Polyphenolic Profiling of Green Waste Determined by UPLC-HDMS^E

Abstract

Valorising green waste will greatly enhance and promote the sustainable management of this large volume resource. One potential way to achieve this is the extraction of high value human health promoting chemicals (e.g., polyphenols) from this material. Our primary aim was to identify the main polyphenols present in four contrasting green waste feedstocks, namely Smyrnium olusatrum, Urtica dioica, Allium ursinum and Ulex europaeus, using UPLC-HDMS^E. Polyphenol-rich Camellia sinensis (green tea) was used as a reference material. Samples were extracted and analysed by UPLC-HDMS^E, which was followed by data processing using Progenesis QI and EZ Info. A total of 77 high scoring polyphenolic compounds with reported benefits to human health were tentatively identified in the samples, with abundances varying across the plant types; A. ursinum was seen to be the least abundant in respect to the polyphenols identified, whereas U. europaeus was the most abundant. Important components with a diverse range of bioactivity, such as procyanidins, (−)-epigallocatechin, naringenin, eriodictyol and iso-liquiritigenin, were observed, plus a number of phytoestrogens such as daidzein, glycitin and genistein. This research provides a route to valorise green waste through the creation of nutritional supplements which may aid in the prevention of disease.

1. Introduction

In most countries, green waste typically represents a high volume, low value resource, with most of this material being composted and subsequently spread back to agricultural land to improve soil quality [1]. However, green waste also represents a promising starting material for the direct extraction of valuable compounds and for the chemical and fermentative conversion of this waste into basic chemicals [2]. One of the main issues in valorising this resource, however, is knowledge of what high value products can be obtained in sufficient quantities from different types of green wastes to make it commercially viable. One area that has drawn particular interest has been the extraction of polyphenols [3]. These plant species may also contain many other non-polyphenolic bioactive chemicals which are also worthy of attention. Co-extraction of these would, of course, further enhance the value of waste materials, though this research was focused on polyphenols. A full cost–benefit analysis and life cycle assessment are required to determine the valorisation benefits relative to other synthesis or extraction procedures.

Polyphenols are a naturally occurring group of secondary metabolites which are relatively abundant in plants and which are purported to have many health benefits [4,5,6]. For example, they are thought to play an important role in disease prevention, resulting not only from their antioxidant ability but also their epigenetic influence and their positive impact on the composition of gut microbiota. Due to their complex chemical structures, many of these plant-derived bioactive polyphenols can be difficult to synthesize in large quantities [7]. Green waste offers a potentially cheap feedstock to extract and purify these compounds; however, this necessitates good knowledge of the polyphenols present in different source materials [2].

Green waste is expected to contain many thousands of chemicals. Consequently, high resolution analytical approaches are needed to enable separation and identification of the myriad chemicals present. One potential solution is the use of ultra-performance liquid chromatography (UPLC) linked to an ion mobility time-of-flight high-definition/high-resolution mass spectrometer (UPLC-HDMS^E). Recent work characterising phenolic compounds in forestry waste has shown that UPLC-HDMS^E can provide an in-depth analysis of the wide suite of phenolics present [8]. This is also supported by the use of UPLC-HDMS^E for the detection and characterisation of bioactive compounds in complex medicinal mixtures and urine [9,10]. The characterisation of polyphenolics in agricultural or municipal green waste via UPLC-HDMS^E has not, to the best of our knowledge, been previously undertaken. The aim of this study was therefore to analyse the diversity of polyphenols in four contrasting but common green waste materials generated in municipal or agricultural settings. These plant-based feedstocks have previously been characterised by a range of analytical techniques, but they have not been subject to the potential benefits of the detailed characterisation provided by UPLC-HDMS^E. As a broad reference material, green tea (Camellia sinensis) was also included in the study. The phenolic chemistry of this plant material has been well characterised [6,11,12], and can therefore act as a validation of this discovery workflow, i.e., the expected polyphenols associated with green tea should be observed.

2. Materials and Methods

2.1. Sample Collection

Representative samples of Alexanders (Smyrnium olusatrum), Stinging Nettle (Urtica dioica), Wild garlic (Allium ursinum) and Gorse (Ulex europaeus) were collected from the Lligwy Bay area of Anglesey, Wales, UK (53°21′14″ N, 4°15′47″ W) in April 2019. The sample of U. europaeus was separated into separate flower and stem samples in order to observe whether there were any significant differences in the characterisation of these two physically connected structures. These plants were chosen due to their contrasting phylogenies and their frequent presence in municipal and agricultural green waste streams.

