Targeted phenolic profile of radler beers by HPLC-ESI-MS/MS: the added value of hesperidin to beer antioxidants

The well-known health beneficial properties of beer are mainly due to phenolic antioxidants. Citrus-flavored beers represent a growing side-market in the beer industry, sparingly investigated to date. The phenolic profile of commercial radler beers (R1, R2) was investigated to evaluate the impact of the lemon juice added to beer in the industrial production. Results were compared to those obtained for opportunely chosen commercial beer (B) and lemonade (L). The study was carried out by an HPLC-MS/MS with an electrospray ionization source in selected ion recording mode, analyzing in a single chromatographic run 26 compounds belonging to the different phenolic classes of hydroxybenzoic, hydroxycinnamic and caffeoylquinic acids, flavonoids and prenylflavonoids. Different phenolic profiles were found for R1 and R2, mainly ascribed to different malt/hop/recipe used for the beer. High to very high level of hesperidin were found in the radlers, so that a major impact on phenolic antioxidants of the radlers was due to the lemon. Similarly, a major impact of the lemon aromas was found, D-limonene being the dominant peak resulting from the GC-MS analysis of the volatile fraction of the radlers. Graphical Abstract


Introduction
The beneficial effects exerted by the daily intake of citrus fruits are well-known and associated to the phytochemical composition of volatile and nonvolatile fractions, including terpene hydrocarbons, flavonoids and phenolic acids, with strong antioxidant and bioactive properties (Alu'datt et al. 2017;Shi et al. 2020;Singh et al. 2020;Mazzotti et al. 2021). The flavonoids composition of citrus fruits has been widely investigated, with distribution and content depending on citrus species, fruit part, fruit development and maturation (Gorinstein et al. 2001;Mazzotti et al. 2021;Mcharek and Hanchi 2017;Singh et al. 2020;Xi et al. 2017). Citrus limon (L.) Burm. is the most widespread Citrus species in the world after orange and mandarin, appreciated by consumers for its flavor and popular as health-promoting fruit (Xi et al. 2017). Among the bioactive flavonoids of lemon, high interest has been recently addressed to hesperidin, because of its high binding affinity to the cellular receptors of SARS-CoV-2 (Bellavite and Donzelli 2020; Santana et al. 2021).
Beer is a beverage largely consumed worldwide, and with coffee, tea, chocolate and wine, represents an important source for the daily intake of polyphenols, whose beneficial long-term effects on health have been largely supported (Rio et al. 2013;Reed and de Freitas 2020). Malt and hops are the main source of polyphenols in beer (Cortese et al. 2020;Gouvinhas et al. 2021); further, the effect of adding polyphenol-rich foods to beer has been recently evaluated (Trovato et al. 2021). Citrus-flavored beers represent a growing side-market in the beer industry, due to an increasing popularity among the consumers that appreciate the citrus fruits fresh-flavor on the beer aromas, the decreased alcohol content and a feeling of well-being on health. In particular, the radler is a beverage composed of beer and lemonade in equal parts, likely originated in Baviera at the beginning of the last century. Though a combined effect of phenolic content from beer and lemon might be expected, this beverage has been sparingly studied, at least up to our knowledge. The effect of the addition of the citrus flavors on the volatile and non-volatile profile of beer has been recently reported (Trovato et al. 2021). The phenolic composition of lemon has been also reported, mainly regarding the peel, rarely the juice, often in comparative studies with other citrus fruits, and in general phenolic compounds of different classes were investigated separately (Alu'datt et al. 2017;Gorinstein et al. 2001;Mazzotti et al. 2021;Mcharek and Hanchi 2017;Singh et al. 2020;Xi et al. 2017). Nowadays, mass spectrometry, coupled with chromatographic separation, is the most used technique for the analysis of complex matrices, suitable for clean, semi-purified or not pretreated samples (Cheiran et al. 2019;Chiarotto et al. 2019;Panusa et al. 2015;Petrucci et al. 2020a;Quifer-Rada et al. 2015).
The present study aimed to investigate the phenolic profile of the radler beverage and to evaluate the impact of the lemon juice on beer antioxidant fraction. For this purpose, two commercial Italian radlers, declared made with Italian malts and lemon juice, were analyzed by high performance liquid chromatography (HPLC) coupled with tandem mass spectrometry (MS/MS), with an electrospray ionization source (ESI) acquiring in selected ion recording (SIR) mode, by using a method previously developed for the identification and quantitation of fourteen phenolic compounds (Petrucci et al. 2020b), and up-gradated for the analysis of twelve phenolic compounds more. One Italian beer and one Italian lemonade, soft drink declared made with Italian lemon, were chosen and analyzed for a comparison.
The volatile fraction of the radler beers was also investigated by gas chromatography (GC) coupled with mass spectrometry (MS) and compared to those ones of beer and lemon juice, with the aim to obtain a comprehensive metabolic profile of the radler beer.

