.

Figure 18. Key COSY (bold lines) and HMBC (blue arrows) correlations of compound 3

The large coupling constants of J7",8" (8.0 Hz) and J 7'",8'" (8.0 Hz) suggested the trans substitutions of H-7" and H-8", and H-7'" and H-8'" in compound 3. The ROESY correlations were not detectable for establishing relative configurations due to the large distance from H-7" and H-

8" to H-7 and H-8, from H-7'" and H-8'" to H-7' and H-8' (6.2-7.1 Å, Figure 19).(124)

proton distance H-7 to H-7" 6.2 Å H-7 to H-8" 6.7 Å H-8 to H-7" 7.1 Å H-8 to H-8" 6.3 Å

Figure 19. Distances between protons of steric centers for compounds 3 (energy minimized by MM2 force field)

There are four theoretical configurations including 3a (7S, 8R, 7'S, 8'R, 7"S, 8"S, 7'"S, 8'"S),

3a' (7R, 8S, 7'R, 8'S, 7"R, 8"R, 7'"R, 8'"R), 3b (7S, 8R, 7'S, 8'R, 7"R, 8"R, 7'"R, 8'"R) and 3b' (7R,

8S, 7'R, 8'S, 7"S, 8"S, 7'"S, 8'"S). Among them, ECD spectra of 3a and 3b were simulated. The

experimental results matched closely with the inverse Cotton effects of 3a (Figure 20), suggesting that compound 3 shares the same configuration with 3a'. Therefore, its structure was established as a new dineolignan, (7R,8S,7'R,8'S,7"R,8"R,7'"R,8'"R)-7,9':7',9:3,7":3',7'"-tetraepoxy-4,8":4',8'"- bisoxy-8,8'-dineolignan-3",4",9",3'",4'",9'"-hexaol, named moricitrin A.

Figure 20. Experimental and calculated ECD spectra of compound 3 in MeOH, a and b represent the model configurations of each compound

3.3.4. Moricitrin B

22 Compound 4 was obtained as brown amorphous powder, with an optical rotation of [α] D -

8.0 (c 0.05, MeOH). Its molecular formula was determined as C36H34O12, based on the HRESIMS peak at m/z 657.1954 [M - H]- (calcd for 657.1967) and 1H and 13C NMR data (Table 1). Compound

4 was found also to be a symmetrical dineolignan that nominally showed half the actual numbers of protons and carbons in its NMR spectra. It was determined as a 3, 3'-bisdemethylpinoresinol derivative expanded with two phenylpropanoid groups. The observation of two sets of alkoxy

signals contributed to this deduction. One set of signals at δC 87.0 (C-7, 7'), 55.4 (C-8, 8') and 72.7

(C-9, 9'), and δH 4.67 (H-7, 7'), 3.06 (H-8, 8'), and 4.19 and 3.81 (H-9a, 9'a and 9b, 9'b) suggested a furofuran-type unit, and another set of signals at δC 77.6 (C-7", 7'"), 79.9 (C-8", 8'") and 62.1 (C-

9", 9'"), and δH 4.79 (H-7", 7'"), 3.98 (H-8", 8'"), and 3.65 and 3.45 (H-9"a, 9'"a and 9"b, 9'"b) suggested a 1,4-benzodioxane-type moiety. Also, the 13C NMR spectrum of 4 showed a pair of adjacent chemical shifts at δC 144.4 (C-3, 3') and 144.8 (C-4, 4'), rather than δC 144.4 and 145.2 as present in compound 3, suggesting a different connection of the two phenylpropanoid groups in compound 4. The adjacent signals were consistent with the signals of C-3' and C-4' in princepin

(124), suggesting that the two dihydroxy-phenyl groups are located on the same side of the molecule as C-4 and C-4', while the two hydroxymethyl groups are near to C-3 and C-3'. These assignments were confirmed by HMBC (optimized coupling constant = 3 Hz) correlations between

H-7" and C-4, H-7'" and C-4', H-9"b (δ 3.45) and C-3, and H-9'"b (δ 3.45) and C-3' (Figure 21). All the signals were assigned with the assistance of HSQC, COSY and HMBC NMR spectra.

Figure 21. Key COSY (bold lines) and HMBC (blue arrows) correlations of compound 4

Similar to compound 3, the trans relationships of H-7" and H-8", and H-7'" and H-8'" in compound 4 were also determined based on the large coupling constant 7.9 Hz for both J7",8" and

J 7'",8'", and the ROESY correlations between H-7/H-8 and H-7"/H-8", and H-7'/H-8' and H-7'"/H-

8'" were not detectable due to the long distance between these two sets of steric centers (6.8-7.7 Å,

Figure 22).

proton distance H-7 to H-7" 7.7 Å H-7 to H-8" 6.8 Å H-8 to H-7" 7.6 Å H-8 to H-8" 7.3 Å

Figure 22. Distances between protons of steric centers for compounds 4 (energy minimized by MM2 force field)

Among the four possible configurations, 4a (7S, 8R, 7'S, 8'R, 7"S, 8"S, 7'"S, 8'"S), 4a' (7R,

8S, 7'R, 8'S, 7"R, 8"R, 7'"R, 8'"R), 4b (7S, 8R, 7'S, 8'R, 7"R, 8"R, 7'"R, 8'"R) and 4b' (7R, 8S, 7'R,

8'S, 7"S, 8"S, 7'"S, 8'"S), the ECD spectra for 4a and 4b were calculated. The experimental data showed inverse Cotton effects with 4b (Figure 23). Thus, the structure of compound 4, named moricitrin B, was established as a novel dineolignan with the same configuration with 4b', namely,

(7R,8S,7'R,8'S,7"S,8"S,7'"S,8'"S)-7,9':7',9:4,7":4',7'"-tetraepoxy-3,8":3',8'"-bisoxy-8,8'- dineolignan-3",4",9",3'",4'",9'"-hexaol.

Figure 23. Experimental and calculated ECD spectra of compound 4 in MeOH, a and b represent the model configurations of each compound

CHAPTER 4 Evaluation of Inhibitory Activities

From the raw extract of noni fruits to the final mixture (F-3-2-2) containing compounds 1-

4, bioactive fractions were traced by inhibition assay against bacterial β-glucuronidase. Once purified 1-4 were obtained, their individual inhibitory activities were evaluated to confirm if they are the bioactive contributors to noni fruits. In the small intestine, there are also other important enzymes like α-amylase, α-glucosidase and pancreatic lipase that are responsible for the digestion of carbohydrates and triglycerides (127). Inhibition on these digestive enzymes may introduce additional adverse gastrointestinal (GI) effects. Therefore, compounds 1-4 were examined further for inhibitory activities against these enzymes to determine whether they are specific inhibitors against bacterial β-glucuronidase with minimal GI side effects.

4.1. Inhibitory Assays on Enzymes

4.1.1. Enzymes and Reagents

All enzymes used in this study, including β-glucuronidase from Escherichia coli (500 kU), porcine pancreatic α-amylase (type VI-B), rat intestinal acetone powder, and pancreatic lipase

(type VI-S), were purchased from Sigma-Aldrich (St. Louis, MO, USA). The substrates 4-methyl umbelliferyl-β-D-glucuronide hydrate, p-nitrophenyl-α- D-glucopyranoside (pNPG), and 4-methyl umbelliferyl oleate (4-MU oleate) were obtained from Sigma-Aldrich, potato starch powder was from Fisher Scientific (Fair Lawn, NJ, USA). Dinitrosalicylic acid reagent (DNS) from Sigma-

Aldrich (St. Louis, MO, USA) acted as an indicator for α-amylase assay. Acarbose (Alfa Aesar,

Ward Hill, MA, USA), D-saccharic acid 1,4-lactone monohydrate (DSA, Sigma-Aldrich, St. Louis,

MO, USA), and orlistat (Sigma-Aldrich, St. Louis, MO, USA) were used as positive controls. The

absorbance and fluorescence of inhibitory assays were measured in EnSpire multimode plate reader (PerkinElmer, Waltham, MA, USA)

4.1.2. Bacterial β-Glucuronidase Assay

The inhibitory activities against bacterial β-glucuronidase were measured using black 96- well microplates (42). Lyophilized samples were dissolved in 50% DMSO at various concentration levels. A portion (1.5 µL) of each was transferred into a well, followed by addition of 60 µL of β- glucuronidase (6.25 U/mL). The mixture was incubated for 5 min at room temperature, then added with 40 µL of 4-methyl umbelliferyl-β- D-glucuronide hydrate (312.5 µM). After further incubation of 30 min at 37 °C, the fluorescence values were measured at an excitation wavelength of 355 nm and an emission wavelength of 460 nm. The results are expressed as percent inhibition compared with solvent control, and DSA was used as positive control. IC50 values were calculated based on inhibition percentages of series dilutions.

