Noni fruits collected in June 2019 from Wuzhishan, Hainan, China, had their pulps harvested at stage 5 (25 days pre-ripening). The fruits were processed and dried, stored in 161 g containers (Lot GH/T 1091, SCI10646900100011) for use in the study [1, 2, 18, 19]. See Figure 1A and B.

The extraction solvents, EtOH, ≥ 99.5% and MeOH, ≥ 99.5% were obtained from Sigma—Aldrich in Shanghai, China. Acetonitrile (AcN, ≥ 99.5%) was procured from Titan Scientific, also based in Shanghai, China. Formic acid (FA, ≥ 99%) and MeOH used for UPLC analysis were purchased from Merk, Shanghai, China. All the reagents used were of standard chromatography grade. The distilled water utilized in the process was treated using a Milli-Q system (Millipore Elix 3, Millipore, Shanghai, China), ensuring high purity [18, 19, 21–24].

The fungal strain of SC (Saccharomyces cerevisiae) and bacterial strains of EC (Escherichia coli), SA (Staphylococcus aureus), and ST (Streptococcus thermophilus) were obtained from the China Industrial Microbial Species Preservation and Management Centre in Shanghai, China. The EC, SA, and ST strains were cultured in Luria-Bertani (LB) medium, while the SC strain was cultured in yeast peptone dextrose (YPD) medium at 30°–37°C in a microbial incubator (Jinghong Equipment Co., Ltd, China). The LB medium was composed of 0.5% yeast extract, 1% peptone, and 1% sodium chloride (NaCl) (w/v), as per standard protocols. The YPD medium contained 1% yeast extract, 2% peptone, and 2% glucose (w/v). To prepare solid medium, 1.5% (w/v) agar was added to each medium, dissolved in distilled water, and then sterilized using an autoclave (Heng Technology Co., Ltd, China) [26–29]. The HT-29 (human colon adenocarcinoma cell), UMUC-3 (human bladder cancer cell), human umbilical vein endothelial cell (HUVEC), and normal human uroepithelium bladder cell line (SVHUC-1) were procured from the Institute of Cell Biology (Shanghai, China). The identity of the cell lines was confirmed through STR (Short Tandem Repeat) analysis and consistent with their known identities in databases of American type culture collection (ATCC), and International cell line authentication committee (ICLAC) for verification in our past studies (data are not shown) [18, 19, 30–36]. Cell lines were maintained in Dulbecco’s modified Eagle’s medium (DMEM) and Ham’s F-12K medium (mixture of DMEM and F-12) sourced from Gibco/Invitrogen, Camarillo, CA, USA. Additionally, Roswell Park Memorial Institute (RPMI) 1640 Medium was utilized, provided by Gibco, ThermoFisher, Shanghai, China. Each culture mediums were supplemented with 10% fetal bovine serum (FBS), obtained from PAN Biotech, Aidenbach, Germany, along with 1% antibiotics consisting of 10,000 U/mL penicillin and 10 mg/mL streptomycin from Solarbio Life Science, Beijing, China. The cells were incubated at a temperature of 37°C and a carbon dioxide (CO₂) concentration of 5% within a cell incubator [18, 19, 30–36].

The noni extracts were obtained through a process involving dissolvable maceration with a 1:4 ratio of water to solvent, aided by the KBE-I5 extractor processor (Shanghai Kunbo automation equipment Co., Ltd). See Figure 1C. Initially, 40 g of dry noni fruit grounds were soaked in 500 mL of hydroalcoholic solvents, either EtOH or MeOH, and preserved at 4°C for 120 h. The samples were then pulverized using the KBE-I5-extractor at a rotation speed of 1,000 rpm in dual cycle mode, with a rate of 1 min per cycle for 3 min. Following this, the sample was then centrifuged for 7 min at 3,000 rpm, removing precipitants. The supernatants were then filtered through a 0.45 µm organic filter twice to ensure purity. The filtered supernatants were poured into petri dishes and placed in a –80°C freezer for 48 h to freeze the samples. Subsequently, the frozen samples were subjected to freeze-drying using a vacuum-sealed chamber dryer equipped with a pump and refrigerator vapor trap system at –55°C for 24 h to concentrate the extracts. The equipment used for the extraction and concentration processes was from reputable manufacturers, including Shanghai Kunbo Automation Equipment Co., Ltd, for the KBE-I5 extractor processor and Thermo Scientific, USA, for the freeze dryer. The dry powder extracts, seen in Figure 1D and E, were stored in a sealed desiccator, protected from light, for future use. Prior to UPLC analysis, samples of both extracts were preconcentrated, filtered twice, and degassed using sonication. To optimize the auto-running sequence and achieve reliable results, multiple trial runs were conducted. All experimental data were recorded in triplicate to ensure accuracy and reliability. Stock samples (4 mL) were prepared from each extract at a concentration of 2 g/mL with 0.01% dimethyl sulfoxide (DMSO). Specifically, 8 g of each extract powder was dissolved in a mixture of DMSO and water at a ratio of 0.1:1,000 (v/v) [13, 14, 17–19].