Samples collection and preparation
Two Italian radlers (R1, R2), one Italian beer (B, same brand of R1) and one Italian lemonade soft drink (L) were purchased from a local supermarket and stored at 4 °C before use. Based on the label, Italian malts were used for B, R1 and R2. The alcohol content by volume (ABV) was 4.7% (B), 2% (R1) and 2% (R2), respectively. R1 contained concentrated lemon juice 3.2% (100% Italian lemon) and 42% beer; R2 contained concentrated lemon juice 2.7% (100% Italian lemon), 40% beer, orange, lime and acerola juice. L contained concentrate lemon juice 12% (100% Sicilian lemon). Data are resumed in Table 1. 20 mL of each sample were degassed for 15 min in ultrasonic bath (Metason 60, Struers), filtered at 0.45 μm and stored at -20 °C. Prior to analysis, the samples were brought back to room temperature and appropriately diluted with the mobile phase (MilliQ water/acetonitrile, both formic acid 5 mM, 95:5 v/v).

Standards solution preparation, calibration curves, quality-of-analysis parameters
A stock solution containing 3,5-DHBA, 2,5-DHBA, 2,6-DHBA, 4CQA, SaA, Cat, My, Q3G, K3R, He, IsoX and X was prepared by dissolving 1 mg/mL of each standard in methanol; working solutions were prepared by appropriate dilution with the mobile phase (A/B, 95:5, v/v, vide infra). The isomeric compounds were analyzed also separately to unambiguously assign the retention time. Calibration curves were calculated with equal-weighted least-squares linear regression analysis of the SIR peak area against the standard nominal concentration, by using 10, 20, 40, 60, and 80 µg/L solutions analyzed in triplicate (20 µL injected). Limit of detection (LOD) and quantitation (LOQ) were obtained as LOD = 3Sa/b and LOQ = 10Sa/b, respectively, where Sa and b are the estimated standard deviation and the slope of the analytical calibration function with a 95% confidence level, respectively (Trani et al. 2015). 40 µg/L solutions were used to evaluate intraday (five injections) and interday (triplicate injections in different days) precision, and results given as percent standard deviation (RSD%); 40 µg/L solutions were used to evaluate accuracy (triplicate injections), and results given as the percent difference between the nominal concentration and the measured one. The recovery test was carried out in duplicate by spiking five levels of concentration (10, 20, 40, 60, and 80 µg/L) into B diluted 1:100 with the mobile phase; the percent recovery was calculated as the ratio (peak area in B prepared by an additional method)/ (peak area in solvent) and reported as RSD%. Matrix effect (ME) was evaluated as [(slope of the calibration curve in the matrix)/(slope of the calibration curve in the solvent) -1]x100. Linearity, sensitivity, precision, accuracy, recovery and ME are shown in Table 2.

HPLC-ESI-MS/MS instrumental conditions
A Waters 1525µ HPLC (Milford, MA) was used for the chromatographic separation, performed with a Waters XBridge C18 (150 × 2.1 mm i.d.) 5 μm analytical column; A (MilliQ water/formic acid 5 mM) and B (acetonitrile/formic acid 5 mM) were used as mobile phase for the elution binary gradient (Petrucci et al. 2020b)

Stir bar sorptive extraction
Stir bar sorptive extraction was carried out by adding a polydimethylsiloxane coated stir bar (10 mm length, 3.2 mm o.d., 0.5 mm thickness, "Twister", Gerstel, Germany) to 10 mL of the samples, in turn. After 1 h stirring at room temperature, the bar was removed, rinsed with MilliQ water, and placed in 350 µL of dichloromethane (Horák et al. 2007). After 1 h stirring, the extract was injected (1 µL) into the GC-MS for the analysis, in duplicate.

GC-MS instrumental conditions
A GC-MS System Clarus 500 MS Turbo (PerkinElmer Instruments LLC, U.S.A.) was used to analyze the volatile fraction of the samples, with a Rtx®-1 capillary column (Restek, Bellefonte, U.S.A.), 60 m, 0.25 mm id, 0.25 μm df, and helium as carrier gas flowing at 1 mL/min, split ratio 1:10; injector temperature 250 °C, analysis program: 35 °C for 10 min, to 100 °C at 5 °C/min, to 201 °C at 3 °C/min, final temperature of 210 °C held for 40 min (Di Matteo G. et al. 2021;. The GC-MS interface temperature was 200 °C, 180 °C the source temperature. The acquisition was carried out in full scan mode, mass range 30 ÷ 200 Da, scan time 0.2 s. Data acquisition, data handling, and instrument control were performed by Turbomass 6.1.0 v PerkinElmer. Compounds were identified by comparison of mass spectra with NIST libraries.