4.1.3. α-Amylase Assay

The α-amylase inhibitory activities were performed based on a previously established method with small modifications (128). The samples and α-amylase (4 U/mL) were mixed and incubated at room temperature for 5 min, then the substrate 0.5% potato starch solution was added and incubated for 3 min at 37 °C. After addition of DNS reagent and reaction of 10 min at 90 °C,

25 µL of the cooled reaction mixture were loaded onto a 96-well plate and diluted with distilled water. The absorbance was measured at 540 nm, and the results are expressed as percent inhibition compared with solvent control, with acarbose as positive control.

4.1.4. α-Glucosidase Assay

Rat α-glucosidase was used to evaluate inhibitory activities (128). The enzyme was extracted from rat intestinal acetone powder with 0.1 M potassium phosphate buffer (pH 6.8). The samples and α-glucosidase (0.025 g/mL) were incubated at 37 °C for 3 min, then 4 mM pNPG were added and incubated for additional 30 min. The released p-nitrophenol from pNPG was measured at 405 nm, which was inversely related to the sample’s inhibition level. The results are expressed as percent inhibition compared with solvent control, with acarbose as positive control.

4.1.5. Pancreatic Lipase Assay

The enzyme inhibition was determined by measuring the released 4-methylumbelliferone from 4-MU oleate (128). In brief, the samples and lipase (50 U/mL) were mixed with the substrate

4-MU oleate (0.1 mM), and incubated for 30 min at 37 °C. The fluorescence was measured at excitation of 355 nm and emission of 460 nm. The results are expressed as percent inhibition compared with solvent control, and orlistat (5 µg/mL) was used as positive control.

4.1.6. Data Analysis

Data from the bioactivity assays are reported as means ± SD, with the statistics analysis performed using SPSS for Windows (version 25, 2017, IBM). IC50 values were calculated by a probit procedure, and the statistical significance between groups was compared by one-way

ANOVA and the Tukey’s test (p < 0.05 or 0.01).

4.2. Specific Inhibition Against Bacterial β-Glucuronidase

As shown in Table 7, the active compounds 1-4 exhibited potent inhibition against bacterial

β-glucuronidase, with IC50 values ranging from 0.62 to 6.91 µM. Their activities were 3.5 to 39.5-

fold more potent than the positive control DSA, which gave an IC50 of 24.30 µM. It is also worth noting that the isolated natural dineolignans 3 and 4 were more active against β-glucuronidase than several synthesized novel inhibitors with IC50 values ranging from 1.7 to 4.8 µM (42). These results may lead to new applications for the bioactive components of M. citrifolia (noni). As a nutritional agent (52, 55), noni fruits could be developed into a safe and readily available dietary source to more effectively prevent irinotecan-induced diarrhea than those synthesized compounds by more potently inhibiting bacterial β-glucuronidase.

Table 7. Inhibitory Activities against Bacterial β-Glucuronidase

a compound IC50 (µM) ± SD 1 6.91 ± 0.16e 2 4.02 ± 0.06d 3 0.62 ± 0.01c 4 0.95 ± 0.01c DSAb 24.30 ± 0.80f a b IC50 values in µM, and SD represents standard deviation (n = 3). DSA (D-saccharic acid 1,4- lactone monohydrate) was used as positive control. Different letters (c-f) represent significant differences with p < 0.01.

Interestingly, the dineolignans (compounds 3 and 4) showed an order of magnitude more potent inhibition than the sesquineolignans (compounds 1 and 2), suggesting a strong structure- activity relationship. Bacterial β-glucuronidase has a loop that is absent in human β-glucuronidase, and selective inhibitors exert their activity by binding to the overlapping loops at the tetramer interface of the enzyme (22, 33). Compared to the sesquineolignans (1 and 2), the dineolignans (3 and 4) possess larger molecular sizes and more number of active hydroxy groups, which may facilitate their binding to the active sites of the enzyme. Although sharing a similar skeleton,

compound 1 showed more potent β-glucuronidase inhibitory activity than compound 2, and the activity of compound 3 was also greater than that of compound 4, suggesting that their activities are also influenced by the substitution variations of hydroxypropenyl and dihydroxyphenyl groups.

The results should provide valuable insights for searching for more effective inhibitors in the future.

4.3. Inhibition on Digestive Enzymes

As shown in Figure 24A, all the compounds showed no or weak effects on intestinal α- amylase. At the concentrations of 0.1 and 0.5 mg/mL, none of them exhibited significant inhibition.

When the concentration was increased to 1.0 mg/mL, compounds 2 (2.02 mM), 3 (1.52 mM) and

4 (1.52 mM) showed weak inhibition on α-amylase (the highest inhibition was 21%) compared to the positive control acarbose, which resulted in 94% inhibition at 10 µg/mL (0.02 mM). Similar to

α-amylase, α-glucosidase was only minimally inhibited by compounds 1-4 at concentrations up to

1 mg/mL (Figure 24B). The highest inhibition of only 12% was from compound 1 at 1.0 mg/mL

(2.02 mM), while acarbose showed 80% inhibitory activity at 0.02 mM. Figure 24C shows that compounds 1-4 exerted modest inhibition on lipase. At 0.1 mg/mL, all compounds exhibited inhibition of less than 56%, and their activities increased in the order from 1 to 4. Up to 1.0 mg/mL, they reached similar suppression values ranging from 85% to 93%, which were close to orlistat

(96%). However, at this inhibition level, 1 (2.02 mM) and 2 (2.02 mM) were at a concentration level 200 times that of orlistat [0.01 mM (5 µg/mL)], while 3 (1.52 mM) and 4 (1.52 mM) were at a concentration 152 times that of orlistat (0.01 mM). Therefore, the inhibition of compounds 1-4 on lipase was relatively weak versus the positive control used.

A a lp h a -A m y la s e

1 0 0 e 0 .1 m g /m L

0 .5 m g /m L

7 5 )

1 .0 m g /m L

(

o 1 0 µ g /m L i

t 5 0

n I 2 5 d d

a b c b c a b c a b a b a b a a a 0 C 1 C 2 C 3 C 4 A c a rb o s e

B a lp h a -G lu c o s id a s e

1 0 0 0 .5 m g /m L

c 1 .0 m g /m L

7 5 )

1 0 µ g /m L

(

n o

i 5 0

n I 2 5

b a b a b a b a b a a b a b 0 C 1 C 2 C 3 C 4 A c a rb o s e

C P a n c re a tic L ip a s e h 1 0 0 g h f g 0 .1 m g /m L f f f f

e 0 .5 m g /m L

) 7 5

% 1 .0 m g /m L (

d n

o 5 µ g /m L i

t 5 0 i c

b i

h b

n I 2 5 a

0 C 1 C 2 C 3 C 4 O rlis ta t

Figure 24. Inhibitory effects of compounds 1-4 against α-amylase (A), α-glucosidase (B) and pancreatic lipase (C). Acarbose (10 µg/mL) and orlistat (5 µg/mL) were used as positive controls. C1-4 represent compounds 1-4. The vertical bars represent the standard deviation (n = 3) for each data point, and different letters represent significant differences (p < 0.05).

CHAPTER 5 Discussion

5.1. Bioassay-guided Isolation

Chemical constituents of natural resources are often more complex than expected in structural types and numbers. There are no fixed processes to do extraction and isolation for those constituents. However, the countless combinations of isolation processes are greatly decided by the purposes of specific projects. To characterize chemical constituents of rarely or newly studied natural materials, a systematic isolation is often performed to obtain as many compounds as possible. Isolation of specific type of constituents needs to be guided by their unique chemical properties. For instance, the basic property and selective reaction with Dragendroff’s reagent can be used to guide the isolation of alkaloids. Bioassay-guided isolation has been widely used for digging out active leads from chemical mixtures, such as paclitaxel from Taxus brevifolia and camptothecin from Camptotheca acuminata, they were successfully discovered following target bioactive assay (1). The aim of this project is to find out active compounds alleviating irinotecan- induced delayed diarrhea, thus the procedure was started with the method establishment of appropriate in vitro biological assay (129). The diarrhea is caused by the intestinal SN-38 that is primarily produced by bacterial β-glucuronidase. Therefore, inhibitory assay against bacterial β- glucuronidase was established to trace active compounds from extracts of noni fruits.

