Detailed investigation of the composition and transformations of phenolic compounds in fresh and fermented Vaccinium floribundum berry extracts by high‐resolution mass spectrometry and bioinformatics

Abstract Introduction Blueberries are known for their very high content of biologically active phenolic compounds; nonetheless, differently from the North American and European species of blueberries, Neotropical blueberries have not been extensively studied yet. Objectives In the present paper, the phenolic composition of Vaccinium floribundum Kunth, which is endemic to the Andean regions and grows 1,600 to 4,500 meters above sea level, was investigated by ultra‐high‐performance liquid chromatography coupled to high‐resolution mass spectrometry (UHPLC‐HRMS). Native and fermented berries were compared in terms of phenolic composition as well as antioxidant activity, total phenolic content, and total anthocyanin content. Materials and Methods V. floribundum native and fermented berries were extracted and analyzed by UHPLC‐HRMS. The acquired datasets were processed by Compound Discoverer 3.1 using a dedicated data analysis workflow that was specifically set up for phenolic compound identification. Results In total, 309 compounds were tentatively identified, including anthocyanins, flavonoids, phenolic acids, and proanthocyanidins. The molecular transformations of phenolic compounds during fermentation were comprehensively investigated for the first time, and by a customized data processing workflow, 13 quinones and quinone methides were tentatively identified in the fermented samples. Compared to other species of the genus Vaccinium, a peculiar phenolic profile is observed, with low abundance of highly methylated compounds. Conclusion Andean berries are a rich source of a wide variety of phenolic compounds. Untargeted MS analyses coupled to a dedicated data processing workflow allowed expanding the current knowledge on these berries, improving our understanding of the fate of phenolic compounds after fermentation.

Conclusion: Andean berries are a rich source of a wide variety of phenolic compounds. Untargeted MS analyses coupled to a dedicated data processing workflow allowed expanding the current knowledge on these berries, improving our understanding of the fate of phenolic compounds after fermentation.

| INTRODUCTION
Increased intake of fruits like berries, rich in nutrients and phytochemicals, is recommended in dietary guidelines for their beneficial health effects. 1 Consumption of berry fruits is not only limited to fresh or frozen forms, as several processed and derived products are prepared, such as dried and fermented berries, yogurts, beverages, and jams. 2 Moreover, in recent years, berry extracts have been increasingly employed as functional food and dietary supplements combined with other vegetable and herbal extracts. 3 Raspberry extracts have been demonstrated to exert antineuroinflammatory effects, 4 while anthocyanin-rich strawberry extracts were shown to protect human dermal fibroblasts against oxidative damage. 5 Berries belonging to the genus Vaccinium, such as European blueberry (also known as bilberry or huckleberry, Vaccinium myrtillus) and North American blueberry (Vaccinium corymbosum), have received increasing interest due to their extremely high content of flavonoids, anthocyanins, phenolic acids, and tannins, which have been demonstrated to exert a wide range of biological activities. 6 In a recent paper by Rutledge et al., 7 blueberry phenolics were associated with cognitive enhancement in healthy older adults. Likewise, Stull 8 reported that consumption of whole blueberries reduces the blood glucose level in vivo. For these reasons, blueberry is often referred to as a "superfruit." The composition of phenolic compounds in North American and European blueberries has been widely studied. 9, 10 Ancillotti et al. 11 reported more than 200 compounds in blueberry hydroalcoholic extracts, comprising mainly anthocyanins, flavonols, and proanthocyanidins (PACs), and described the differences among V. myrtillus, V. corymbosum, and Vaccinium uliginosum. Other than V. myrtillus, there are several other lesser-known wild species of the genus Vaccinium, such as Vaccinium floribundum Kunth, known with the trivial names of Andean blueberry or mortiño. V. floribundum is a woody perennial shrub that is endemic to the Andean region in South America, ranging from Venezuela to Bolivia, and can be found between 1,600 to 4,500 meters above sea level. 12 Mortiño is known to play a significant environmental and ecological role, being one of the first species that recover after bouts of deforestation and human-made fires. 12 Moreover, fruits have high concentrations of phenolic compounds with potential beneficial effects on human health and are widely used by the local population in native and fermented form or for the preparation of traditional drinks, ice creams, preserves, and wines. 13 Despite the growing interest in the bioactive compounds in berries from South America, 14 there is still a lack of knowledge on the phenolic compound composition of V. floribundum. The few papers dealing with the identification of phenolic compounds are limited to specific subclasses and use low-resolution techniques. [15][16][17][18] In the present article, V. floribundum Kunth from the Peruvian Andean Region was characterized by high-resolution mass spectrometry (HRMS), which is the foremost technique for untargeted analysis of phenolic compounds. 19 Moreover, as wild edible berries are often fermented before consumption to enhance phenolic compounds' bioavailability, 20 V. floribundum berries were subjected to lactic fermentation by Lactobacillus plantarum to compare its biological activities to those of the native berry. Moreover, the transformations of the phenolic molecular species were comprehensively evaluated for the first time. For this purpose, a semi-automated data processing workflow for the identification of phenolic degradation products was set up.