2.2. Sample Preparation

After collection, each sample was thoroughly washed in LC-MS grade water (Optima) and then freeze-dried (48 h) before being ground to a fine powder. A total of 0.5 g of powder was then placed in a glass beaker containing 10 mL of ethanol and sonicated in an ultrasonic water bath for 30 min before being left to stand for 24 h at 4 °C before being sonicated again for a further 30 min. After the solids had settled out, the supernatant was transferred to a polypropylene tube and centrifuged (10,000 rev min⁻¹, 30 min). Through heating to 60 °C, the resultant ethanol solution of extracted components was concentrated to 1 mL; that is to say that 1 mL of extract was equivalent to 1 g of initial sample. The sample produced was stored at −20 °C. The plant samples were prepared in quadruplicate.

2.3. Analytical Instrumentation

HDMS^E mode is a data-independent acquisition in which data for all gas phase parent ions, and also fragments (product ions) created, are recorded. This was recorded and saved as continuum data. In addition, a drift cell was used to collect ion mobility data. The Synapt G2-Si (Waters UK, Wilmslow, Cheshire, UK) can be described as a quadrupole time-of-flight mass spectrometer (Q-ToF) which has ion mobility capability added to the ion path. Analytes, which have been separated by the UPLC, were infused into a Z-Spray^TM source (Waters UK, Wilmslow, Cheshire, UK). Simultaneously, leucine enkephalin (Tyr-Gly Gly Phe-Leu) was infused via a separate probe, which provided the lock mass data to correct the mass axis drift which occurs during an acquisition. Baffle switching allowed for the selection of which infusion, analyte or lock mass entered the MS.

2.4. UPLC Conditions

A Waters I-class UPLC was used for analyte separation with a Waters Cortecs Shield RP18 (2.7 μm × 2.1 mm × 100 mm, Waters UK, Wilmslow, Cheshire, UK) solid core column installed. This column provides high selectivity for phenolic compounds due to the use of imbedded polar carbamate technology. A guard column of the same stationary phase was used for protection of the analytical column. Water with 0.1% acetic acid in reservoir A and MeOH with 0.1% acetic acid in reservoir B were used as the mobile phase. The flow rate, column temperature and injection volume were 0.5 mL min⁻¹, 40 °C and 1.0 μL, respectively. The starting composition for this eluent was 90% A and 10% B, with a linear change to 1% A with 99% B over the course of 4 min. At the end point, the initial conditions were returned to over a time of 0.2 min.

2.5. Synapt G2-Si Conditions

Negative ion data, in a mass range of 50 to 1200 Da, was acquired using resolution mode. The scan time was 0.2 s, with an average of 3 scans and a mass window of ± 0.5 Da. mode. The cone voltage was 40 V. The method was set to acquire the lockmass (leucine enkephalin, 554.2615 Da) at regular intervals (30 s). This was not used for immediate mass correction throughout the run but stored for later use in the data processing phase.

2.6. Data Processing

Progenesis QI software (NonLinear Dynamics Ltd., Newcastle upon Tyne, UK) was used to process these data. An experimental design was chosen (between subjects) by creating individual groups for the various plant extracts, plus one for the reference (green tea) and one group for blank extracts. As this detail has previously been published [8], only a brief overview is provided here. Post deconvolution, a 5 ppm precursor tolerance was used to compare ions to the ChemSpider Polyphenols database [13] using isotope similarity scores above 90%, an elemental composition of C, O, H only and in an silico fragmentation tolerance of 90%. Similarly, filters were used to reveal only analytes with ANOVA p values ≤0.01 and where blanks were the lowest mean, and, also, scores with a value above 40 were selected for further evaluation. Multivariate analysis (MVA) was conducted through the use of EZInfo (Umetrics, Umeå, Sweden). Matlab (MatWorks Inc., Natlick, MA, USA) was used to create the heat map.

3. Results and Discussion

3.1. Bioactive Phenolic Compounds in the Green Waste Extracts

An Excel file in the supplementary materials (plant extracts XL SM v6) provides a detailed summary of the 77 high-scoring components elucidated through this analysis of green waste. Details such as retention time (min.), normalised abundance, drift time (ms) and accurate mass values are provided for parent ions and their main product ions. Very low ANOVA-p and q values are seen, which indicates a false discovery rate (FDR) close to zero. Mass errors of ≤5 ppm are observed in these data, plus scores of over 80 for isotope similarity.