During the isolation process, it is notable that activity of an extract or fraction is tested negative or weak doesn’t mean that it is not worth further investigation. The principle of “like dissolves like” concludes that polar/ionic solvents dissolve polar/ionic solutes and non-polar solvents dissolve non-polar solutes. Selection of extraction solvents greatly impact on what constituents can be obtained from the same natural material. Therefore, it is necessary to change

solvents or adjust ratio of solvent combination to obtain bioactive extractions in some cases. A famous example is the discovery of artemisinin (Qinghaosu). The success is attributed to the change of extract solvent from hot water to low-temperature ethyl ether by Tu Youyou who was awarded Nobel Prizes in 2015 (130, 131). 50% aqueous acetone was used to extract noni fruit powder in this project. It is a good solvent mixture to dissolve both nonpolar and polar substances, which ensured the chemical diversity and inclusion of active compounds for noni fruits extract.

The weak activities of some fractions may be caused by relatively lower concentration of active components, which can be enhanced by enriching active compounds. As seen in this project, the raw extract of noni fruits showed only about 50% inhibition against bacterial β-glucuronidase at

0.125 mg/mL. Along with continuous separation that get rid of inactive components, the inhibitory activities of refined fractions were increased up to 90%. Eventually, the strong inhibitors, compounds 1-4, were acquired with the assistance of bioassay. However, in some opposite cases, high activity of an extraction is a result of synergistic effects of low active components, which may turn out losing active targets after further separation (132, 133). This should be considered when using bioassay-guided isolation.

5.2. Structure Elucidations

Based on the characteristics of compounds 1-4, the commonly shared 1,4-benzodioxane moiety makes it challenging to elucidate their structures. Firstly, the two oxygen atoms impede the informative HMBC correlations across the1,4-dioxane ring, which was also observed in the reported compounds, such as isoamericanin A, americanol A, isoamericanol A, princepin, isoperincepin, isoamericanol B1, B2, C1, and C2 (124), haedoxancoside A (134), and obovatalignan A (135). To determine the locations of hydroxymethyl and dihydroxy-phenyl groups,

the parameters of HMBC experiment can be optimized to achieve longer correlations (124). As shown in this investigation, the expected HMBC correlations across the 1,4-dioxane ring were presented after optimizing the proton-carbon coupling constant to 3 Hz. Secondly, the rigid plane of 1,4-benzodioxane separates the two steric centers of compounds 1-4 with theoretical distances in the range 6.2-8.0 Å (Figure 13, 16, 19 and 22). NOESY and ROESY correlations could be observed for protons separated typically by less than 5 Å (136, 137), thus, they are not applicable to establish the relative configurations for this type of molecules.

ECD computation is often based on confirmed relative configurations, but also available for known possible relative configurations (138). Since the relative configurations were not decided, each of compounds 1-4 has eight configurations theoretically. Because the hydroxymethyl and dihydroxy-phenyl groups in each compound were assigned as trans substitutions, the numbers of possible configurations were reduced to four. Just based on these four possibilities, their absolute configurations were established through ECD calculation.

5.3. In vitro Study

In this study, purified compounds from noni fruits showed potent and specific inhibition against β-glucuronidase of Escherichia coli. However, there are still several essential steps needed to extrapolate this promising work to the clinical application alleviating irinotecan-induced delayed diarrhea.

We conclude the inhibitory selectivity of our compounds 1-4 based on their none or weak inhibition against digestive enzymes. However, their inhibition on other lysosomal β- glucuronidases of mammalian host may also need to be investigated. For instance, human β- glucuronidase has been considered as reliable biomarker for diagnosing diseases like different

cancers, and potential target for the development of anti-inflammatory and anticancer drugs (139).

Wallace et al. elucidated the crystal structure of the first bacterial β-glucuronidase from E. coli, and found that it has a 45% of sequence similarity with human β-glucuronidase (33). We don’t expect that our compounds 1-4 have side effects on glycosyl hydrolysis in host tissues and biofluids because of this similarity. It can be verified by measuring their inhibition against bovine liver β-glucuronidase, modified E. coli β-glucuronidase or other lysosomal β-glucuronidases (33).

Wallace et al. also revealed that β-glucuronidase of E. coli possesses a unique “bacterial loop” of

17 residues that is widely distributed in other human gut microbiota, while lacked in human β- glucuronidase (33). Molecular docking study are used to investigate structure-activity relationship

(SAR) between inhibitors and human β-glucuronidase (140-142). The inhibitory selectivity of compounds 1-4 can also be predicted through docking study that shows binding interactions of compounds with the key active sites of the bacterial loop.

In another investigation by Wallace et al., β-glucuronidases from Streptococcus agalactiae and Clostridium perfringens of Firmicutes, Escherichia coli of Proteobacterium and Bacteroides fragilis of Bacteroidetes were compared together (11). These orthologous enzymes shared very similar structures but showed different catalytic properties and were variously inhibited by designed compounds. It suggests that our selective inhibitors against E. coli β-glucuronidase may also interact very differently with enzymes of other bacteria, which result in different net effect on the overall deconjugation of SN-38G. Therefore, we need to measure inhibition of our compounds on the enzyme of as more representative bacteria as possible, to evaluate their potential impacts on overall microbial β-glucuronidases. In addition, we need the most powerful data of animal and clinical experiments to convince our hypothesis that compounds 1-4 can alleviate irinotecan- induced delayed diarrhea.

5.4. In vivo Study

Balb/cJ mice are often used in cancer studies because of their tendency to develop different kinds of tumor (Figure 25). For instance, tumor-bearing Balb/cJ mice have been utilized for the discovery of irinotecan (CPT-11) (143). Unlike human β-glucuronidase, research on bacterial β- glucuronidase inhibitors is a relatively new field that was firstly reported in 2010, and few of these inhibitors have reached animal experiment (139). In both investigations of Wallace et al. and

Cheng et al., healthy Balb/cJ female mice were used to evaluate bacterial β-glucuronidase inhibitors’ effects in alleviating irinotecan-induced delayed diarrhea (33, 44). Therefore, healthy

Balb/cJ model could also be used to assess our specific inhibitors. The experimental mice can be assigned into four groups and treated with 1) solvent of irinotecan intraperitoneally (i.p.) + solvent of inhibitor via oral gavage, 2) irinotecan + solvent of inhibitor, 3) solvent of irinotecan + inhibitor,

4) irinotecan + inhibitor. The dose of inhibitor needs to be calculated carefully according to literature and preliminary experiment.

Figure 25. Balb/cJ mouse

Besides of inhibitors’ effects on delayed diarrhea, their influence on irinotecan’s anticancer ability is another important consideration need to be figured out by animal experiment. Gupta et

al. showed that there was an obvious secondary peak of serum SN-38 level in patients infused intravenously with irinotecan (17). Kehrer et al. found that SN-38 could be transported across the membrane of colonic Caco-2 monolayers from apical to basolateral (144). These observations indicate that intestinal SN-38 may be reabsorbed into plasma to exert anticancer effect. In other words, inhibition of bacterial β-glucuronidase may reduce plasma SN-38 level and subsequently impair the therapeutic efficacy of irinotecan. As shown in the research on inhibitor pyrazolo[4,3- c]quinoline derivative (TCH-3562) by Cheng et al. (44), monitoring the plasma pharmacokinetics of SN-38 is a practicable approach to learn the influence of our specific inhibitors on irinotecan’s efficacy. In addition, the inhibitors’ influence on the growth of gut bacteria can be determined by incubating our inhibitors with live bacteria, such as E. coli cells (44). It can also be examined by profiling entire gut microbial composition using16S rRNA sequencing (145).

Up to date, there has no clinical trial been reported for inhibitors of E. coli β-glucuronidase in alleviating irinotecan-induced delayed diarrhea. Nevertheless, as mentioned above, some synthesized compounds have been shown strong inhibition on the enzyme. Animal studies have demonstrated their capacities in reducing delayed diarrhea and maintaining efficacy of irinotecan, while with limited impact on normal gut microbial community and the growth of epithelial cells.

These researches have established a basic framework to develop inhibitors of bacterial β- glucuronidase. Our compounds 1-4 possess potent and specific inhibition against the enzyme, and with minimized side effects on the intestinal digestion, which have been a solid foundation for further research and development. Moreover, our specific inhibitors are discovered from noni fruits, which is a widely distributed and dietary safe natural resource. Therefore, the noni fruits and their active compounds are promising to be developed as a dietary supplement or therapeutic approach to alleviate the irinotecan-induced delayed diarrhea.

APPENDIX

HRESIMS (Figure A1-A4), and 1H, 13C, COSY, HSQC, HMBC and ROESY NMR spectra for

(7S,8S,7'R,8'R)-isoamericanol B (1, Figure A5-A11), americanol B (2, Figure A12-A18), moricitrin A (3, Figure A19-A25), and moricitrin B (4, Figure A26-A32).