Statistical analysis
All samples were analyzed in triplicate and results reported in Table 3 as mean values ± standard deviation (SD). Data were analyzed by using the one-way analysis of variance (ANOVA). The significance of differences (p < 0.05) among samples was determined by the Tukey test.

Phenolic profile by HPLC-ESI-MS/MS-SIR mode
26 phenolic compounds were investigated in two commercial radlers (R1 and R2, different brand), made with Italian malts and lemon juice, according to the label. Other citrus fruits besides lemon were labelled for R2. One beer (B, same brand of R1) and one lemonade soft drink (L), containing Sicilian lemon juice, according to the label, were analyzed for a comparison (Table 1).
The hydroxycinnamic acids (HCAs) were identified in all samples, except CA 10, found only in B and R1 (same brand). FA 16 was predominant in all alcoholic samples, as expected since it is typical of beer, while SA 15 was more abundant in L. R1 and R2 resulted significantly different (p < 0.05) for HCAs, R1 in general richer than R2, analogously to what observed for HBAs, described above. B and R1 resulted significantly different for HCAs, except for CuA 14: the general decrease observed in R1 compared to B was likely due to the dilution step occurring during the radler production, that prevails over the possible intake from the lemon juice.
CuA, SA and FA (14-16) might be typical of the Italian lemon, used to prepare L, R1 and R2. Noteworthy, some authors reported the presence of HCAs 10 and 14-16 in fresh lemon, mainly in the peel (Gorinstein et al. 2001); others reported the presence of CA 10 in different fruit parts of lemon cultivars, juice included though in lower amounts, while very low content of FA 16 was found only in the peel, not in the juice (Xi et al. 2017).
Flavonoids were differently distributed among the analyzed samples.
Prenylflavonoids (pFs) from hops X 26 and IsoX 25 are characteristic of beer and were absent in L. Similar content of 25 was found in the alcoholic samples; conversely, a low quantity of 26 was found in R2, suggesting, once more, differences in the beer recipes used for R1 and R2, likely regarding hop, thermic treatments or pH (Zambrzycka-Szelewa et al. 2020). Fig. 1 Comparative view of the distribution of each phenolic class in the analyzed beverages (B, R1, R2 and L), reported as percentage of each phenolic class respect to the total phenolic content calculated as the sum of all classes. Green: pFs, prenylflavonoids; blue: HBAs, hydroxybenzoic acids; brown: HCAs, hydroxycinnamic acids; grey: Fs, flavonoids Among flavonoids (Fs),Cat 6 was the most abundant in B, R1 and R2, in agreement with literature reporting 6 as the typical flavonoid of beer (Cheiran et al. 2019;Gouvinhas et al. 2021), and it was absent in L.
Q 23 and K 24 are typical of beer too (Di Gouvinhas et al. 2021;Petrucci et al. 2020bPetrucci et al. , 2021Quifer-Rada et al. 2015), mainly coming from hops. Quercetin was found in the free form (23) and in the bonded form, identified as Q3G 19; conversely, kampferol was present mainly as the bonded form, identified as K3R 20, though it was detected also in the free form (24) in B and R1. B, R1 and R2 resulted not significantly different (p < 0.05) with regard to 20. Noteworthy, K3R 20 was reported as a discriminant compound for lager beer (Cheiran et al. 2019).
My 18 was found in trace in B and L, and it was quantitated in R1, in which the higher content was likely due to the intake from both B and the lemon juice. It was not detected in R2.
Conversely, Ru 17 and He 22 were found typical of the Italian lemon juice; particularly, He 22 was present in L in very high amount, in agreement with literature (Mcharek and Hanchi 2017;Singh et al. 2020;Xi et al. 2017). Ru 17 and He 22 were absent in B, consequentially their presence in R1 and R2 was due to the Italian lemon juice.
He 22, Ru 17, Q3G 19 and My 18 were found in L, with a content following the order 22 > > 17 > 19 > 18. These data agree with literature reporting the biosynthesis and the accumulation of high level of flavonoid glycosides in citrus (Owens and McIntosh 2011). Free quercetin 23 was not detected in L, the same occurring for K 24 and its rutinoside derivative 20. He 22 and Ru 17 were reported in the literature as the main representative flavonoids of lemon (Alu'datt et al. 2017;Mchareck and Hanchi, 2017;Xi et al. 2017).
Resuming, a different phenolic profile was observed for R1 and R2, mainly regarding the absence of HBAs in R2 and a different distribution of IsoX (25) and X (26). These results may be ascribed to differences in malt, hop, and brewing process for the production of the beers used to prepare R1 and R2. This is confirmed by the correlation found between R1 and B, same brand, that had substantially the same phenolic profile, except for hesperidin.
Conversely, R2 was found much richer than R1 in hesperidin 22: since similar content of the concentrate Italian lemon juice is labelled for R1 and R2 (3.2% vs. 2.7%, respectively, see Table 1), such a higher content of He 22 in R2 was likely due to the presence of the other citrus fruits, orange and lime, reported on the label (see Table 1).
Summing up, the targeted phenolic profile of the Italian lemon juice has been firstly investigated, at least up to The impact of the lemon juice on the phenolic profile of beer was evaluated for the analyzed samples. The total content of the phenolic compounds of B, R1, R2 and L were summarized for classes (hydroxybenzoic acids HBAs, hydroxycinnamic acids HCAs, flavonoids Fs and prenylflavonoids pFs), and comparative results are plotted in Fig. 1. Radlers resulted strongly strengthened in antioxidants content respect to beer from both qualitative and quantitative aspects, with high level of hesperidin.