In cases where it was not possible to distinguish analytes from species of the same accurate mass, the various possibilities are listed. Examples of total ion chromatograms (Figures S1–S6) and example molecular structures of the 29 identified components, with scores of 50.0 and above, are shown in Figure S7. Mass spectra (Figure S8) are available in the supplementary materials.

Green tea, which was used as a reference material, exhibited the expected polyphenols generally associated with this plant in the published literature. These include (−)-epigallocatechin, (−)-epigallocatechin gallate, theaflavin and procyanidin (B1, B2, B3 or B4). Another key component of green tea’s polyphenol profile, (−)-epicatechin, was also observed at 1.23 min and m/z 289.0732, with an average abundance of score of 38.1, 6782 average abundance, a mass error of 4.8 ppm and very low ANOVA-p and q values (1.31 × 10⁻⁸ and 3.47 × 10⁻⁹). A confirmatory product ion at m/z 245.0808, due to the loss of 44 Da (CH₃CHO), was observed [14]. This added further validity to this discovery method. An abundance profile of (−)-epicatechin across the plant extracts can be seen in the supplementary materials showing the greatest abundance in the green tea extract (Figure S9).

A heat map was created using averaged abundances for each sample type over the 77 identified polyphenol components (Figure 1). This visual overview of polyphenol identifications shows that abundances are higher in U. europaeus flower and stem extracts than they are in the extracts of S. olusatrum, U. dioica and A. ursinum. Furthermore, it can be seen that A. ursinum leaf has the lowest abundance of the components identified here. It is also noted that individual plant extracts have their own pattern or fingerprint of abundances, with components showing much variation been plant types.

A loadings bi-plot was created as another overview of the data. Pareto scaling was used in this unsupervised principal component analysis (PCA). Sample replicates are tightly clustered and sample types show clear separation. The relationship between sample types and the data swarm of m/z values (x-variables) can also be seen (Figure 2).

3.2. Tentative Identification of Polyphenols

Due to the lack of fragmentation data contained in databases, in silico predication has been used here. As the fragmentation mechanisms of polyphenols are well documented, this can be quite an effective approach. Retro-Diels-Alder (RDA) reactions [15] are the general category here, which also applies to the sugar moiety, which is often seen in polyphenols. Additionally, the mass spectra of conjugated phenolic compounds showed the aglycone ion as result of the loss of sugar moieties such as hexosyl ([M–162]⁻) or pentosyl ([M–132]⁻). An example of this is seen with the two isomers of isorhamnetin glucoside (3 or 7) that are identified at 1.81 and 2.63 min. Their [M-H]⁻ ions undergo the loss of the hexoside moiety to create product ions of m/z 315.0503 and m/z 315.0508. It is also noted that these isobaric species are not only separated by chromatographic retention time but also by drift time, further confirming that these molecules are different in terms of their collision cross section (CCS). The isomer at 1.81 min is seen in high abundance in U. dioica and the other isomer at 2.63 min is high in S. olusatrum. Both are also fairly abundant in the U. europaeus flower extract. Isorhamnetin-3-glucoside is a major component of Salicornia herbacea, which is traditionally used in Asian medicine and is thought to exhibit multiple nutraceutical and pharmaceutical properties [16]. Isorhamnetin-7-glucoside is suggested to have antitumour activity in a review of topopoisons from weeds [17].

The highest scoring component, gallic acid, was identified by its [M-H]⁻ ion at 1.43 min and the loss of 44 Da to create the product ion m/z 125.0238 [M-H-CH₃CHO]⁻. Although this had an expected high abundance in the green tea extract, it was only present at low abundance in the plant extracts. The identification of (−)-epigallocatechin was achieved by observation of the [M-H₂O-H]⁻ adduct at m/z 287.0562, eluting at 1.41 min and the main MS/MS fragments of m/z 257.0437 [C₁₄H₁₀O₅-H]⁻ and m/z 151.0387 [C₈H₈O₃-H]⁻. This is one of the key health components of green tea [18] and was present in this extract in high abundance. Although (−)-epigallocatechin was detected in all extracts, it was found to be in particularly high abundance in the U. europaeus flower and stem extracts. At m/z 179.0343 and 1.76 min, trans-caffeic acid was identified. The product ion m/z 135.0441 [C₈H₇O₂+e]⁻ was used for additional confirmation. This component is well documented in the literature as a potent antioxidant which can aid in the fight against cellular injury due to reactive oxygen species [19], and was present in this study, with the highest abundance being in U. dioica and, to a lesser degree, in S. olusatrum.