Figure A1. HRESIMS spectrum of (7S,8S,7'R,8'R)-isoamericanol B (1)

H4_negative #23-38 RT: 0.62-1.03 AV: 16 NL: 3.33E6 T: FTMS - c ESI Full ms [150.00-1500.00] 493.1486 C 27 H25 O9 = 493.1493 -1.5236 ppm

3200000

3000000

2800000

2600000

2400000

2200000

2000000

1800000

1600000 Intensity

1400000

1200000

1000000 494.1520

800000

600000

400000

200000 495.1558 491.5239 492.1371 492.5731 493.2663 493.9382 494.5396 496.1555 497.0287 497.9535 498.4118 498.8196 499.4073 500.0748 0 491.5 492.0 492.5 493.0 493.5 494.0 494.5 495.0 495.5 496.0 496.5 497.0 497.5 498.0 498.5 499.0 499.5 500.0 m/z

Figure A2. HRESIMS spectrum of americanol B (2)

H6_negative #16-24 RT: 0.42-0.64 AV: 9 NL: 6.62E6 T: FTMS - c ESI Full ms [150.00-1500.00] 657.1952 C 36 H33 O12 = 657.1967 -2.1968 ppm 6500000

6000000

5500000

5000000

4500000

4000000

3500000 Intensity 3000000

659.2018 2500000

2000000

1500000

1000000 693.1719 C 36 H34 O12 Cl = 693.1733 -2.0668 ppm 720.1907 500000 687.2058 695.1690 671.1747 685.1903 725.1825 664.5959 679.1772 703.2005 712.0649 717.2159 733.2105 742.1724 750.2009 755.1622 0 650 660 670 680 690 700 710 720 730 740 750 760 m/z

Figure A3. HRESIMS spectrum of moricitrin A (3)

H7_negative #19-27 RT: 0.50-0.71 AV: 9 NL: 7.16E6 T: FTMS - c ESI Full ms [150.00-1500.00] 657.1954 C 36 H33 O12 = 657.1967 -1.8533 ppm 7000000

6500000

6000000

5500000

5000000

4500000

4000000

3500000 Intensity 658.1988 3000000 C 43 H30 O7 = 658.1986 0.3440 ppm

2500000

2000000

1500000

1000000 659.2018 C 25 H39 O20 = 659.2029 655.1800 -1.6330 ppm 500000 662.1179 663.5367 C 36 H31 O12 = 655.1810 657.6971 C H O = 662.1172 C H O = 663.5347 -1.4611 ppm 22 30 23 44 71 4 656.6897 660.7556 0.9991 ppm 3.0066 ppm 0 655.0 655.5 656.0 656.5 657.0 657.5 658.0 658.5 659.0 659.5 660.0 660.5 661.0 661.5 662.0 662.5 663.0 663.5 664.0 m/z Figure A4. HRESIMS spectrum of moricitrin B (4)

1 Figure A5. H NMR (500 MHz, CD3OD) spectrum of (7S,8S,7'R,8'R)-isoamericanol B (1)

13 Figure A6. C NMR (125 MHz, CD3OD) spectrum of (7S,8S,7'R,8'R)-isoamericanol B (1)

Figure A7. COSY (CD3OD) spectrum of (7S,8S,7'R,8'R)-isoamericanol B (1)

Figure A8. HSQC (CD3OD) spectrum of (7S,8S,7'R,8'R)-isoamericanol B (1)

Figure A9. HMBC (CD3OD) spectrum of (7S,8S,7'R,8'R)-isoamericanol B (1)

Figure A5. HMBC (600 MHz, CD3OD, coupling constant = 3 Hz) spectrum of (7S,8S,7'R,8'R)- isoamericanol B (1)

Figure A6. ROESY (CD3OD) spectrum of (7S,8S,7'R,8'R)-isoamericanol B (1)

1 Figure A7. H NMR (500 MHz, CD3OD) spectrum of americanol B (2)

13 Figure A8. C NMR (125 MHz, CD3OD) spectrum of americanol B (2)

Figure A9. COSY (CD3OD) spectrum of americanol B (2)

Figure A10. HSQC (CD3OD) spectrum of americanol B (2)

Figure A11. HMBC (CD3OD) spectrum of americanol B (2)

Figure A12. HMBC (600 MHz, CD3OD, coupling constant = 3 Hz) spectrum of americanol B (2)

Figure A18. ROESY (CD3OD) spectrum of americanol B (2)

1 Figure A19. H NMR (500 MHz, CD3OD) spectrum of moricitrin A (3)

13 Figure A20. C NMR (125 MHz, CD3OD) spectrum of moricitrin A (3)

Figure A13. COSY (CD3OD) spectrum of moricitrin A (3)

Figure A14. HSQC (CD3OD) spectrum of moricitrin A (3)

Figure A15. HMBC (CD3OD) spectrum of moricitrin A (3)

Figure A16. HMBC (600 MHz, CD3OD, coupling constant = 3 Hz) spectrum of moricitrin A (3)

Figure A17. ROESY (CD3OD) spectrum of moricitrin A (3)

1 Figure A18. H NMR (500 MHz, CD3OD) spectrum of moricitrin B (4)

13 Figure A19. C NMR (125 MHz, CD3OD) spectrum of moricitrin B (4)

Figure A28. COSY (CD3OD) spectrum of moricitrin B (4)

Figure A29. HSQC (CD3OD) spectrum of moricitrin B (4)

Figure A30. HMBC (CD3OD) spectrum of moricitrin B (4)

Figure A31. HMBC (600 MHz, CD3OD, coupling constant = 3 Hz) spectrum of moricitrin B (4)

Figure A32. ROESY (CD3OD) spectrum of moricitrin B (4)

REFERENCES

1. Wall ME, Wani MC. Camptothecin and taxol: from discovery to clinic. J

Ethnopharmacol. 1996;51:239-54.

2. Hsiang Y-H, Hertzberg R, Hecht S, Liu L-F. Camptothecin induces protein-linked DNA breaks via mammalian DNA topoisomerase I. J Biol Chem. 1985;260:14873-8.

3. Armand J, Ducreux M, Mahjoubi M, Abigerges D, Bugat R, Chabot G, Herait P, De

Forni M, Rougier P. CPT-11 (irinotecan) in the treatment of colorectal cancer. Eur J Cancer.

1995;31:1283-7.

4. Giovanella BC, Stehlin JS, Wall ME, Wani MC, Nicholas AW, Liu LF, Silber R,

Potmesil M. DNA topoisomerase I-targeted chemotherapy of human colon cancer in xenografts.

Science. 1989;246:1046-8.

5. Hwang J, Shyy S, Chen AY, Juan C-C, Whang-Peng J. Studies of topoisomerase-specific antitumor drugs in human lymphocytes using rabbit antisera against recombinant human topoisomerase II polypeptide. Cancer Res. 1989;49:958-62.

6. Hsiang Y-H, Lihou MG, Liu LF. Arrest of replication forks by drug-stabilized topoisomerase I-DNA cleavable complexes as a mechanism of cell killing by camptothecin.

Cancer Res. 1989;49:5077-82.

7. Masuda N, Kudoh S, Fukuoka M. Irinotecan (CPT-11): pharmacology and clinical applications. Crit Rev Oncol Hematol. 1996;24:3-26.

8. Bailly C. Irinotecan: 25 years of cancer treatment. Pharmacol Res. 2019;148:e104398.

9. Femke M, Goey AK, van Schaik RH, Mathijssen RH, Bins S. Individualization of irinotecan treatment: a review of pharmacokinetics, pharmacodynamics, and pharmacogenetics.

Clin Pharmacokinet. 2018;57:1229-54.

10. Deboever G, Hiltrop N, Cool M, Lambrecht G. Alternative treatment options in colorectal cancer patients with 5–fluorouracil-or capecitabine-induced cardiotoxicity. Clin

Colorectal Cancer. 2013;12:8-14.

11. Wallace BD, Roberts AB, Pollet RM, Ingle JD, Biernat KA, Pellock SJ, Venkatesh MK,

Guthrie L, O’Neal SK, Robinson SJ. Structure and inhibition of microbiome β-glucuronidases essential to the alleviation of cancer drug toxicity. Chem Biol. 2015;22:1238-49.

12. Koselke E, Kraft S. Chemotherapy-Induced Diarrhea: Options for Treatment and

Prevention. J Hematol Oncol Pharm. 2012;2:143-51.

13. Stein A, Voigt W, Jordan K. Chemotherapy-induced diarrhea: pathophysiology, frequency and guideline-based management. Ther Adv Med Oncol. 2010;2:51-63.

14. Bleiberg H, Cvitkovic E. Characterisation and clinical management of CPT-11

(irinotecan)-induced adverse events: the European perspective. Eur J Cancer. 1996;32:S18-S23.