The agar plate disc diffusion assay was conducted to assess the antibacterial activity, following the guidelines established by the Institute of Clinical and Laboratory Standards (CLSI) and the European Committee on Antimicrobial Susceptibility Testing (EUCAST) [18, 19, 37–39]. The EC, SA, and ST strains were first pre-incubated in 10 mL of LB, and the SC was pre-incubated in YPD liquid medium at 30°–37°C overnight. Sterilized agar culture medium (20 mL) was prepared and solidified onto petri dishes. A 200 μL aliquot of the individual strains’ standard inoculum was daubed onto the agar surface. The extract stock samples were serially diluted to the assigned test concentrations before being used in each test. A 30 μL aliquot of each pre-diluted extract sample was applied to filter paper discs placed on top of the inoculum agar surface, and the samples were then incubated overnight for 24 h. Kanamycin sulfate (ϫ = 10 mg/mL) was used as a positive control for the selected test strains. The MIC (minimum inhibitory concentration) of each extract was determined by measuring the inhibition zones [18, 19, 21, 26–29].

A total of 3 × 10⁴ cells of SVHUC-1 and HUVEC cell lines and 2 × 10⁴ cells of UMUC-3 and HT-29 were cultured in triplicate using 96-well plates at 37°C in 5% CO₂ level. The cells were incubated until they reached 80% confluence, and then the test groups were treated with NE and NM extracts at concentrations between 200 and 1,500 µg/mL (w/v). A 0.1% DMSO sample was used as a control. After 72 h, cell viability was assessed using the CCK-8 (cell counting kit-8) assay, 10 μL per well of reagent in accordance with the manufacturer’s instructions (Dojindo, Tokyo, Japan). The absorbance was read at 450 nm with a Genios multimode reader (Tecan GENios Pro, Tecan Group Ltd, Mannedorf, Switzerland). The IC₅₀ (half inhibitory concentration) was calculated using GraphPad Prism, indicating the toxicity of the extracts [30]. For the study on the morphology of HT-29 and UMUC-3 cell lines, the cells were cultured in a 24-well plate at a density of 8 × 10⁶ cells/well. Each group of extract samples was treated with its respective IC₅₀ values. After 72 h of treatment with NE and NM, the HT-29 and UMUC-3 cell samples, along with the control group, were rinsed three times using phosphate-buffered saline (PBS). The samples were resuspended in 1 mL of PBS and observed under white light using a Nikon A1R confocal laser scanning microscope, with an excitation wavelength of 488 nm and emission wavelength: 510–540 nm [18, 30, 34–36, 40–42].

Fluorescence staining with AO (acridine orange) stain (Sigma Aldrich, USA) was utilized to examine cell autophagy in HT-29 and UMUC-3 cell lines. The lysosomotropic dye AO emits a green/yellow to orange/red fluorescence in a pH-dependent manner, as it protonates to form aggregates that emit bright red fluorescence at low pH [40]. The cells were cultured in a 24-well plate at a density of 8 × 10⁶ to 12 × 10⁶ cells/well and treated with the respective IC₅₀ values of each test group for the NE and NM extract samples in triplicate. Following a 72 h treatment, the cancer cells were washed three times with PBS and then stained with 1 mL AO (1 μg/mL AO in PBS) for 30 min at 37°C in the dark. The samples were subsequently observed using a Nikon A1R confocal laser scanning microscope with an excitation wavelength of 488 nm and an emission wavelength of 510–540 nm at a scale of 20 μm [41].