To link together some key phytoestrogens, dihydrodaidzein (CSID154076) was elucidated via its [M-H]⁻ adduct (m/z 255.0663) at 1.87 min. Major fragments of m/z 121.0282 [C₇H₆O₂-H]⁻ and m/z 119.0495 [C₈H₈O-H]⁻ were observed in the high energy MS/MS signal, although other product ions were apparent. Daidzein was observed at 2.58 min and m/z 253.0515, with confirmatory fragments m/z 224.0465 [C₁₄H₉O₃-H]⁻ and m/z 208.0520 [C₁₄H₈O₂+e]⁻ and genistein eluted at 2.73 min, with a parent ion at m/z 251.0336 [M-H₂O-H]⁻ and primary fragment at m/z 223.0396 [C₁₄H₁₀O₄-H₂O-H]⁻. The [2M-H]⁻ ion revealed glycitin with an m/z ratio of 891.2362 and an MS/MS fragment at m/z 729.1821 [C₁₈H₁₁O₅+M+e]⁻. These components, which were shown to be highly abundant in U. europaeus flower, U. europaeus stem and U. dioica, have been associated with protection against major adverse cardiac events in women [20]. The findings support the use of a plant-based diet and the need for future randomized prospective studies examining the influence of glycitin and genistein, as well as daidzein and dihydrodaidzein diets on cardiovascular outcomes. A literature review on the benefits of phytoestrogen supplementation on human cognition was conflicting, with less than half of the included studies showing beneficial effects, though supplementation with soy isoflavones for less than 6 months, irrespective of dose and mode, can improve cognitive performance, with greater impact on women than men. Equol is seen as an important product of daidzein metabolism in this process, though only 30%–50% of the general population are equol producers. This is thought to affect individual responses to isoflavone intervention [21]. Two isomeric glucosides of daidzein, which are known as daidzin (m/z 397.0933, 3.15 min) and puerarin (m/z 397.0938, 2.97 min), are also a soy isoflavones known to convert to equol. On top of this, identified at a lower score, is cis-4-hydroxyequol (m/z 257.0807, 2.32 min) which has activity similar to that of equol itself. Detail is available on glucuronidation via microbial metabolic processes, providing a possible route for how polyphenol glucuronides are formed outside of mammalian metabolism [22,23]. Though the 2 isomers of equol O-glucuronide (7 and 4′) were identified at 1.43 and 2.18 min with m/z 417.1198 and m/z 835.2454 parent ions, respectively, literature is scarce on how these two components, found in high amounts in U. europaeus stem and flower extracts, would contribute to overall blood plasma equol levels post transformation via gut microbial enzymes.

Naringenin, a flavonoid common in citrus fruits, was found to be another high scoring component, with high abundance in the U. europaeus flower and stem extracts. Identified by its [M-H]⁻ adduct (m/z 271.0617) at 1.31 min and MS/MS product ion at m/z 243.0654 [C₁₄H₁₂O₄-H]⁻, many beneficial biological effects have been linked to naringenin, including antioxidant, cardioprotective, antitumor, antiadipogenic, antiviral, antibacterial and anti-inflammatory effects [24]. In light of the current global pandemic, positive claims have been made as to the therapeutic potential of naringenin in the treatment of COVID-19 [25,26]. Two positional isomers of naringenin were also observed: 3′,4′,7-trihydroxyisoflavanone ([M-H₂O-H]⁻, m/z 253.0518, 1.79 min, dt 2.3328 ms) and 7,3,4-trihydroxyflavanone, also known as butin ([M-H]⁻, m/z 271.0604, 1.76 min, dt 2.5498 ms), the former being prominent in U. europaeus flower and stem and the latter being of greatest abundance in green tea, but also present in smaller amounts in the other extracts. These had similar retention times but were clearly separated by drift time, and are therefore different molecules. Molecules of this type are often associated with red wine as a contribution from the barrel wood [27] and have been shown to have potential health benefits [28]. In the context of this, naringin, the rhamnoglucoside of naringenin, was likewise identified in U. europaeus flowers and stem samples ([M-H₂O-H]⁻, m/z 561.1596, 2.00 min), but at much lower abundance than naringenin. The major confirmatory MS/MS fragment was observed at m/z 515.1190 [C₂₅H₂₆O₁₃-H₂O-H]⁻, showing partial fragmentation of the rhamnoglucoside group. The sugar moiety is a major determinant of the absorption of dietary flavonoid glycosides in mammals [29] and future research into comparative potency would be of interest here. The identification of naringenin 7-glucuronide ([2M-H]⁻, m/z 895.1921, 2.32 min) in high abundance in U. europaeus flower extract brings to mind a similar question focused on the potency of the glucuronides of polyphenols.