15. Hecht JR. Gastrointestinal toxicity of irinotecan. Oncology. 1998;12:72-8.

16. Kawato Y, Aonuma M, Hirota Y, Kuga H, Sato K. Intracellular roles of SN-38, a metabolite of the camptothecin derivative CPT-11, in the antitumor effect of CPT-11. Cancer

Res. 1991;51:4187-91.

17. Gupta E, Lestingi TM, Mick R, Ramirez J, Vokes EE, Ratain MJ. Metabolic fate of irinotecan in humans: correlation of glucuronidation with diarrhea. Cancer Res. 1994;54:3723-5.

18. Mathijssen RH, Van Alphen RJ, Verweij J, Loos WJ, Nooter K, Stoter G, Sparreboom A.

Clinical pharmacokinetics and metabolism of irinotecan (CPT-11). Clin Cancer Res.

2001;7:2182-94.

19. Benson III AB, Ajani JA, Catalano RB, Engelking C, Kornblau SM, Martenson Jr JA,

McCallum R, Mitchell EP, O'Dorisio TM, Vokes EE. Recommended guidelines for the treatment of cancer treatment-induced diarrhea. J Clin Oncol. 2004;22:2918-26.

20. Andreyev J, Ross P, Donnellan C, Lennan E, Leonard P, Waters C, Wedlake L,

Bridgewater J, Glynne-Jones R, Allum W. Guidance on the management of diarrhoea during cancer chemotherapy. Lancet Oncol. 2014;15:e447-e60.

21. Tang L, Li X, Wan L, Xiao Y, Zeng X, Ding H. Herbal medicines for irinotecan-induced diarrhea. Front Pharmacol. 2019;10:e00182.

22. Pellock SJ, Creekmore BC, Walton WG, Mehta N, Biernat KA, Cesmat AP, Ariyarathna

Y, Dunn ZD, Li B, Jin J. Gut microbial β-glucuronidase inhibition via catalytic cycle interception. ACS Cent Sci. 2018;4:868-79.

23. Naz H, Islam A, Waheed A, Sly WS, Ahmad F, Hassan MI. Human β-glucuronidase: structure, function, and application in enzyme replacement therapy. Rejuvenation Res.

2013;16:352-63.

24. Little MS, Pellock SJ, Walton WG, Tripathy A, Redinbo MR. Structural basis for the regulation of β-glucuronidase expression by human gut Enterobacteriaceae. Proc Natl Acad Sci

U S A. 2018;115:e152-e61.

25. Takasuna K, Hagiwara T, Hirohashi M, Kato M, Nomura M, Nagai E, Yokoi T,

Kamataki T. Involvement of β-glucuronidase in intestinal microflora in the intestinal toxicity of the antitumor camptothecin derivative irinotecan hydrochloride (CPT-11) in rats. Cancer Res.

1996;56:3752-7.

26. Takasuna K, Hagiwara T, Hirohashi M, Kato M, Nomura M, Nagai E, Yokoi T,

Kamataki T. Inhibition of intestinal microflora β-glucuronidase modifies the distribution of the

active metabolite of the antitumor agent, irinotecan hydrochloride (CPT-11) in rats. Cancer

Chemother Pharmacol. 1998;42:280-6.

27. Flieger D, Klassert C, Hainke S, Keller R, Kleinschmidt R, Fischbach W. Phase II clinical trial for prevention of delayed diarrhea with cholestyramine/levofloxacin in the second- line treatment with irinotecan biweekly in patients with metastatic colorectal carcinoma.

Oncology. 2007;72:10-6.

28. Cummings JH, Macfarlane GT. Role of intestinal bacteria in nutrient metabolism. Clin

Nutr. 1997;16:3-11.

29. Guarner F, Malagelada J-R. Gut flora in health and disease. Lancet. 2003;361:512-9.

30. Settle C, Wilcox M. antibiotic‐induced Clostridium difficile infection. Aliment

Pharmacol Ther. 1996;10:835-41.

31. Stamp D. Antibiotic therapy may induce cancers in the colon and breasts through a mechanism involving bile acids and colonic bacteria. Med Hypotheses. 2004;63:555-6.

32. Sears S, McNally P, Bachinski MS, Avery R. Irinotecan (CPT-11) induced colitis: report of a case and review of Food and Drug Administration MEDWATCH reporting. Gastrointest

Endosc. 1999;50:841-4.

33. Wallace BD, Wang H, Lane KT, Scott JE, Orans J, Koo JS, Venkatesh M, Jobin C, Yeh

L-A, Mani S. Alleviating cancer drug toxicity by inhibiting a bacterial enzyme. Science.

2010;330:831-5.

34. Levvy G. The preparation and properties of β-glucuronidase. 4. Inhibition by sugar acids and their lactones. Biochem J. 1952;52:464-72.

35. Karunairatnam M, Levvy G. The inhibition of β-glucuronidase by saccharic acid and the role of the enzyme in glucuronide synthesis. Biochem J. 1949;44:599-604.

36. Fittkau M, Voigt W, Holzhausen H-J, Schmoll H-J. Saccharic acid 1.4-lactone protects against CPT-11-induced mucosa damage in rats. J Cancer Res Clin Oncol. 2004;130:388-94.

37. Mori K, Kondo T, Kamiyama Y, Kano Y, Tominaga K. Preventive effect of Kampo medicine (Hangeshashin-to) against irinotecan-induced diarrhea in advanced non-small-cell lung cancer. Cancer Chemother Pharmacol. 2003;51:403-6.

38. Sakata Y, Suzuki H, Kamataki T. Preventive effect of TJ-14, a kampo (Chinese herb) medicine, on diarrhea induced by irinotecan hydrochloride (CPT-11). Gan To Kagaku Ryoho.

1994;21:1241-4.

39. Narita M, Nagai E, Hagiwara H, Aburada M, Yokoi T, Kamataki T. Inhibition of β- glucuronidase by natural glucuronides of Kampo, medicines using glucuronide of SN-38 (7- ethyl-10-hydroxycamptothecin) as a substrate. Xenobiotica. 1993;23:5-10.

40. Takasuna K, Kasai Y, Kitano Y, Mori K, Kobayashi R, Hagiwara T, Kakihata K,

Hirohashi M, Nomura M, Nagai E. Protective effects of kampo medicines and baicalin against intestinal toxicity of a new anticancer camptothecin derivative, irinotecan hydrochloride (CPT‐

11), in rats. Jpn J Cancer Res. 1995;86:978-84.

41. Redinbo MR, Mani S, Williams A, Scott J, Yeh L-A, Wallace BD, Lane KT, inventors;

Selective beta-glucuronidase inhibitors as a treatment for side effects of camptothecin antineoplastic agents. US patent 9,334,288. 2013.

42. Ahmad S, Hughes MA, Lane KT, Redinbo MR, Yeh L-A, Scott JE. A high throughput assay for discovery of bacterial β-glucuronidase inhibitors. Curr Chem Genomics. 2011;5:13-20.

43. Cheng K-W, Tseng C-H, Yang C-N, Tzeng C-C, Cheng T-C, Leu Y-L, Chuang Y-C,

Wang J-Y, Lu Y-C, Chen Y-L. Specific inhibition of bacterial β-glucuronidase by pyrazolo [4, 3-

c] quinoline derivatives via a pH-dependent manner to suppress chemotherapy-induced intestinal toxicity. J Med Chem. 2017;60:9222-38.

44. Cheng K-W, Tseng C-H, Tzeng C-C, Leu Y-L, Cheng T-C, Wang J-Y, Chang J-M, Lu

Y-C, Cheng C-M, Chen I-J. Pharmacological inhibition of bacterial β-glucuronidase prevents irinotecan-induced diarrhea without impairing its antitumor efficacy in vivo. Pharmacol Res.

2019;139:41-9.

45. Bhat SV, Nagasampagi BA, Sivakumar M. Chemistry of natural products. New York:

Springer Science & Business Media; 2005.

46. Cooper R, Nicola G. Natural Products Chemistry: Sources, Separations and Structures.

New York: CRC Press; 2014.

47. Anulika NP, Ignatius EO, Raymond ES, Osasere O-I, Abiola AH. The chemistry of natural product: Plant secondary metabolites. Int J Technol Enhanc Emerg Eng Res. 2016;4:1-8.

48. Khan RA. Natural products chemistry: The emerging trends and prospective goals. Saudi

Pharm J. 2018;26:739-53.

49. Yuan H, Ma Q, Ye L, Piao G. The traditional medicine and modern medicine from natural products. Molecules. 2016;21:e559.

50. Dias DA, Urban S, Roessner U. A historical overview of natural products in drug discovery. Metabolites. 2012;2:303-36.

