Chemical fingerprint analysis of fermented Morinda citrifolia L. (Noni) juice by UHPLC Q-TOF/MS combined with chemometric analysis

Morinda citrifolia L. (Noni) has been widely used in traditional medicine in tropical zones and has become increasingly popular globally owing to its health benefits. Most noni fruits are consumed as juice, which is traditionally produced by the natural fermentation of noni fruits. In this study, the metabolic profiles of noni fruit juice (NJ1) and fermented noni fruit juices (NJ2 and NJ3) was compared. A total of 74, 83, and 91 compounds including anthraquinones, coumarins, flavonoids, phenolic acids, phenolics, terpenoids, and miscellaneous (acids, carbohydrates, vitamins, fatty acids, etc.) were tentatively identified from NJ1, NJ2, and NJ3 in both positive and negative electrospray ionization modes. The phenolic compound composition differed significantly between noni juice and fermented noni juice. The results of the unsupervised principal component analysis and hierarchical clustering analysis showed that the non-fermented juice group clustered with the fermented juice groups. Asperulosidic acid, isoasperulosidic acid, and rutin levels were higher in the NJ1 group than those in the NJ2 group. Deacetylasperulosidic acid and monotropein contents in NJ2 were higher than those in NJ1. Similarly, NJ1 had higher asperulosidic acid and isoasperulosidic acid than those in NJ3. The findings from this study have the potential to enhance the quality of fermented noni juice.


Introduction
Noni (Morinda citrifolia L) is widely cultivated in tropical and subtropical regions such as Australia, Polynesia, Hawaii, and other Pacific islands.The fruits and leaves have been used in traditional folk medicine for the treatment of several diseases, such as high blood pressure, inflammation, and diabetes [1].Modern scientific research has shown that noni fruits possess antioxidant, anti-inflammatory, liver-protective, and immunomodulatory effects [2][3][4].This fruit contains various bioactive compounds, such as flavonoids, lignans, iridoids, coumarins, anthraquinones, polysaccharides, terpenoids, sterols, fatty acids, organic acids, vitamins, and minerals [2,5].Phenolic compounds play a key role in the therapeutic properties of plants; for example, rutin, β-sitosterol, asperuloside, and ursolic acid are important biologically active compounds present in noni [6].Flavonoids isolated from noni fruits, such as kaempferol, quercetin, catechin, and epicatechin have shown antidepressant and antioxidant activities [7].Previous studies have shown that fermentation can significantly affect the composition and bioactivity of phenolic compounds.For example, the phenolic composition of Dendrobium candidum substantially changes after fermentation, and the contents of syringic acid, 4-hydroxybenzoic acid, and p-hydroxycinnamic acid increase significantly [8].Traditionally, noni juice is produced by fermenting noni fruits in sealed jars or barrels for approximately 2 months or longer and then recovering the juice through drip extraction and/or mechanical pressure.Deng et al. reported that the phytochemical fingerprints of 13 commercial noni-fermented juices on the global market and found that they all contained scopolamine, rutin, and quercetin [9].
Fermentation is a natural process that converts sugars into products that are useful to humans using several microorganisms [10].In addition, fermentation can lead to increased chemical changes in organic substances through the action of enzymes [11].Lactic acid bacteria (LAB) are used in dairy and non-dairy foods such as yogurt, tea, and fruits.Recently, the demand for non-dairy probiotic products has increased because of an increase in their immune function [12].An increasing number of investigators are focusing on the biotransformation of phytochemical substances in vegetable and fruit juices using LAB to produce functional beverages [13].Vegetables and fruits are enriched in phenolics, organic acids, and sugars, which can be metabolized by LAB strains and therefore improve the sensory, nutritional, and functional qualities, as well as extend the shelf life of fermented products.Currently, the noni juice market is growing continuously because of its potential biological activities.Moreover, there are attempts to commercialize noni fruit through various processing methods such as fermentation and drying, and among them, fermented noni fruit juice stands out as the most popular in the market [14].However, there is very little information on the phytochemical components of fermented noni juice in the modern scientific literature.Therefore, the objective of this study was to explore an extensive investigation of the phytochemical composition of noni juice and fermented noni juice through the application of untargeted metabolomics.

Preparation of standard solutions
The individual compounds were dissolved in methanol to a concentration of 100 µg/mL.Mixed working standard solutions were prepared at concentrations of 1 µg/ mL and 0.01 µg/mL in ACN/IPA/water (3:3:2, v/v/v) from each stock solution.The stock solution of the internal standard was prepared by dissolving 0.4 mg labeled L-[ 13 C 6 ] phenylalanine in 1.0 mL of ACN/IPA/Water (3:3:2, v/v/v).All the standard solutions were stored in a refrigerator at -20 °C, and filtered through a 0.22-µm nylon syringe filter prior to analysis.