Tetrahydroxyflavanones have been shown to reduce the heme group in cytochrome c, which is a necessary component of the electron transport system and is also involved in apoptotic pathways. The oxidation state of the iron in the heme group is crucial to its specific functions [30]. These compounds can be seen as the addition of a phenol group to naringenin, with two being observed in these data: 3′,4′,5,7-tetrahydroxyisoflavanone ([2M-H]⁻, m/z 575.1184, 1.20 min) and eriodictyol ([M-H]⁻, m/z 287.0554, 2.19 min). 3′,4′,5,7-tetrahydroxyisoflavanone was seen significantly only in the reference green tea extract and will therefore receive no further discussion. On the other hand, eriodictyol was present in all of the extracts analysed, with a high abundance in U. europaeus flower and stem, but reasonable amounts were also present in U. dioica and A. ursinum leaf. The potential health benefits of eriodictyol are plentiful [8], with further research showing increasing evidence of its function and benefits. For example, eriodictyol may have therapeutic potential for the treatment of rheumatoid arthritis [31], and a study also shows that eriodictyol may provide a new therapeutic strategy for the management of diabetic retinopathy through the inhibition of high glucose-induced oxidative stress and inflammation in retinal ganglial cells [32].

3′-Hydroxy-O-desmethylangolensin was identified by its [M-H]⁻ adduct, m/z 273.0770, at 1.95 min and confirmed by its MS/MS fragment m/z 119.0497 [C₈H₈O-H]⁻. This was only seen in significant amounts in the green tea reference and therefore will not be discussed in further detail. This was also the case for 6′-hydroxy-O-desmethylangolensin, which was identified by its [M-H]⁻ adduct, m/z 273.0758, at 1.77 min, p-hydroxymandelic acid ([M-H]⁻, m/z 167.0344, 1.18 min), (−)-epigallocatechin gallate ([M-H₂O-H]⁻, m/z 439.0679, 1.76 min), (−)-epicatechin-3-O-gallate ([M-H₂O-H]⁻, m/z 423.0708, 1.95 min), daidzein 4′-O-glucuronide ([M-H₂O-H]⁻, m/z 411.0700, 1.42 min) and isopropyl 3-(3,4-dihydroxyphenyl)-2-hydroxypropanoate ([M-H₂O-H]⁻, m/z 221.0812, 1.77 min). The strong presence of the heavily documented component of green tea, (−)-epigallocatechin gallate, further validates this discovery workflow [6,27,28]. Furthermore, the product ions m/z 289.0727 [C₁₅H₁₃O₆–H]⁻, m/z 169.0146 [C₇H₅O₅+e]⁻ and m/z 125.0242 [C₆H₅O₃+e]⁻ were consistent with the loss of gallate (169 Da), producing an epigallocatechin minus the C-ring hydroxyl group (289 Da), with further fragmentation causing the loss of the B-ring (125 Da), which was also confirmed by library spectra [33].

Violanone was observed via its [M-H₂O-H]⁻ adduct, with a score of 55.8 at m/z 297.0759 and 1.87 min. The main product ion m/z 121.0282 was due to [C₇H₆O₂+e]⁻. Violanone was predominant in U. europaeus flower and stem extracts, but is also an important component of the fragrant Chinese rosewood (Dalbergia odorifera), which is used in traditional medicine, with claims broadly relating to its antioxidant qualities [34]. Furthermore, it is shown to selectively inhibit phytopathogenic fungi [35] and to be key component in the roots of Pongamia pinnata, which exhibits antioxidant, anticancer, antimicrobial and anti-inflammatory properties through its use in folk medicine [36]. It is noted, too, that the references to violanone in traditional medicinal plants, found in South Asia and China [34,35,36], also encompass the benefits of many other components identified in these plant extracts from Britain, including daidzein, eriodictyol, naringenin, genistein, isoliquiritigenin ([M-H]⁻, m/z 255.0671, 2.72 min) and tectorigenin ([M-H₂O-H]⁻, m/z 281.0454, 2.35 min). This begs the question as to whether the published benefits of polyphenols are plant specific, specific to certain phytochemical combinations or specific to a particular molecule in isolation? Isoliquiritigenin is found in very high abundance in U. europaeus flower and, to a lesser extent, in U. europaeus stem and S. olusatrum, with tectorigenin being high in U. europaeus stem and lower amounts in U. europaeus flower. Isoliquiritigenin is an important bioactive ingredient of traditional Chinese medicine, often extracted from the roots of liquorice plant species, including Glycyrrhiza uralensis, Mongolian glycyrrhiza and Glycyrrhiza glabra, with research exhibiting significant pharmacological properties, including both the prevention and treatment of tumours [37]. The transformation of tectoridin ([M-H]⁻, m/z 461.1107, 1.55 min), also found in high abundance in U. europaeus flower and stem, to the bioactive compound tectorigenin is efficiently achieved by gut microflora, resulting in anaphylaxis inhibitory action [38]. Additionally, recent research suggests that tectorigenin inhibits airway inflammation and pulmonary fibrosis in allergic asthma [39].