| Fermentation process
In total, 166 g of frozen V. floribundum Kunth berries was disinfected with a diluted sodium hypochlorite solution and washed three times with distilled water. The berries were added to 500 mL of ultra-pure water, liquefied, and placed in an amber bottle. Sterilization of the berries was performed by the thermal shock method: the homogenized material was subjected to a temperature of 70 C for 10 min,

| Determination of the antioxidant activity
The ABTS antioxidant assay was carried out as previously described with slight modifications. 22 ABTS was dissolved in water to a concentration of 7 mM, and potassium persulfate was dissolved in water to a concentration of 2.45 mM. The stock solutions were mixed in a 1:1 (v/v) ratio and kept at room temperature for 12-16 h in the dark to prepare the ABTS reaction solution. A volume of 2.8 mL was diluted to 65 mL in acetate buffer at pH 4.5 to obtain the ABTS working solution. The absorbance was measured at 734 nm. Trolox was used as the standard and distilled water as the blank control. The results are expressed as μmol of Trolox equivalents (TE) per g of dry weight (dw) of the sample. Details are reported in Table S1.

| Total phenolic content
The total phenolic content (TPC) was determined using Folin-Ciocalteu reagent as previously described with minor modifications. 23 Briefly, berry extracts (10 μL + 490 μL H 2 O) were reacted with 1 N Folin-Ciocalteu reagent (250 μL, for 5 min) and then neutralized with 1.2 N sodium carbonate (1.25 mL). After 30 min, the absorbance of the resulting solution was measured at 755 nm. Gallic acid was used as the standard. The TPC is expressed as mg of gallic acid equivalents (GAE) per g dw of the sample. Details are reported in Table S2.

| Total anthocyanin content
The total anthocyanin content (TAC) was determined by the previously reported pH differential method. 24 Briefly, two different dilutions were prepared by diluting 1 mL of each extract to 10 mL: one at pH 1 with potassium chloride buffer and the other at pH 4.5 with sodium acetate buffer. The absorbance of the samples was measured at 510 and 700 nm (to correct for haze) against a blank sample consisting of ultra-pure water. The TAC is expressed as mg of cyanidin 3-glucoside equivalents per g of sample. Therefore, the maximum absorbance of cyanidin 3-glucoside (510 nm) and its molar absorptivity (ε = 26,900) were employed. Details are reported in Table S3.

| UHPLC-HRMS analysis
Phenolic compound chromatographic separation and MS analysis were carried out using a previously reported platform based on RP separation using a Kinetex core-shell C 18 column (100 mm Â 2.1 mm i.d.) and MS analysis with a TOP 5 data-dependent acquisition (DDA) mode for both low-and high-molecular-weight phenolic compounds.
Details are provided in the Supporting Information. All samples were run in triplicate.

| Phenolic compound identification
Raw data obtained from three consecutive injections and the blank sample were processed by Compound Discoverer 3.1 (Thermo Fisher Scientific) using a customized method specifically dedicated to phenolic compound analysis. 25 Details are provided in the Supporting Information. The identification data for the tentatively identified compounds are discussed in the following sections and summarized in Tables S4-S6 with the related confidence level according to Schymanski et al. 26 Whenever the annotated compounds could be compared to available standards in terms of retention time, accurate mass, and MS/MS spectrum, the compounds were considered identified (confidence level 1). Otherwise, the compounds are tentatively identified (confidence level 2-3).

| Identification of phenolic compound degradation products
To gain knowledge on the degradation processes of phenolic compounds after the fermentation process, a customized data processing workflow was set up in Compound Discoverer based on the Expected   (Table S4). In total, 31 anthocyanins were tentatively identified in V. floribundum berries, a much larger number compared to previous studies, 15,17,18 but still lower than reported for V. myrtillus. sugar is bound to position 3 (on the non-aromatic C-ring of the flavonol structure), the homolytic cleavage is favored. 30 Based on these shreds of evidence, whenever the radical aglycone ion had a higher abundance than the aglycone ion, the compounds were described as Despite being efficiently ionized in both positive and negative ion mode, PACs have been only analyzed in negative polarity for the higher clarity of the MS 2 spectra and the minor interference of contaminants and noise. In Table S6, detailed data on the 45 tentatively identified PACs are reported. For simplicity, species presenting at least one A-type interflavanoid bond are commonly defined as A-type PACs. 34 Nevertheless, it is worth specifying that, while dimers are clearly distinguished (as there is either one A-type bond or one B-type bond), in the case of the other oligomers, there are more than two possibilities, i.e., in the case of trimers, two B-type bonds, two A-type bonds, and one for each kind, with the latter two both defined as A-type PACs. Compared to other PAC-rich matrices, such as tea, 35 strawberry, 21 and even bilberry, 11 the identified compounds are mostly A-and B-type procyanidins, meaning that they derived from only catechin and epicatechin. A-type procyanidins were more abundant than B-type ones in terms of the number of identifications (29 vs. 18) and the total peak area (73% vs. 27%), in good agreement with previous findings for bilberry. Oligomers in the range of 2-6 were tentatively identified, with A-type trimers and B-type dimers as the main species.