51. Patridge E, Gareiss P, Kinch MS, Hoyer D. An analysis of FDA-approved drugs: natural products and their derivatives. Drug Discov Today. 2016;21:204-7.

52. Pawlus AD, Kinghorn AD. Review of the ethnobotany, chemistry, biological activity and safety of the botanical dietary supplement Morinda citrifolia (noni). J Pharm Pharmacol.

2007;59:1587-609.

53. Kamiya K, Tanaka Y, Endang H, Umar M, Satake T. Chemical constituents of Morinda citrifolia fruits inhibit copper-induced low-density lipoprotein oxidation. J Agric Food Chem.

2004;52:5843-8.

54. Chan-Blanco Y, Vaillant F, Perez AM, Reynes M, Brillouet J-M, Brat P. The noni fruit

(Morinda citrifolia L.): A review of agricultural research, nutritional and therapeutic properties. J

Food Compost Anal. 2006;19:645-54.

55. Potterat O, Hamburger M. Morinda citrifolia (Noni) fruit-phytochemistry, pharmacology, safety. Planta Med. 2007;73:191-9.

56. Heinicke R. The pharmacologically active ingredient of Noni. Bull Pacific Trop Bot

Garden. 1985;15:10-4.

57. Dixon AR, McMillen H, Etkin NL. Ferment this: the transformation of Noni, a traditional

Polynesian medicine (Morinda citrifolia, Rubiaceae). Econ Bot. 1999;53:51-68.

58. McClatchey W. From Polynesian healers to health food stores: changing perspectives of

Morinda citrifolia (Rubiaceae). Integr Cancer Ther. 2002;1:110-20.

59. Aalbersberg WG, Hussein S, Sotheeswaran S, Parkinson S. Carotenoids in the leaves of

Morinda citrifolia. J Herbs Spices Med Plants. 1993;2:51-4.

60. Peerzada N, Renaud S, Ryan P. Vitamin C and elemental composition of some bushfruits. J Plant Nutr. 1990;13:787-93.

61. Shovic AC, Whistler WA. Food sources of provitamin A and vitamin C in the American

Pacific. Tropical Science. 2001;41:199-202.

62. Bui AKT, Bacic A, Pettolino F. Polysaccharide composition of the fruit juice of Morinda citrifolia (Noni). Phytochemistry. 2006;67:1271-5.

63. Simonsen J. Morindone. J Chem Soc. 1918;113:766-74.

64. Simonsen JL. LIX.—Note on the constituents of Morinda citrifolia. J Chem Soc.

1920;117:561-4.

65. Sang S, Ho C-T. Chemical components of noni (Morinda citrifolia L.) root. In: Wang M,

Sang S, Hwang LS, Ho C-T, editors. Herbs: Challenges in Chemistry and Biology. Washington,

DC: ACS Publications; 2006. p. 185-94.

66. Sang S, Wang M, He K, Liu G, Dong Z, Badmaev V, Zheng QY, Ghai G, Rosen RT, Ho

C-T. Chemical components in noni fruits and leaves (Morinda citrifolia L.). In: Ho C-T, Zheng

QY, editors. Quality Management of Nutraceuticals. Washington, DC: ACS Publications; 2002. p. 134-50.

67. Singh B, Sharma RA. Indian Morinda species: A review. Phytother Res. 2020;34:924-

1007.

68. West BJ, Palmer SK, Deng S, Palu AK. Antimicrobial activity of an iridoid rich extract from"Morinda citrifolia" fruit. Curr Res J Biol Sci. 2012;4:52-4.

69. Jayaraman SK, Manoharan MS, Illanchezian S. Antibacterial, antifungal and tumor cell suppression potential of Morinda citrifolia fruit extracts. Int J Integr Biol. 2008;3:44-9.

70. Kamata M, Wu RP, An DS, Saxe JP, Damoiseaux R, Phelps ME, Huang J, Chen IS. Cell- based chemical genetic screen identifies damnacanthal as an inhibitor of HIV-1 Vpr induced cell death. Biochem Biophys Res Commun. 2006;348:1101-6.

71. Krishnaiah D, Bono A, Sarbatly R, Anisuzzaman S. Antioxidant activity and total phenolic content of an isolated Morinda citrifolia L. methanolic extract from Poly-ethersulphone

(PES) membrane separator. J King Saud Univ Eng Sci. 2015;27:63-7.

72. Fletcher H, Dawkins J, Rattray C, Wharfe G, Reid M, Gordon-Strachan G. Morinda citrifolia (Noni) as an anti-inflammatory treatment in women with primary dysmenorrhoea: a randomised double-blind placebo-controlled trial. Obstet Gynecol Int. 2013;2013:e195454.

73. Palu AK, West BJ, Jensen J. Noni-based nutritional supplementation and exercise interventions influence body composition. N Am J Med Sci. 2011;3:552-6.

74. Horsfal A, Olabiyi O, Osinubi A, Noronha C, Okanlawon A. Anti diabetic effect of fruit juice of Morinda Citrifolia (Tahitian Noni Juice®) on experimentally induced diabetic rats.

Nigerian J Health Biomed Sci. 2008;7:34-7.

75. Saraswathi C, Prakash WS, Kunal P. Antiarthritic activity of Morinda citrifolia L. fruit juice in Complete Freund’s adjuvant induced arthritic rats. J Pharm Res. 2012;5:1236-9.

76. Saludes JP, Garson MJ, Franzblau SG, Aguinaldo AM. Antitubercular constituents from the hexane fraction of Morinda citrifolia Linn.(Rubiaceae). Phytother Res. 2002;16:683-5.

77. Masuda M, Murata K, Fukuhama A, Naruto S, Fujita T, Uwaya A, Isami F, Matsuda H.

Inhibitory effects of constituents of Morinda citrifolia seeds on elastase and tyrosinase. J Nat

Med. 2009;63:267-73.

78. Younos C, Rolland A, Fleurentin J, Lanhers M-C, Misslin R, Mortier F. Analgesic and behavioural effects of Morinda citrifolia. Planta Med. 1990;56:430-4.

79. Abou Assi R, Darwis Y, Abdulbaqi IM, Vuanghao L, Laghari M. Morinda citrifolia

(Noni): A comprehensive review on its industrial uses, pharmacological activities, and clinical trials. Arab J Chem. 2017;10:691-707.

80. Arpornsuwan T, Punjanon T. Tumor cell‐selective antiproliferative effect of the extract from Morinda citrifolia fruits. Phytother Res. 2006;20:515-7.

81. Clafshenkel WP, King TL, Kotlarczyk MP, Cline JM, Foster WG, Davis VL, Witt-

Enderby PA. Morinda citrifolia (Noni) juice augments mammary gland differentiation and reduces mammary tumor growth in mice expressing the unactivated c-erbB2 transgene. Evid

Based Complement Alternat Med. 2012;2012:e487423.

82. Gupta RK, Banerjee A, Pathak S, Sharma C, Singh N. Induction of mitochondrial- mediated apoptosis by Morinda citrifolia (Noni) in human cervical cancer cells. Asian Pac J

Cancer Prev. 2013;14:237-42.

83. Taşkın Eİ, Akgün‐Dar K, Kapucu A, Osanç E, Doğruman H, Eraltan H, Ulukaya E.

Apoptosis‐inducing effects of Morinda citrifolia L. and doxorubicin on the Ehrlich ascites tumor in Balb‐c mice. Cell Biochem Funct. 2009;27:542-6.

84. Masuda M, Itoh K, Murata K, Naruto S, Uwaya A, Isami F, Matsuda H. Inhibitory effects of Morinda citrifolia extract and its constituents on melanogenesis in murine B16 melanoma cells. Biol Pharm Bull. 2012;35:78-83.

85. Liu G, Bode A, Ma W-Y, Sang S, Ho C-T, Dong Z. Two novel glycosides from the fruits of Morinda citrifolia (noni) inhibit AP-1 transactivation and cell transformation in the mouse epidermal JB6 cell line. Cancer Res. 2001;61:5749-56.

86. Lv L, Chen H, Ho C-T, Sang S. Chemical components of the roots of Noni (Morinda citrifolia) and their cytotoxic effects. Fitoterapia. 2011;82:704-8.

87. Kamiya K, Hamabe W, Tokuyama S, Hirano K, Satake T, Kumamoto-Yonezawa Y,

Yoshida H, Mizushina Y. Inhibitory effect of anthraquinones isolated from the Noni (Morinda citrifolia) root on animal A-, B-and Y-families of DNA polymerases and human cancer cell proliferation. Food Chem. 2010;118:725-30.

88. Hirazumi A, Furusawa E. An immunomodulatory polysaccharide‐rich substance from the fruit juice of Morinda citrifolia (noni) with antitumour activity. Phytother Res. 1999;13:380-7.