4-O-Feruloyl-D-quinic acid was identified by its [M-H]⁻ adduct and parent ion m/z 367.1050 at 0.43 min. The product ion m/z 291.0883 was explained by the fragment [C₁₅H₁₆O₆–H]⁻. 3-Feruloylquinic acid was identified by its [M-H]⁻ adduct and parent ion m/z 367.1024 at 1.60 min. The product ion m/z 295.0802 was explained by the fragment [C₁₄H₁₅O₇+e]⁻. Chlorogenic acid (3-O-Caffeoylquinic acid) was identified by its [2M-H]⁻ adduct and parent ion m/z 707.1802 at 2.47 min. The product ions m/z 663.1927 and m/z 459.1292 were explained by the fragment [C₁₅H₁₇O₇+e+M]⁻ and [C₇H₅O+e+M]⁻, respectively. p-Coumaroyl quinic acid was identified by its [M-H]⁻ adduct and parent ion m/z 337.0923 at 1.54 min, and the product ion m/z 307.0820 was explained by the fragment [C₁₅H₁₆O₇-H]⁻. These four molecules are known as chlorogenic acids, which are a polyphenolic family of quinic acid esters of hydroxyl-cinnamic acids such as p-coumaric acid, caffeic acid and ferulic acid. A placebo-controlled double-blind pilot study, conducted in Japan, found that chlorogenic acids decreased arterial stiffness, which is a characteristic of the progression of arteriosclerosis [40]. 4-O-Feruloyl-D-quinic acid was found in high amounts in U. europaeus flower and A. ursinum leaf extracts, whereas 3-feruloylquinic acid was found to be abundant in S. olusatrum and U. dioica. Chlorogenic acid was found to be most abundant in U. europaeus flower and U. europaeus stem, with its stereoisomer, neochlorogenic acid, also being observed ([M-H]⁻, m/z 353.0876, 1.51 min), which was seen in high abundance in S. olusatrum and U. dioica as was p-coumaroyl quinic acid.

Theaflavin ([M-H]⁻, m/z 563.1182, 2.56 min) is created when epicatechin and epigallocatechin come in contact with polyphenol oxidase. This analyte is used as an indicator of production quality in the tea industry, and can be present in low quality green tea but is more common in oolong and black teas. It is shown to be one of many components that are responsible for the antioxidant properties of tea [41]. Here, it is seen in the green tea reference but also in significant abundance in the alexander extract.

The two isomers of procyanidin were identified by their [M-H]⁻ adduct, one at 1.18 min (m/z 577.1376) and the other at 1.30 min (m/z 577.1370), with drift times of 4.5570 ms and 4.6113 ms, respectively. These were either B1, B2, B3 or B4 isomers, but which two could not be determined by the technique as described, although, with external standards of the 4 isomers, retention time and drift time comparison could reveal which is which. Both isomers detected were described by the same three product ions, namely [C₂₂H₁₈O₉-H]⁻, [C₂₂H₁₆O₈-H]⁻ and [C₁₅H₁₃O₆+e]⁻. As expected, these components were dominant in green tea [6,27,28], and both isomers were shown to be present in U. europaeus flower and stem extracts. These important phytochemicals, which give rise to the red positively charged cyanidin pigment associated with grapes and berries, exhibit beneficial health effects, including anti-inflammatory, anti-proliferative and antitumor activities, with many reports suggesting procyanidin as a promising lead compound for cancer prevention and treatment [42].