The HT-29 and UMUC-3 cell lines were cultured in 6-well plates at a density of 120 × 10⁶ cells/well. These cells were then treated with the respective IC₅₀ values determined for each test group, specifically for NE and NM application, with all samples processed in triplicate. Following treatment, cell lysis and protein extraction, an essential step for subsequent western blotting analysis, were performed as previously described [33, 34, 40–42]. Following treatment with NE and NM extract compounds, the cell lines were subject to lysis using 200 mL lysis buffer that contains 1.19 g of 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 0.028 g magnesium chloride (MgCl₂), 50 mM tris (hydro methyl)-aminomethane buffer (Tris/pH 7.4), 100 mM NaCl, 5 mM disodium salt dyhydrate (Na/EDTA), 0.1% sodium dodecyl sulfate (SDS), 1% Triton X-100 and 2% glycerol. Supplements to the lysis buffer include: 0.1 M phenylmethylsulphonyl (PMSF) 1 µL, 1 M dithiothreitol (DTT) 10 µL, 1 M sodium 3-glycerophosphate (β-GP) 20 µL, 1 mM sodium vanadate (Na₃VO₄), and 10 µL protein inhibitor for every 1 mL of the lysis buffer. The protein concentration was determined using the Bradford assay with BSA as a standard. 50 μg protein extracted from the lysates was separated on 12% SDS-PAGE (polyacrylamide gel electrophoresis). After they were separated, they were transferred to polyvinylidene fluoride (PVDF) membranes (Buckinghamshire, UK) [24, 30, 33–36, 40, 41, 43]. The target proteins were blocked with 5% skimmed milk and then blotted with specific primary antibodies as follows: Beta Actin antibody (β-Actin, ab8227, Abcam, UK), and rabbit polyclonal antibody (LC3A I/II, 4599, Cell Signaling Technology, USA). Subsequently, the membrane was washed three times with tris-buffered saline-Tween 20 (TBST) solution and incubated with goat anti-rabbit IgG horseradish peroxidase (HRP)-coupled secondary antibodies (111-036-003, Jackson Immuno Research Laboratories, USA). The enhanced chemiluminescent (ECL) protein imprint detection kit (Tian Gen Biotech, Shanghai, China) was used to detect the immunoreactive proteins, and the signal density was measured using the Tanon 6200 (Tanon Science & Technology Co., Ltd, Shanghai, China) [33–36, 39, 40].

The UPLC-Q-exactive orbitrap-MS/MS analysis was conducted using a Nexera UPLC system. The system is equipped with a column oven, sample management with a flow-through needle, binary solvent management, a UV spectroscopy detector, and two individual distribution pumps (Shimadzu Corporation, Japan). A Fourier transform ion cyclotron resonance mass spectrometer (FTMS) with magnetic mass spectroscopy and an electrostatic field mass spectrometer, both supported by an electrospray ionization (ESI) source, is located in the Q-exactive orbitrap. The fragmentation data-dependent spectra were scanned using a low collision-induced dissociation (LCID) and a high collision-induced dissociation (HCID) energy source, as well as ESI sources, all of which were supplied by ThermoFisher Scientific, USA. The mixture was fractionated on a Waters Acquity UPLC HSS-T3 end-capped column (100 mm × 2.1 mm, 1.8 µm, Waters, Ireland). The temperature of both the column and the sample was maintained at 45°C. To optimize the auto-run procedures and results, a 0.1 mg/mL (w/v) extract sample from each NE and NM extract was diluted with MeOH, filtered, degassed, and then used for the analytical tests. An injection volume of 3 µL was injected into the system mobile phase flow rate set at 0.35 mL/min. Eluate analysis was monitored at 365 nm UV within a range of 200–400 nm [18, 19, 21–24]. The two mobile phases were filtered through a 0.22 µm organic filter and degassed by sonication for 1 h prior to use. The linear elution gradient was configured as follows: an initial hold at 5% B from 0.01 to 2 min, 5% B (2–2.5 min), followed by another 5% B from 2 to 2.5 min, then a linear increase to 15% B from 2.5 to 7 min, further increase to 55% B from 7 to 9.5 min, then to 90% B from 9.5 to 12 min, held at 90% B from 12 to 13 min, and finally 5% B from 13–14 min. The instrument control, data processing, and recording were performed using Xcalibur^® data software (Version 4.2.47, ThermoFisher Scientific, USA) [18, 23, 24]. In both positive (ESI+) and negative (ESI–) settings, the ionic fragments that controlled the UPLC system remained stable, showing increased intensity and a well-defined peak. Consequently, ESI was utilized for continuum data acquisition with a scan time of 0.2 sec over a mass-to-charge range of 50–1,000 m/z in both ESI+ and ESI– modes. To assess the mass mode, dual mass acquisition mode scan functions were employed. Specifically, in the trap cell, the LCID scan was set to a collision voltage of 10 V, whereas in the transfer cell, the HCID scan was configured with a collision voltage ranging between 20 and 60 V. The mass spectra acquired during the full range acquisition encompass ESI source parameters and ion scanning at a normalized HCID collision energy for fragmentation, which is dependent on a data-dependent scan mode. Nitrogen gas was used as both the collision gas and the source of ionization. Additional parameters used include a desolvation gas temperature of 550°C, a source temperature of 120°C, and a capillary voltage of 1.5 V; others, such as gas flow, auxiliary gas flow rate, capillary temperature, and cone voltage. The auxiliary gas flow rate is set at 10 arbitrary (Arb) units for both ESI+ and ESI– modes. The capillary temperature is maintained at 320°C; the cone voltage is set to 3.8 kV. The full scan resolution is 70,000 m/z, and the HCD-MS/MS (high energy collision C-trap dissociation tandem mass spectrometer) resolution is 17,500 m/z, with a mass scan range between 100 to 1,200 m/z. The sheath gas flow rates are 40 and 35 Arb units for ESI+ and ESI– modes, respectively, and the spray voltage is set at 3.5 kV for ESI– and –3.0 kV for ESI– [18, 19, 21–24].