Preparation of sample solutions and quality controls
Noni juice samples (0.1 g) were accurately weighed into a 2-mL centrifuge tube and mixed with 1.5 mL of ACN/ IPA/water (3:3:2, v/v/v) containing 10 µg/mL of internal standards (L-phenylalanine-13 C 6 ).The mixture was then vortexed for 5 min and sonicated for 1 h in an ice bath.After sonication, the mixture was centrifuged at 12,298 x g for 10 min at 4 °C, the supernatant was passed through a 0.22-µm nylon syringe filter (Whatman, Maidstone, UK).After that, 990 µL of supernatant was transferred to a new centrifuge tube and 10 µL of internal standard solution (0.4 mg/mL) was added before analysis.All samples were prepared in quintuplicate.Quality control (QC) samples (n = 4) were prepared by pooling equal volumes of an aliquot of each sample extract and analyzing every five samples of the run.

Data processing
Data acquisition and processing were performed using Agilent MassHunter Qualitative Analysis software (Version 10.0, Agilent Technologies).Raw MS data files acquired from the UHPLC-Q-TOF/MS analysis were identified using the METLIN database B 08.00, and their authentic compounds.The amount of each metabolite obtained by UHPLC-Q-TOF/MS was determined as the relative metabolite abundance, which was calculated by dividing all data values for each sample by the chromatographic peak area of the internal standards added to the metabolite extract of each sample.

Chemometric analysis
Principal component analysis (PCA), hierarchical clustering analysis (HCA), partial least-squares discriminant analysis (PLS-DA), and orthogonal partial least-squares discriminant analysis (OPLS-DA) were performed using the SIMCA-P software package (Umetrics, Umea, Sweden).Unsupervised data analysis, including PCA and HCA, was applied as an exploratory data analysis to visualize the analytical connections among the samples.PLS-DA and OPLS-DA were used for the modeling of samples in the retention of settled classes Y.In the PLS-DA output, the variable important in projection (VIP) is an important screening index of metabolites that changed between different non-juice samples.The PLS-DA models were also validated using R 2 Y and Q 2 from a random permutation test (n = 200) in SIMCA-P.The quality of the PLS model was depicted by the cross-validation parameters R 2 and Q 2 , which represent the explained variance and predictive capability of the model, respectively.

Identification of bioactive compounds in noni juice and fermented noni juices
UHPLC-Q-TOF/MS was performed to analyze secondary metabolites in noni and fermented noni juice.As a result of our analysis, 130 compounds, including anthraquinones, coumarins, flavonoids, phenolic acids, phenolics, terpenoids, and miscellaneous com-pounds (acids, carbohydrates, vitamins, fatty acids, etc.) were tentatively identified from the samples (NJ1-NJ3) in both ESI + and ESI-modes (Fig. 1A and B).A total of 74, 83, and 91 compounds were detected in NJ1, NJ2, and NJ3, respectively.According to previous studies, anthraquinones, flavonoids, phenolic acids, phenolics, terpenoids, fatty acids, and carbohydrates are the major compounds in noni juice [15].