| Transformation of the phenolic molecular species after fermentation
The study on the changes in phenolic compounds following fermentation processes is essential to improve the knowledge on a commonly used practice for wild berry consumption 20 since fermentation processes are believed to enhance the bioavailability of phenolic compounds by converting polymeric polyphenol to simpler compounds. 36 It has been demonstrated that lactic acid fermentation of fruits and vegetables produces bioactive compounds with antibacterial, antiinflammatory, antioxidant, and antiviral properties. 37 However, while several studies have evaluated the biological activities as well as the total phenolic, flavonoid, and anthocyanin content of berries subjected to fermentation processes, 38 the issue of determining the dynamic changes in phenolic molecular species has been dealt with rarely. Recently, Wu et al. 39 have evaluated the concentration of some selected anthocyanins, phenolic acids, and organic acids in blueberry and blackberry during fermentation over time. Our untargeted approach, which allowed the tentative identification of more than 300 phenolic compounds, was applied to native and fermented samples. Therefore, it was possible to evaluate the dynamic changes of the identified compounds by ratios of the single peak areas between the two sets of samples. Thanks to the Fill gaps tool present in Compound Discoverer, compounds present with low abundances in one of the two sets of samples (and higher abundance in the other set) could still be identified. Whenever the peaks were absent in one of the two sets of samples, the noise level was chosen as the area, making it still possible to evaluate the ratios (none of the two peak area values were therefore zero). The ratios are reported in Tables S4-S6 for all tentatively identified compounds. In Figure S3, two exemplary chromatograms of the native and fermented berry extracts are shown.
For a more comprehensive evaluation of the results, the antioxidant activity, TPC, and TAC of both native and fermented berries were measured. These are reported in Table 1.
The TPC and TAC of the native berry are in line with previously obtained results for hydro-organic phenol extracts of other Vaccinium species. 40,41 However, the results reported in the literature are often significantly discordant. Therefore, the results are discussed by comparing the native and fermented berries rather than in relation to previous studies. Indeed, the fermentation process enhances the antioxidant power while lowering the polyphenol content due to hydrolysis and oxidation reactions. 42 Anthocyanins are mainly degraded after the fermentation process, resulting in the brownish color of the fermented berry compared to the bright red color of the native one. In Figure 1, the total peak areas of the two main flavonoid classes are reported for native and fermented berry, along with their ratios on a logarithmic scale.
Confirming the results obtained by the spectrophotometric assay, anthocyanins were extensively degraded during fermentation, with ratios of À1.4 and À1 for conjugates and aglycones in logarithmic scale, which correspond to a 25-and 10-fold decrease, respectively.
Conversely, flavonol conjugates showed a much lower ratio of À0.41  Figure S4. 43 The extremely high increases of some phenolic acids were believed to be mostly the result of these degradation reactions rather than deriving from hydrolysis of phenolic acid conjugates. Phloroglucinol carboxylic acid, which was produced by all anthocyanins, showed a 26-fold increase, protocatechuic acid (which is derived from cyanidin) presented a 19-fold increase, gallic acid (deriving from delphinidin) showed a 4-fold increase, methyl-protocatechuic acid (deriving from peonidin) showed a 7-fold increase, and methyl-gallic acid (deriving from T A B L E 1 Total phenolic content (TPC), total anthocyanin content (TAC), and antioxidant activity (ABTS), measured for native and fermented V. floribundum Kunth berry TPC (mg gallic acid/g dw) TAC (mg cyanidin 3-glucoside/g dw) ABTS (μmol TE/g dw) Native berry 45 ± 3 14.5 ± 1.9 276 ± 2 Fermented berry 38 ± 3 1.0 ± 0.5 288 ± 2 F I G U R E 1 Bar chart representing the total peak areas in the logarithmic scale of anthocyanin and flavonol conjugates and aglycones tentatively identified in native and fermented berry and the logarithm of the ratio fermented/native for each class petunidin) showed a 13-fold increase. Similar to flavonoids, phenolic acid conjugates were extensively hydrolyzed, with a subsequent increase of free hydroxycinnamic acids comparable to that of free flavonols.
In acid, hydroxybenzoic acid, dihydroxybenzoic acid, and gallic acid). Raw data files were re-processed, and the extracted expected compounds were manually identified. In Table 3 and Table S7,