Xcalibur was used to acquire raw data from the analyses conducted in full-scan mode. The Metabolomics Innovation Center (TMIC) provided quantitative metabolomics services for biomarker discovery and validation kits, utilizing a chemical reference library of standards generated from the HMDB (Version 5.0) and FOODB (Version 1.0). This library assists in the identification of confirmed compounds. Peak selection, adduct deconvolution, and feature annotation were employed to process the data. For metabolite annotation, a tolerance of 5 ppm was applied to both precursors and fragment mass, and an isotopic similarity threshold of over 60% relative to the HMDB was used to verify metabolite identities. Additionally, the cosine similarity algorithm integrated into Mass Bank was utilized to compare the calculated tandem MS/MS fragmentation spectra with corresponding reference spectra. This cosine similarity method assigns fragmentation scores, ranging from 0% (indicating a poor match) to 100% (indicating an excellent match), based on how closely the evaluated MS/MS spectrum aligns with the reference spectra. Due to non-matching or missing peaks in the evaluated tandem mass spectrum compared to the referenced spectrum, the overall score-which was initially 30, can vary between 60% in a worst-case scenario and 100% in the best-case scenario. To evaluate both the overall score and metabolite identification, five different properties are considered for each feature: isotope distribution, mass, retention time (RT, t_r = min), collisional cross section (CCS) value, and fragmentation score [18, 19, 21–24]. Each of these individual scores ranges from 0 to 100, where any property that does not match the reference data is assigned a score of 0. The overall score is then calculated as the mean of these five values. When discrepancies such as missing or non-matching peaks are observed between the evaluated dual mass spectrum and its reference, adjustments to the overall score occur. For instance, although an initial overall score 30% might have been recorded, these discrepancies can result in a score of 60% under the worst-case scenario or even 100% under optimal conditions. To assess both the overall score and the accuracy of metabolite identification, each feature is evaluated using 5 key properties: isotope distribution, mass accuracy (mass quantity), RT, CCS value, and fragmentation score. Each property is scored on a 0 to 100 scale, with a score of 0 being assigned if a property does not match the corresponding reference data. The final overall score is calculated as the mean of these five individual scores, and the compounds are ultimately selected based on this comprehensive scoring process combined with fragmentation analysis. The identity of each confirmed compound was determined to be more reliable with higher scores and fragmentation scores when searching the MS/MS spectrum against the HMDB, a comprehensive library that includes metabolites from various mixtures with referenced LC-MS/MS data and predicted nuclear magnetic resonance (NMR) spectra. This library offers precise predictions of mass, retention indices, and collision cross-section data, thereby enhancing compound identification [18]. The hyperlinked databases of HMDB, FOODB, the Small Molecule Pathway Database (SMPDB), Drug Bank, and the Toxin and Toxin Target Database (T3DB) provide valuable support for the efficiency and reliability of phytocompound identification [23, 24, 44]. The dataset comprised a comprehensive array of information on compound identification, including each compound’s ID, adducts, molecular formula, score, fragmentation score, mass error (within 5 ppm), isotope similarity (greater than 60%), theoretical isotope distribution, a web link directing users to the HMDB and FOODB, a detailed compound description, RT (min), and m/z values. For each individual compound, those numbered in order of elution RT were selected based on their high fragmentation scores, as well as their overall scores and isotopic similarity scores. Furthermore, the mass ratio (m/z) detected in the database was evaluated in Excel data format, and the mass error was assessed to be within 5 ppm [18].