GC-MS fingerprinting of the volatile fraction
The volatile fraction profile of R1, R2, B and L was investigated by GC-MS analysis (Di Matteo G. et al. 2021;) of the extracts obtained by a first stir bar extraction followed by a back-extraction in dichloromethane (Horák et al. 2007).
The untargeted analysis evidenced 23 peaks (27-49 in Table 4), tentatively assigned by comparison of the fragmentation spectra with NIST libraries: 8 compounds resulted typical of B, 13 compounds resulted typical of L, and 2 compounds were found in both B and L. Most of the 23 compounds were found in the volatile profile of R1 and R2.
Heptane 27 and m-xylene 30 were found in all the samples.
A major impact of lemon aromas was found on the radlers, D-limonene being the dominant peak, as shown in the total ion chromatograms (TIC) in Fig. 2. Qualitative data are resumed in Table 4.

Conclusion
The present work aimed to characterize the phenolic fraction of radler beer, beverage composed of equal parts of beer and lemonade, growing in popularity because of the citrus fruits fresh-flavor on the beer aromas, the decreased alcohol content and a feeling of well-being on health. For this purpose, two independent commercial radlers, composed of lager beers made with Italian malts and Italian lemons, were analyzed by liquid chromatography (HPLC) coupled with a tandem mass spectrometer (MS/MS) working in Selected Ion Recording (SIR) mass spectrometry technique, by using a modified method previously developed. The validation parameters of the herein up-gradated method are given. One lager beer and one lemonade, made with Italian malts and Italian lemons, respectively, were also investigated for a comparison. From the targeted analysis of 26 phenolic compounds including 10 hydroxybenzoic acids (HBAs), 6 hydroxycinnamic acids and derivatives (HCAs), 8 flavonoids (Fs) and 2 prenylflavonoids (pFs), 20 compounds were identified in at least one of the analyzed sample, and quantitated in most cases. Noteworthy, the phenolic profile of the Italian lemon juice has been firstly investigated, at least up to our knowledge, including HBAs, HCAs, Fs and pFs: coumaric, sinapic and ferulic acids (14-16, respectively), rutin, myricitrin, quercetin-3-O-glucoside (17-19, respectively) and hesperidin 22, were identified and quantitated where possible. A very high content of hesperidin was found in the lemonade, as expected because it is typical of citrus fruits.
Some differences were observed between the two samples of radlers, mainly regarding hydroxybenzoic acids, hydroxycinnamic acids and prenyflavonoids, that means the typical phenolic compounds of beer: such differences might be due to malt, hop and/or brewing process used to produce the beers for the radlers. Since sinapic acid, rutin and quercetin-3-O-glucoside, besides coumaric and ferulic acids at a very lesser extent, were found also in the lemonade, their content in the radlers might be due to both beer and lemon. High to very high level of hesperidin were found in the radlers, so that a major impact on phenolic antioxidants of the radlers was due to the lemon. A major impact of the lemon aromas was also found, D-limonene being the dominant peak resulting from the GC-MS analysis of the volatile fraction of the radlers. Besides the beneficial antioxidants of beer, the radler resulted strongly enriched by the citrus fruits flavonoids, mainly hesperidin, and by lemon aromas as D-limonene.