Many glucuronides of potent polyphenols have been identified and, as previously mentioned, most research relates to the creation of these via mammalian metabolism rather than their benefits as a result of ingestion. Of the remaining components, the highlights will be discussed, though further details can be found in the supplementary materials. U. europaeus flower extracts were found to have the highest abundance of hesperetin 3′-O-glucuronide and genistein 4′-O-glucuronide, to name but two examples of many. U. europaeus stem had a high abundance of the glucuronide of cis-resveratrol [43], which was the component that initially sparked great interest in polyphenols. Additionally, dihydrodaidzein 7-glucuronide was seen with the highest abundance in U. europaeus stem. 4′-Methyl epigallocatechin-3′-glucuronide was observed primarily in U. dioica extract, although the amount was fairly low. Irisolidone 7-O-glucuronide and feruloyl C1-glucuronide were both seen in very high amounts in S. olusatrum, with lower amounts of both being detected in the A. ursinum leaf extract.

Benzoic acid (m/z 365.1031, 3.66 min) and salicylic acid (m/z 413.0867, 2.42 min) were both identified via their [3M-H]⁻ adducts. Benzoic acid was further confirmed by its product ion m/z 347.0926 [C₇H₅O-H+2M]⁻, as was the case for salicylic acid MS/MS fragment m/z 369.0980 [C₆H₅O-e+2M]⁻. These analytes are commonplace in plants and have been used medicinally for many years, primarily for pain relief and as anti-inflammatory agents [44,45]. These analytes were seen in their highest abundance in U. europaeus stem and, to a lesser extent, U. europaeus flower. Additionally, U. dioica extract contained significant amounts of salicylic acid. Benzoic acid is the functional group in salicylic acid and its derivatives which are responsible for inducing stress tolerance in plants [46].

The high scoring component 5,7-Dihydroxy-8,4′-dimethoxyisoflavone (m/z 329.0660, 2.89 min) is seen as a bioactive component in mung beans (Vigna radiata) [47], although very little else is published on this specific methoxyisoflavone. The 2 product ions m/z 311.0552 ([C₁₇H₁₂O₆-H]⁻) and m/z 293.0456 ([C₁₇H₁₁O₅-H]⁻) helped to confirm the identity of this component. Although its score is high, its abundance is seen to be low across all of the extracts, with small amounts in U. europaeus flower and stem. On the other hand, the methoxyisoflavone, irisolidone (m/z 314.0786, 2.69 min), is well documented for health benefits, including anti-helicobacter pylori activity, inhibition of prostaglandin E2 production, hepatoprotective effects, anticancer, estrogenic activity, inhibition of JC-1 virus gene expression and anti-inflammatory activity. Irisolidone is created by the intestinal bacterial transformation of kakkalide (ChemSpider ID 4590337), which is a component of traditional medicine [48]. Irisolidone is evident in good abundance in U. europaeus stem and, to a lesser extent, U. europaeus flower. The product ion m/z 281.0451 ([C₁₆H₉O₅+e]⁻) helped to confirm the identity of irisolidone. Two methoxyisoflavanones (without the double bond on the C-ring) were seen lower on the list of identifications ordered by score. 7,2′-Dihydroxy-4′-methoxyisoflavanone (m/z 267.0663, 2.41 min) and 7-Hydroxy-4′-methoxyisoflavanone (m/z 269.0829, 2.07 min) were seen mainly in U. europaeus flower, U. europaeus stem and U. dioica extracts. 7,2′-Dihydroxy-4′-methoxyisoflavanone has been shown to exist in many plants and is an important component of Tibetan medicine [49]. Isoflavanones are rare compared to isoflavones, with 7-Hydroxy-4′-methoxyisoflavanone being the focus of little research in recent years [50], but it could conceivably have similar activity to its more studied analogues.

The isomers, neohesperidin (m/z 591.1716, 2.77 min) and hesperidin (m/z 591.1724, 2.42 min) were identified with a high score by their [M-H₂O-H]⁻ adducts. These are both derivatives of hesperetin, differing only in the configuration of their rhamnoglucoside moiety and, therefore, cannot, in truth, be specifically identified without external standards, which is made possible due to differences in chromatographic retention times, as well as drift time separation in ion mobility. Hesperidin and derivatives, which are commonly associated with citrus fruits, play an important role in plant defense systems to combat pathogens. It is thought that they may be useful for humans, as they possess antibacterial, antiviral and antifungal activities [51].

Furthermore, hesperidin and derivatives have been shown to have strong activity against the formation of advanced glycation end products which result in the accumulation of random damage in extracellular proteins. This process is known to have deleterious effects on biological functions which are associated with aging and diabetes, such as cataracts, nephropathy, vasculopathy, proliferative retinopathy and atherosclerosis [52]. Large amounts of hesperidin were observed in the U. europaeus stem extracts.