All values were given as a mean ± SD, and the data analyses were conducted using GraphPad Prism software (version 8.2.0) with the Microsoft Excel 2017 data analysis add-in. A Student’s t-test with a significance level of ***P < 0.001 was used to identify statistically significant differences between the test groups and the control group. The experiments were repeated three times, and the data were presented as the mean ± SD (for each group, n = 3). The chemical structures of the phytocompounds were drawn using the ChemDraw software (Revvity Signals Software, Inc., 2023).

The net weight of the NE extract was 6.72 g in Figure 1D, while that of the NM extract was 10.30 g, as seen in Figure 1E. These values were obtained after processing. The extrapolated recovery rates for both extracts were 16.8% for NE and 25.8% for NM, respectively.

The antibacterial efficacy of the substance was evidenced by the formation of a clearly discernible inhibition zone in the disk diffusion assay [18, 19]. Seen in Figure 2, the NE extract versus A: EC, B: SC, C: SA, and D: ST. For the extract NM against strains in E: EC, F: SC, G: SA, and H: ST. The inhibition zone diameters (mm) were assessed, and the MIC values of each extract on the tested strains were also estimated. For the NE extract, inhibition zone diameters (mm) corresponding to the MIC values were recorded as follows: EC (10.06 ± 0.08), SC (11.15 ± 0.21), SA (10.09 ± 0.14), and ST (10.05 ± 0.09). For the NM extract, the inhibition zones were as follows: EC (11.75 ± 0.07), SC (10.63 ± 0.15), SA (9.84 ± 0.17), and ST (9.65 ± 0.10). The MIC value of the NE extract against EC strain is 177 mg/mL, SC 52 mg/mL, SA 388 mg/mL, and ST 283 mg/mL. The MIC of the NM extract against EC strain is 105 mg/mL, SC 47 mg/mL, SA 312 mg/mL, and ST 135 mg/mL, respectively.

Using the CCK-8, the cytotoxicity of NE and NM extracts on HT-29, UMUC-3, HUVEC, and SVHUC-1 cell lines was evaluated. The data showed that both extracts reduced the viability in Figure 3, A: HT-29 and C: UMUC-3 cells after a 72 h test period. The cell growth of B: HUVEC and D: SVHUC-1 was exponentially normal after the test, similar to NE and NM extracts. The IC₅₀ extrapolated for each extract showed NE with a value at 758 µg/mL and NM valued at 1,231 µg/mL against HT-29. The NE value at 899 µg/mL and the NM value at 1,173 µg/mL against UMUC-3. Additionally, microscopic observations of cell organelles morphology after IC₅₀ treatment groups of each extract after 72 h under white light showed deformed and irregular-shaped cellular structures in HT-29 and UMUC-3. Seen in Figure 4 are the control groups of A to D, and the treated groups are E to H.

The AO fluorescent staining was used to investigate the effect of NE and NM extracts on HT-29 and UMUC-3 cells, in which the test group with extract was investigated for the induce and accumulation of autophagy vesicles observed under the cancer cell lines by the florescence emission in Figure 5. As a result of cell death, based on the decrease of cell viability there autophagic bodies were observed among the treated group seen in Figure 5E and F, inflicted by compounds NE and NM compared to the control groups Figure 5A and B for the HT-29 cell. There were also autophagic bodies were also observed in the treated group in Figure 5G and H by extract NE and NM in comparison to the control groups, Figure 5C and D for the UMUC-3. In addition, the effect of the extract NE and NM on autophagy-related proteins in HT-29 and UMUC-3 cells was then examined using western blot in Figure 6A. The ratio of LC3A II/I in cells treated with both extracts individually was clearly expressed and evaluated in relation to those in their control groups, see Figure 6B.