89. Furusawa E, Hirazumi A, Story S, Jensen J. Antitumour potential of a polysaccharide‐ rich substance from the fruit juice of Morinda citrifolia (Noni) on sarcoma 180 ascites tumour in mice. Phytother Res. 2003;17:1158-64.

90. Prapaitrakool S, Itharat A. Morinda citrifolia Linn. for prevention of postoperative nausea and vomiting. J Med Assoc Thai. 2011;93:S204-S9.

91. Issell BF, Gotay CC, Pagano I, Franke AA. Using quality of life measures in a Phase I clinical trial of noni in patients with advanced cancer to select a Phase II dose. J Diet Suppl.

2009;6:347-59.

92. Okamoto H. Morinda citrifolia (Noni) in the treatment of psoriasis. Open General Int

Med J. 2012;5:1-2.

93. West BJ, Jensen CJ, Westendorf J. Noni juice is not hepatotoxic. World J Gastroenterol.

2006;12:3616-9.

94. Westendorf J, Effenberger K, Iznaguen H, Basar S. Toxicological and analytical investigations of noni (Morinda citrifolia) fruit juice. J Agric Food Chem. 2007;55:529-37.

95. Mueller BA, Scott MK, Sowinski KM, Prag KA. Noni juice (Morinda citrifolia): hidden potential for hyperkalemia? Am J Kidney Dis. 2000;35:310-2.

96. Millonig G, Stadlmann S, Vogel W. Herbal hepatotoxicity: acute hepatitis caused by a

Noni preparation (Morinda citrifolia). Eur J Gastroenterol Hepatol. 2005;17:445-7.

97. Stadlbauer V, Fickert P, Lackner C, Schmerlaib J, Krisper P, Trauner M, Stauber RE.

Hepatotoxicity of NONI juice: report of two cases. World J Gastroenterol. 2005;11:4758-60.

98. Yüce B, Gülberg V, Diebold J, Gerbes AL. Hepatitis induced by Noni juice from

Morinda citrifolia: a rare cause of hepatotoxicity or the tip of the iceberg? Digestion.

2006;73:167-70.

99. Adamovics JA. Regulatory considerations for the chromatographer. In: Adamovics JA, editor. Chromatographic analysis of pharmaceuticals. New York: Routledge; 2017. p. 1-17.

100. Coskun O. Separation techniques: chromatography. North Clin Istanb. 2016;3:156-60.

101. Zhang Q-W, Lin L-G, Ye W-C. Techniques for extraction and isolation of natural products: a comprehensive review. Chin Med. 2018;13:e20.

102. Décosterd LA, Dorsaz A-C, Hostettmann K. Application of semi-preparative high- performance liquid chromatography to difficult natural product separations. J Chromatogr A.

1987;406:367-73.

103. Hoffmann RW. Classical Methods in Structure Elucidation of Natural Products. Zürich,

Switzerland: John Wiley & Sons; 2018.

104. Todd L. Robert Robinson (1886–1975). Nat Prod Rep. 1987;4:3-11.

105. Bentley K. Sir Robert Robinson–his contribution to alkaloid chemistry. Nat Prod Rep.

1987;4:13-23.

106. Steinbeck C. Recent developments in automated structure elucidation of natural products.

Nat Prod Rep. 2004;21:512-8.

107. Barjat H, Morris GA, Swanson AG. A three-dimensional DOSY–HMQC experiment for the high-resolution analysis of complex mixtures. J Magn Reson. 1998;131:131-8.

108. Wang S, Li Y, Liu J, Fu Y, Liu D, Wang Y, Li L, Huo C, Zhang M, Shi Q. Historical story on natural medicine chemistry: Application of UV, IR, MS, and NMR spectra in structure elucidation of natural products. Zhong Cao Yao. 2016;47:2779-96.

109. IUPAC. Compendium of chemical terminology. Gold Book Version 2.3. 3 ed; 2014.

110. Mándi A, Kurtán T. Applications of OR/ECD/VCD to the structure elucidation of natural products. Nat Prod Rep. 2019;36:889-918.

111. Nguyen LA, He H, Pham-Huy C. Chiral drugs: an overview. Int J Biomed Sci.

2006;2:85-100.

112. Lin G-Q, Zhang J-G, Cheng J-F. Chiral drugs: Chemistry and biological action. In: Lin

G-Q, You Q-D, Cheng J-F, editors. New Jersey: John Wiley & Sons; 2011. p. 3-28.

113. Ling-Yi K, Peng W. Determination of the absolute configuration of natural products.

Chin J Nat Med. 2013;11:193-8.

114. Polavarapu PL. Determination of the absolute configurations of chiral drugs using chiroptical spectroscopy. Molecules. 2016;21:e1056.

115. Nugroho AE, Morita H. Circular dichroism calculation for natural products. J Nat Med.

2014;68:1-10.

116. Jensen F. Introduction to computational chemistry. New Jersey: John wiley & sons; 2017.

117. Grauso L, Teta R, Esposito G, Menna M, Mangoni A. Computational prediction of chiroptical properties in structure elucidation of natural products. Nat Prod Rep. 2019;36:1005-

30.

118. Goh GB, Hodas NO, Vishnu A. Deep learning for computational chemistry. J Comput

Chem. 2017;38:1291-307.

119. Jawiczuk M, Gorecki M, Masnyk M, Frelek J. Complementarity of electronic and vibrational circular dichroism based on stereochemical studies of vic-diols. Trends Analyt Chem.

2015;73:119-28.

120. Taniguchi T. Analysis of molecular configuration and conformation by (electronic and) vibrational circular dichroism: theoretical calculation and exciton chirality method. Bull Chem

Soc Jpn. 2017;90:1005-16.

121. Polavarapu PL. Vibrational spectra: principles and applications with emphasis on optical activity. The Netherlands: Elsevier; 1998.

122. Mandi A, Mudianta IW, Kurtan T, Garson MJ. Absolute configuration and conformational study of psammaplysins A and B from the Balinese marine sponge Aplysinella strongylata. J Nat Prod. 2015;78:2051-6.

123. Zhao H, Chen G-D, Zou J, He R-R, Qin S-Y, Hu D, Li G-Q, Guo L-D, Yao X-S, Gao H.

Dimericbiscognienyne A: A meroterpenoid dimer from Biscogniauxia sp. with new skeleton and its activity. Org Lett. 2017;19:38-41.

124. Waibel R, Benirschke G, Benirschke M, Achenbach H. Sesquineolignans and other constituents from the seeds of Joannesia princeps. Phytochemistry. 2003;62:805-11.

125. Nguyen P-H, Yang J-L, Uddin MN, Park S-L, Lim S-I, Jung D-W, Williams DR, Oh W-

K. Protein tyrosine phosphatase 1B (PTP1B) inhibitors from Morinda citrifolia (Noni) and their insulin mimetic activity. J Nat Prod. 2013;76:2080-7.

126. Takahasi H, Yanagi K, Ueda M, Nakade K, Fukuyama Y. Structures of 1, 4- benzodioxane derivatives from the seeds of Phytolacca americana and their neuritogenic activity in primary cultured rat cortical neurons. Chem Pharm Bull (Tokyo). 2003;51:1377-81.

127. Whitcomb DC, Lowe ME. Human pancreatic digestive enzymes. Dig Dis Sci. 2007;52:1-

17.

100

128. Zhang L, Hogan S, Li J, Sun S, Canning C, Zheng SJ, Zhou K. Grape skin extract inhibits mammalian intestinal α-glucosidase activity and suppresses postprandial glycemic response in streptozocin-treated mice. Food Chem. 2011;126:466-71.

129. Vlietinck A, Pieters L, Vander Berghe D. Bioassay-guided isolation and structure elucidation of pharmacologically active plant substances. In: Arnason JT, Mata R, Romeo JT, editors. Phytochemistry of Medicinal Plants. Boston: Springer; 1995. p. 113-35.

130. Tu Y. The discovery of artemisinin (qinghaosu) and gifts from Chinese medicine. Nat

Med. 2011;17:1217-20.

131. Tu Y. Artemisinin—a gift from traditional Chinese medicine to the world (Nobel lecture). Angew Chem Int Ed. 2016;55:10210-26.

132. Eloff JN. Quantifying the bioactivity of plant extracts during screening and bioassay- guided fractionation. Phytomedicine. 2004;11:370-1.

133. Senejoux F, Demougeot C, Kerram P, Aisa HA, Berthelot A, Bévalot F, Girard-Thernier

C. Bioassay-guided isolation of vasorelaxant compounds from Ziziphora clinopodioides

Lam.(Lamiaceae). Fitoterapia. 2012;83:377-82.