Urolithins have been shown to be effective in cancer chemoprevention [53] though, due to differences in gut microbiota, urolithin production capacity from ellagic acid varies amongst individuals [54]. Urolithin C ([M-H]⁻, m/z 243.0288, 2.15 min) and urolithin A ([M-H]⁻, m/z 227.0343, 1.75 min) were identified in the extracts studied, with both being significant in U. europaeus flowers and urolithin A being noticeably present in A. ursinum leaf.

Vanillin is an important flavour and fragrance component in the food industry, and is also used in the pharmaceutical and cosmetic industries. The two components, 3,4-dihydroxybenzadehyde ([2M-H]⁻, m/z 275.0555, 1.18 min) and vanillic acid ([3M-H]⁻, m/z 503.1197, 2.35 min), identified in this characterisation of plant extracts, play a key role in the biosynthetic pathway that produces vanillin [55]. 3,4-dihydroxybenzadehyde was found here in good abundance in U. europaeus flower, whereas vanillic acid was at its highest abundance in U. europaeus stem.

Lignans have been shown to have anticarcinogenic properties. 7-Hydroxymatairesinol has been shown to be effective in a prostate cancer model in vivo [56]. A study in the Netherlands showed that some plant lignans, including 7-hydroxymatairesinol ([M-H]⁻, m/z 373.1285, 2.32 min), matairesinol ([M-H]⁻, m/z 357.1336, 2.14 min) and iso-lariciresinol ([M-H]⁻, m/z 359.1499, 1.90 min) can be converted by intestinal microflora into enterolignans, e.g., enterolactone and enterodiol, and may reduce the risk of certain types of cancer, as well as cardiovascular diseases, through anti-oxidant and anti-estrogenic actions [57]. Once again, their highest abundance was to be found in U. europaeus flower and stem.

The flavonoid, pseudobaptigenin, can be seen at [M-H]⁻, m/z 281.0460, 3.10 min and was present in high abundance in U. europaeus stem and in lower quantities in U. europaeus flower. It is believed that this phytochemical could be used as a prototype for synthesizing new molecules against diabetic cataracts [58]. Although only observed at low abundance in U. europaeus stem, 3′,4′,7-trihydroxyisoflavan ([3M-H]⁻, m/z 773.2626, 2.50 min) was present, but little is published about this trihydroxyisoflavan, which can be produced microbially but is included here for completeness. Found to be high in S. olusatrum, 4′,4″-dimethylepigallocatechin gallate was observed via its [M-H]⁻ adduct at m/z 485.1076 and 1.46 min. Research shows that methylation of epigallocatechin gallate alters its potency, with a study showing reduced inhibitory effects in macrophages [59], with a further study on cell surface binding abilities showing that dimethylation prevented surface binding completely, suggesting that the hydroxyl groups on the 4′-position in the B ring and the 4′′-position in the gallate are crucial for the cell surface binding activity of epigallocatechin gallate [60].

Finally, we observed three related phytochemicals which all contain the chromone structure and that are very prominent in the polyphenol literature. Firstly, quercetin ([M-H]⁻, m/z 301.0363, 2.30 min), secondly, isoquercetin ([M-H₂O-H]⁻, m/z 445.0772, 2.40 min) and, finally, rutin ([M-H]⁻, m/z 609.1482, 1.43 min). Isoquercetin is quercetin 3-O-glucoside and rutin is quercetin 3-O-rhamnoglucoside, with the sugar moiety being an important determinant in dietary flavonoid glycosides absorption in humans [29]. Quercetin is present in all the plant samples analysed here, with very high amounts in U. europaeus flower and high amounts in S. olusatrum. Quercetin is a powerful antioxidant that has a well-documented role in reducing different human cancers, and is one of the most abundant antioxidants in the human diet [61]. In fact, isoquercetin, which was found to be most abundant in S. olusatrum, and rutin, which was found in very high amounts in U. europaeus flowers, have also been shown to have powerful anti-mutagenic activity [6].

It is noted from a review relating to the drug discovery potential of these components that the chromones of the previous paragraph have large differences in their chemistry and bioactivity compared to the chroman-4-ones, all due to the C₂–C₃ double bond [62]. Naturally occurring chroman-4-ones, such as naringenin, naringin and eriodictyol, were evident in abundance across these plant extracts, providing, together with the other identified polyphenols, a diverse range of bioactivity, with much potential to benefit human and animal health [63].

4. Conclusions

The effective use of UPLC-HDMS^E for the detailed analysis of four common, but low value, green waste materials is demonstrated here. This discovery mode characterisation, which led to the identification of 77 polyphenols with well-documented health potential, opens the door for these plants to provide naturally occurring treatments for disease.