Both NE and NM extract chromatogram results showed the presence of compounds by several peak counts at RT against relative abundance, and both in selective ion polarized mode scans, according to the selective ions and elution order obtained from the UPLC-Q-exactive orbitrap-MS/MS analysis graphs seen in Figures 7 and 8. A mass spectrometer with dual mass-to-charge ratio scan, a primary mass score range (max 60%), and a secondary fragmentation maximum score (max 100%) was acquired to identify all compounds. As seen in Figure 7A (top ESI–) and Figure 7B (bottom ESI+), base peak intensity (BPI) graphs showed migration time and abundance relative intensity (AUD) of individual compound peaks. NE extract showed 10 peaks, and they were identified by searching the chemical reference library of standards from the FOODB and HMDB databases, as listed in Table 1. The HMDB database hyperlinks databases of the SMPDB, Drug Bank, the T3DB, and the FOODB, aiding efficiency and reliability in identifying each compound [44, 45].

Herein, the identified compounds in NE and NM extract with chemical formular (CF) and molecular weight (MW, g·mol^⁻¹) as follows. Peak 1 (t_r = 0.89) Gentiotriose (CF: C₁₈H₃₂O₁₆, MW: 504.4371), also known as isomaltotriose, is an oligosaccharide and has been detected in fruits [45]. Peak 2 (t_r = 1.69) Wyomin (CF: C₃₃H₄₀O₂₀, MW: 756.6587), belongs to the class of flavonoid-7-O-glycosides compounds. It was one of the phenolic compounds that have been detected in several different foods, such as barley (Hordeum vulgare), breakfast cereal, cereals, cereal products, common wheat (Triticum aestivum), and kinds of wheat (Triticum) [45]. Peak 3 (t_r = 3.25) 1-O-Sinapoylglucose (CF: C₁₇H₂₂O₁₀, MW: 386.3506) belongs to the class of organic compounds known as hydroxycinnamic acid glycosides and has been detected in a few different foods, such as brassicas, Brussel sprouts (Brassica oleracea var. gemmifera), and other green vegetables [45]. Peak 4 (t_r = 4.17) Edulisin VI (CF: C₂₀H₂₂O₉, MW: 406.3833) belongs to the class of organic compounds known as furanocoumarins, and it has been detected in green vegetables [45]. Peak 5 (t_r = 4.57) 1-Octen-3-yl primeveroside (CF: C₁₉H₃₄O₁₀, MW: 422.4673) also seen off NM, was known as fatty acyl glycosides of mono- and disaccharides [45]. Peak 6 (t_r = 0.79) D-1-[(3-Carboxypropyl) amino]-1-deoxyfructose (CF: C₁₀H₁₉NO₇, MW: 265.2604) is a gamma-amino acid and derivative compound and has been detected in fruits [45]. Peak 7 (t_r = 0.89) Umbelliferose (CF: C₁₈H₃₂O₁₆, MW: 504.4371), also identified at (t_r = 4.16) of NM extract, belongs to the class of organic compounds known as O-glycosyl compounds, and it has been detected in several different foods, such as carrots (Daucus carota ssp. sativus), fats and oils, green vegetables, herbs, and spices [45]. Peak 8 (t_r = 1.66) 4-Hydroxycinnamic acid (CF: C₉H₈O₃, MW: 164.158), also known as p-Coumaric acid, which was also seen to appear at (t_r = 3.23) in the NM sample. It was a coumaric corrosive and plays a significant part as a plant metabolite. On average, this compound has been found in the highest concentration in pineapples, sunflowers, and pepper (Capsicum frutescens), and in a lower concentration in spinach, kiwis, and sweet oranges [45]. Importantly, this compound was reported to show promising anticarcinogenic potential for the development of new antitumor agents, adhering to the trend of drugs with greater tolerance and biological effectiveness and the synthesis of derivatives with good antiparasitic profiles [46]. Adiposity caused by a high-fat diet, non-alcoholic fatty liver disease, other metabolic disturbances, and acute lung injury are all ameliorated by the compound’s anti-inflammatory and antioxidant properties [47]. Peak 9 (t_r = 3.22) 2-Hydroxycinnamic acid (CF: C₉H₈O₃, MW: 164.158), also known as O-coumaric acid, has been found in a few different foods, such as corns, hard wheats, and olives and in a lower concentration in pomegranates, cranberries, peanuts, carrots, soybeans, ryes, rye bread, and turmeric [45]. Moreover, it was shown to be an antioxidant [48] and a potential among phenolic acids used as effective in the inhibition of antibiotic-resistant pathogenic bacteria [49]. Peak 10 (t_r = 4.16) Inulobiose (CF: C₁₂H₂₂O₁₁, MW: 342.2965), also known as difructan or inulin, belongs to the class of C-glycosyl compounds.