134. Chen C, Zhu H, Zhao D, Deng J, Zhang Y. Two new lignans from Phryma leptostachya

L. Helv Chim Acta. 2013;96:1392-6.

135. Seo KH, Lee DY, Jung JW, Lee DS, Kim YC, Lee YH, Baek NI. Neolignans from the fruits of Magnolia obovata inhibit NO production and have neuroprotective effects. Helv Chim

Acta. 2016;99:411-5.

136. Karunakaran C, Santharaman P, Balamurugan M. In: Karunakaran C, editor. Spin

Resonance Spectroscopy Principles and Applications. Oxford, UK: Elsevier; 2018. p. 49-110.

101

137. Edison AS, Schroeder FC. Modern Methods in Natural Products Chemistry. In: Liu H-

WB, Mander L, editors. Comprehensive Natural Products II: Chemistry and Biology. Oxford,

UK: Elsevier Science; 2010. p. 169-96.

138. Li X-C, Ferreira D, Ding Y. Determination of absolute configuration of natural products: theoretical calculation of electronic circular dichroism as a tool. Curr Org Chem. 2010;14:1678-

97.

139. Awolade P, Cele N, Kerru N, Gummidi L, Oluwakemi E, Singh P. Therapeutic significance of β-glucuronidase activity and its inhibitors: A review. Eur J Med Chem.

2020;187:e111921.

140. Taha M, Ismail NH, Imran S, Selvaraj M, Rahim A, Ali M, Siddiqui S, Rahim F, Khan

KM. Synthesis of novel benzohydrazone–oxadiazole hybrids as β-glucuronidase inhibitors and molecular modeling studies. Biorg Med Chem. 2015;23:7394-404.

141. Abdullah N, Taha M, Ahmat N, Wadood A, Ismail N, Rahim F, Ali M, Abdullah N,

Khan K. Novel 2, 5-disubtituted-1, 3, 4-oxadiazoles with benzimidazole backbone: a new class of β-glucuronidase inhibitors and insilico studies. Bioorg Med Chem. 2015;23:3119-25.

142. Salar U, Taha M, Ismail NH, Khan KM, Imran S, Perveen S, Wadood A, Riaz M.

Thiadiazole derivatives as new class of β-glucuronidase inhibitors. Biorg Med Chem.

2016;24:1909-18.

143. Kaneda N, Nagata H, Furuta T, Yokokura T. Metabolism and pharmacokinetics of the camptothecin analogue CPT-11 in the mouse. Cancer Res. 1990;50:1715-20.

144. Kehrer DF, Yamamoto W, Verweij J, de Jonge MJ, de Bruijn P, Sparreboom A. Factors involved in prolongation of the terminal disposition phase of SN-38: clinical and experimental studies. Clin Cancer Res. 2000;6:3451-8.

102

145. Bhatt AP, Pellock SJ, Biernat KA, Walton WG, Wallace BD, Creekmore BC, Letertre

MM, Swann JR, Wilson ID, Roques JR. Targeted inhibition of gut bacterial β-glucuronidase activity enhances anticancer drug efficacy. Proc Natl Acad Sci U S A. 2020;117:7374-81.

103

ABSTRACT

DEVELOPMENT OF SPECIFIC NATURAL BACTERIAL BETA-GLUCURONIDASE INHIBITORS FOR REDUCING IRINOTECAN-ASSOCIATED DIARRHEA

FEI YANG

December 2020

Advisor: Dr. Kequan Zhou

Major: Nutrition and Food Science

Degree: Doctor of Philosophy

ABSTRACT:

Irinotecan is a derived compound from the plant alkaloid camptothecin (CPT). It specifically inhibits eukaryotic DNA enzyme topoisomerase I. Irinotecan was approved by the

FDA as the second-line therapy for metastatic colon or rectal cancer in 1996. However, one of its leading side effects is diarrhea. It has been reported that up to 88% of treated patients were suffering from diarrhea and 31% of cases with grade 3 or 4 diarrhea. Irinotecan is metabolized into

SN-38G in the liver, then SN-38G is excreted to the intestinal tract and deconjugated back to SN-

38 by bacterial β-glucuronidase. The free SN-38 in the gut leads to the delayed diarrhea by damaging the intestinal mucosa. Therefore, selectively inhibiting bacterial β-glucuronidase has been an attractive strategy to alleviate irinotecan-induced delayed diarrhea. We preliminarily screened about 50 extracts and found that an extract of noni (Morinda citrifolia) fruits showed potent inhibitory activity on gut bacterial β-glucuronidase. In this study, four bacterial β-

104

glucuronidase inhibitors were obtained following bioactive assay-guided isolation, including two sesquineolignans, (7S,8S,7'R,8'R)-isoamericanol B (1) and americanol B (2), and two dineolignans, moricitrin A (3) and B (4). Compounds 2-4 are new, and the absolute configuration of compound

1 was determined for the first time. Their chemical structures were elucidated through HRESIMS and NMR spectra, and their absolute configurations were established via the comparison of the experimental and calculated electronic circular dichroism (ECD) spectra. These compounds showed potent inhibition against gut bacterial β-glucuronidase with IC50 values in the range 0.62-

6.91 µM. The inhibition presented specificity for β-glucuronidase as all the compounds showed no or weak effects on digestive enzymes such as α-amylase, α-glucosidase and lipase, suggesting that their gastrointestinal side effects could be minimized. These specific inhibitors as naturally occurring dietary compounds may be developed as promising candidates to alleviate irinotecan- induced diarrhea.

105

AUTOBIOGRAPHICAL STATEMENT Fei Yang EDUCATION 9/2016 - 12/2020, Ph.D. in Nutrition & Food Science, Wayne State University, Detroit, USA 9/2007 - 6/2010, M.S. in Medicinal Chemistry, China Pharmaceutical University, Nanjing, China 9/2003 - 7/2007, B.S. in Traditional Chinese Pharmacy, Shandong University of Traditional Chinese Medicine, Jinan, China

WORK EXPERIENCES 4/2015 - 12/2016, Research Assistant in the Department of Nutrition & Food Science, Wayne State University, Detroit, United States 5/2014 - 3/2015, Scientist in the Department of Discovery Chemistry, Hutchison MediPharma Limited, Shanghai, China 8/2010 - 4/2014, Laboratory Technician in the Department of Food Science & Engineering, Shanghai Jiao Tong University, Shanghai, China

AWARDS & HONORS 2020, Summer 2020 Dissertation Award, Graduate School, Wayne State University 2018-2019, Graduate Research Assistant Award, Graduate School, Wayne State University 2013, “Excellent” Annual Appraisal, Shanghai Jiao Tong University 2008 - 2009, Merit Student, China Pharmaceutical University

PUBLICATIONS 1. Yang, F., Zhu, W., Sun, S., Ai, Q., Edirisuriya, P., Zhou, K.* Isolation and Structural Characterization of Specific Bacterial β-Glucuronidase Inhibitors from Noni (Morinda citrifolia) Fruits. Journal of Natural Products, 2020, 83 (4): 825-833. 2. Yang, F.; Shi H. M.; Zhang X. W.; Yu, L. L.* Two Novel Anti-inflammatory 21-Nordammarane Saponins from Tetraploid Jiaogulan (Gynostemma pentaphyllum). Journal of Agricultural and Food Chemistry, 2013, 61: 12646-12652. 3. Yang, F.; Shi, H. M.; Zhang, X. W.; Yang, H. S.; Zhou, Q.; Yu, L. L.* Two new saponins from tetraploid jiaogulan (Gynostemma pentaphyllum), and their anti-inflammatory and α-glucosidase inhibitory activities. Food Chemistry, 2013, 141(4): 3606-3613. 4. Yang, F.; Chen, R.;* Feng, L.; Li, H. D.; Zhang, H. Two new alkaloids from the aerial part of Peganum nigellastrum. Helvetica Chimica Acta, 2011, 94(3):514-518. 5. Yang, F.; Chen, R.; Feng L.; Li, H. D.; Zhang, H.; Liang J. Y.* Chemical Constituents from the Aerial Part of Peganum nigellastrum. Chinese Journal of Natural Medicine, 2010, 8(3):199-201. 6. Yang, F.; Feng, L.; Li, H. D.; Zhang, H.; Chen, R.* A new flavone glycoside from the aerial part of Peganum nigellastrum. Chemistry of Natural Compounds, 2010, 46(4): 520-522.

POSTER PRESENTATION 1. Isolation and Structure Characterization of Specific Bacterial β-Glucuronidase Inhibitors from Noni (Morinda citrifolia L.) Fruit. The American Society of Pharmacognosy (ASP) Annual Meeting, July 13-17, 2019, Madison, Wisconsin, USA.