Metabolomic profiles of different noni juice samples
Untargeted UHPLC-Q-TOF/MS chemical fingerprinting coupled with chemometric analysis was performed to determine the comprehensive metabolite profiles of the different noni juice samples.The normalized data of NJ1, NJ2, and NJ3 in both positive and negative modes for both UHPLC-Q-TOF/MS are shown in Fig. 1.After data normalization, unsupervised PCA and HCA were performed using the normalized data of different noni juice samples to reveal the same sample clustering pattern, obtain an overview of the trend, and determine putative outliers (Fig. 2A and B).The PCA score plot showed a clear separation according to the three groups of noni juice samples including 'NJ1, ' 'NJ2, ' and 'NJ3' with no significant sample outliers.The plot was built using the first two principal components (PCs), PC1 and PC2, of the total variance (PC1 = 40.7%and PC2 = 34.1%).HCA was performed to elucidate the similarities and dissimilarities between the metabolite profiles of different samples.The HCA dendrogram indicated a clear separation between the non-fermented juice group (Group 1) and the fermented juice group (Group 2).These results imply that the major metabolite profiles of noni juice may be affected mainly by fermentation with the mixture of probiotics used in this study.
Supervised PLS-DA was performed to calculate models that differentiated between groups and to select the metabolites responsible for the different noni juice sample-dependent clustering based on VIP values > 1.
The PLS-DA score plot showed clear separation of the three clusters representing the different noni juice samples (Fig. 2C).The observed PLS-DA model had good coefficient fractions (p < 0.05), with R 2 Y = 0.989 and Q 2 = 0.979 (positive predictive ability), and the variables explained 0.749 (R 2 X) of the total variation.The y intercepts of R 2 and Q 2 in the permutation test were 0.187 and − 0.398, respectively, indicating a valid model (Fig. 2D).The recommended values for good fitting of models have been described as R 2 Y intercept < 0.3 and Q 2 Y intercept < 0.05.Differential metabolites were selected according to the criteria of p < 0.05 and VIP score > 1 from the PLS-DA model (Table 1).Hence, a combination of chromatographic fingerprints based on differential metabolite profiling with PLS-DA provides more comprehensive and insightful information regarding the dissimilarities between different noni juice samples.
Investigation of the OPLS-DA score plot and S-line plots was conducted to identify peaks that differed by sample type (Fig. 3).Within the S-line plots, iridoid compounds, flavonoids, and sucrose that showed substantial differences between each noni juice sample by matching them with an in-house library (Table SS4).In comparison between NJ1 and NJ2, the contents of asperulosidic acid (m/z 431.1202 [M ̶ H] ¯), isoasperulosidic acid (m/z 431.1199 [M ̶ H] ¯), and rutin (m/z 611.1612 [M + H] + ) were higher in NJ1 than NJ2, whereas the contents of deacetylasperulosidic acid (m/z 389.1099 [M ̶ H] ¯) and monotropein (m/z 391.1234 [M + H] + ) from NJ2 were higher than those from NJ1.Similarly, NJ1 had higher asperulosidic acid, and isoasperulosidic acid contents in NJ1 than those in NJ2.A previous study reported that the deacetylasperulosidic acid content was higher in fermented Morinda citrifolia L. extract (MCE) than that in non-fermented MCE, whereas the asperulosidic acid content was lower in fermented MCE than that in non-fermented extract [24].These results imply that asperulosidic acid may be acetylated during noni juice fermentation.A previous study also reported that an increase in the contents of polysaccharides, free anthraquinones, rubiadin, monotropein, and fructose in noni juice fermented by Bacillus sp.DU-106 and Lactobacillus plantarum [24].Dendrobium officinale fermented by Bacillus sp.DU-106 showed enhanced the immunostimulatory activity [26].In addition, monotropein induced immune activation in NCM460 cells [27].Noni juice fermented by LAB showed the enhancement of immune activities compared to nonfermented noni juice [28].It appears that changes in metabolite in noni juice caused by LAB fermentation could affect immune activity of noni juice.
Noni fruits are typically fermented before consumption due to their unpleasant odors and taste [29].In addition, commercially available noni products are processed in various ways by adding sugar, fruits, and condiments to reduce undesirable odor [30].Citrus fruits, which contain linalool and limonene-very floral and strong odors-can be added as additives to noni juices to mask flavor or eliminate odor [31].Adding fruit juices to noni juice is considered a convenient way to create value-added fruit drinks that excel in both sensory and nutritional quality.In this study, NJ3, made by adding coconut pollen sugar, calamansi juice, and maltodextrin, was used as a fermented noni juice to reflect commercial blended noni juice.NJ3 had higher contents of 3,5,7,4ʹ-tetramethoxyflavone (m/z 341.1055 [M ̶ H] ¯), 5-hydroxy-6,7,8,3ʹ,4ʹpentamethoxyflavone (m/z 373.1284 [M ̶ H] ¯), and sucrose (m/z 341.1093 [M ̶ H]¯) than those in NJ1 and NJ2 (Table SS4).Coconut blossom sugar is produced from the phloem sap of the blossom of the coconut palm tree (Cocos nucifera L.), which is heated until it begins to form a thick syrup [32].Sugar contains little fructose and has a lower glycemic index than conventional refined cane or beet sugars [33].Citrus fruits such as calamansi are rich sources of polymethoxylated flavones that could be responsible for their antioxidant, anti-inflammatory, and antiviral potentials [34][35][36][37].The high polymethoxyflavone and sucrose contents in NJ3 could be attributed to organic coconut blossom sugar and calamansi juice, respectively.However, significant variations in metabolite profiles are expected based on the specific ingredients added to noni juice, indicating the need for further research on metabolites in commercial noni juice formulations.
In conclusion, we confirmed the metabolic profiles of noni fruit juice and fermented noni fruit juices by using UHPLC-Q-TOF/MS.A total of 74, 83, and 91 compounds were identified in NJ1, NJ2, and NJ3, respectively.Untargeted UHPLC-Q-TOF/MS chemical fingerprinting coupled with chemometric analysis showed the comprehensive metabolite profiles of the different noni juice samples.These results can be used as basic information for developing products with fermented noni juice.Finally, further research is ongoing to improve the quality of noni juice according to fermentation conditions.

Fig. 1
Fig. 1 Classification information of differential metabolites identified between organic noni juice (NJ1) and fermented noni juices (NJ2 and NJ3) (A) and heatmap visualization of significant metabolites in positive mode (B).The horizontal bars of green, blue, and yellow describe the number of differential metabolites identified between noni juice and fermented noni juices.The color gradient of dark blue and deep red colors in the heat map indicates low and high intensities

Fig. 2
Fig. 2 Unsupervised score plots of the PCA model (A) and HCA dendrogram (B) and supervised score plot of the PLS-DA model (C).Two hundred permutation test (D) on PLS model from chemical fingerprinting of three noni juices

Table 1
Variable importance in projection (VIP) scores of partial least squares-discriminant analysis (PLS-DA)