NM extract, as seen in Figure 8A (top ESI–), and Figure 8B (bottom ESI+), are both the BPI graph, which showed 11 identified compound peaks, of which 4 were identical and were present in NE analyzed extracts. The 7 remaining unique compounds present in the NM extracts were listed and categorized accordingly in Table 2. Peak 11 (t_r = 0.86) Sakebiose (CF: C₁₂H₂₂O₁₁, MW: 342.2965), also known as nigerose, belongs to O-glycosyl compounds. It has been detected in fruits [45]. Peak 12 (t_r = 1.73) Riboflavin (CF: C₁₇H₂₀N₄O₆, MW: 376.3639) or vitamin B2 is an easily absorbed, water-soluble micronutrient with a key role in maintaining human health [50]. The disparate findings of numerous studies attempting to link B vitamins to the development of cancer are clarified by previous studies, which demonstrated that riboflavin plays an evident role in immune cell regulation [51]. Interestingly, a recent report mentioned that riboflavin is a safer option for critically ill COVID-19 patients, especially those in high-risk categories, established as an alternative [52]. It was known as an antioxidant and anti-inflammasome that regulates the inflammatory response [53]. Riboflavin deficiency offered new insights into the role in esophageal carcinogenesis [54]. Peak 13 (t_r = 0.78) N-(1-Deoxy-1-fructosyl) alanine (CF: C₉H₁₇NO₇, MW: 251.2338) is a class of organic compounds known as C-glycosyl compounds, and it is detected in green vegetables and root vegetables [45]. Peak 14 (t_r = 0.87) is β-D-Galactopyranosyl-(1->3)-β-D-galactopyranosyl-(1->6)-D-galactose (CF: C₁₈H₃₂O₁₆, MW: 504.4371), which is known as an oligosaccharide and has been detected in fruits [45]. Peak 15 (t_r = 1.62) 1-Oxo-1H-2-benzopyran-3-carboxaldehyde (CF: C₁₀H₆O₃, MW: 174.1528), also known as 3-formylisocoumarin or artemidinal, belongs to the class of iso-coumarins and derivatives [45]. Peak 16 (t_r = 5.88) 1-(β-D-Glucopyranosyloxy)-3-octanone (CF: C₁₈H₃₂O₁₆, MW: 504.4371) is a fatty acyl glycoside of mono- and disaccharides [45]. Peak 17 (t_r = 6.46) Armillaripin (CF: C₂₄H₃₀O₆, MW: 414.4914) belongs to the class of organic compounds and is known as melleolides and analogues of lipids [45].

Using UPLC-Q-exactive orbitrap-MS/MS, we identified a total of 17 compounds from the alcoholic extracts of Hainan dry noni fruit. They are, respectively, coumarins, flavonoids, phenolic, oligosaccharides, flavonoids of both glycosides’ substrates, namely O-glycosyl and C-glycosyl, coumaric acids, phenol lipids, melleolides, and fatty acyl glycoside compounds [44, 45]. To confirm their individual molecular structures and other physical and chemical properties from the HMDB and FOODB databases, the compounds were cross-referenced to the National Institute of Standards and Technology (NIST) and Open Chemistry database-National Library of Medicine-National Center of Biotechnology Information (PubChem-NCBI) of the identifiers. Herein, the identified compounds in NE and NM extract are presented with their molecular structures: compounds (1) to (7) in Figure 9, compounds (8) to (17) in Figure 10. Among the 17 compounds, other 14 phytocompounds were identified from noni fruit for the first time, except 4-Hydroxycinnamic acid (peak 8), 2-Hydroxycinnamic acid (peak 9), and Riboflavin (peak 12).