Caffeoylquinic Acids and Flavonoids of Fringed Sagewort (Artemisia frigida Willd.): HPLC-DAD-ESI-QQQ-MS Profile, HPLC-DAD Quantification, in Vitro Digestion Stability, and Antioxidant Capacity

Fringed sagewort (Artemisia frigida Willd., Compositae family) is a well-known medicinal plant in Asian medical systems. Fifty-nine hydroxycinnamates and flavonoids have been found in A. frigida herbs of Siberian origin by high-performance liquid chromatography with diode array and electrospray triple quadrupole mass detection (HPLC-DAD-ESI-QQQ-MS). Their structures were determined after mass fragmentation analysis as caffeoylquinic acids, flavone O-/C-glycosides, flavones, and flavonol aglycones. Most of the discovered components were described in A. frigida for the first time. It was shown that flavonoids with different types of substitution have chemotaxonomic significance for species of Artemisia subsection Frigidae (section Absinthium). After HPLC-DAD quantification of 16 major phenolics in 21 Siberian populations of A. frigida and subsequent principal component analysis, we found substantial variation in the selected compounds, suggesting the existence of two geographical groups of A. frigida. The antioxidant activity of A. frigida herbal tea was determined using 2,2-diphenyl-1-picrylhydrazyl free radical (DPPH•) and hydrophilic/lipophilic oxygen radical absorbance capacity (ORAC) assays and DPPH•-HPLC profiling, revealing it to be high. The effect of digestive media on the phenolic profile and antioxidant capacity of A. frigida herbal tea was assessed under simulated gastrointestinal digestion. We found a minor reduction in caffeoylquinic acid content and ORAC values, but remaining levels were satisfactory for antioxidant protection. These results suggest that A. frigida and its food derivate herbal tea could be recommended as new plant antioxidants rich in phenolics.


Introduction
Traditional healing treatments refer to the collective knowledge, skills and practices that are based on the theories, values and personal experiences developed and used by indigenous people of different cultures to improve health, avoid and reduce diseases and their spread, or as complete cures

Introduction
Traditional healing treatments refer to the collective knowledge, skills and practices that are based on the theories, values and personal experiences developed and used by indigenous people of different cultures to improve health, avoid and reduce diseases and their spread, or as complete cures of both physical and mental health conditions [1,2]. The investigation of traditional healing procedures in unique regions such as Siberia is of particular interest due to the large variety of ethnic groups. The territory of Siberia extends eastwards from the Ural Mountains to the watershed between the Pacific and Arctic drainage basins. The local residents mostly passed traditional healing knowledge on orally from one generation of healers to the next. The study of the use of local flora of a particular region or culture by native people is termed ethnobotany [3]. The native populations of different regions of Siberia (yakuts, buryats, tuvans, soyots) have used the plants in their environs for different purposes since ancient times.
Artemisia frigida Willd. (fringed sagewort) is one of the plant species most used by nomadic people for its therapeutic properties ( Figure 1). As a member of the Compositae family, it is widespread in the steppe regions of Siberia and Mongolia, as well as on the prairies of North America [4]. There are reports of both external and internal uses of A. frigida herbs by nomads [5][6][7][8][9], as well as the addition of A. frigida to cold baths with water taken from mineral hot and cold springs known as "arshans" [7]. Cold spring water is heated in barrels using hot stones, then A. frigida herbs are added. This treatment is believed to prevent colds. The application of fringed sagebrush as an herbal tea is of the greatest interest. Herbal tea from aerial parts of A. frigida is referred to as sagaan aya tea. Sagaan aya tea is used to treat many diseases, in particular those caused by oxidative stress-a process that can trigger cell damage [10]. In particular, during surveys of nomadic populations, we found that sagaan aya tea was used to treat a number of diseases in the pathogenesis of which oxidative stress is thought to be involved, such as hypertension, diabetes, cardiovascular diseases, etc. (supplementary materials Table S1). Chemical investigations of Artemisia species often focus on artemisinin and other sesquiterpenes, ignoring phenolic compounds despite their well-known potent antioxidant properties [11]. According to literature data, a chemical study of phenolic compounds was carried out on A. frigida growing in North America [12,13], China [14][15][16][17], Inner Mongolia [18,19] and Russia [20]. As a result, only 37 substances, such as flavonoids of the flavone, flavonol and biflavone groups; coumarins and phenylpropanoids, were isolated and characterized (Table S2)  Chemical investigations of Artemisia species often focus on artemisinin and other sesquiterpenes, ignoring phenolic compounds despite their well-known potent antioxidant properties [11]. According to literature data, a chemical study of phenolic compounds was carried out on A. frigida growing in North America [12,13], China [14][15][16][17], Inner Mongolia [18,19] and Russia [20]. As a result, only 37 substances, such as flavonoids of the flavone, flavonol and biflavone groups; coumarins and phenylpropanoids, were isolated and characterized (Table S2). Derivatives of luteolin, 6-hydroxyluteolin and 3 ,5 -demethoxytricin in the form of aglycones and glycosides are the main structural types of flavones in A. frigida. Rare O-glucuronides and 2 -O-glucuronyl-glucuronides were detected only in raw materials collected in China and Inner Mongolia [17][18][19]. In specimens of Russian origin, only C-glycosides of apigenin were found [20], and luteolin-7-O-glucoside (cynaroside) was revealed in North American samples [12,13]. The only biflavone glycoside (8-O-8 -biluteolin-7, 7 -O-glucuronide) was detected in the aerial part of A. frigida from the vicinity of the Tongliao district (China) [18]. North American samples of A. frigida were characterized by the ability to accumulate derivatives of luteolin and heptahydroxyflavone, whereas the glycosides of tricetin were found in specimens from China and Mongolia. Caffeoylquinic acids are widely distributed in species of the genus Artemisia [21], but their presence in A. frigida has not been established.
Thus, we decided to characterize caffeoylquinic acids and flavonoids of A. frigida using high-performance liquid chromatography (HPLC) with both diode array detection (DAD) and electrospray ionization triple quadrupole mass spectrometry (ESI-QQQ-MS) to compare the cumulative content of flavonoids and caffeoylquinic acids in the herb of the Siberian populations of fringed sagewort, according to habitat, by HPLC-DAD quantification. At the second stage of research, based on the potential bioactivity of phenolic compounds, we investigated the digestive stability of caffeoylquinic acids and flavonoids from A. frigida herbal tea. Sagaan aya tea was processed through simulated gastric and small intestinal digestion, mimicking the physicochemical and biochemical changes that occur in the upper gastrointestinal tract. Finally, because antioxidant properties of A. frigida herbal tea may be an important means of protecting the gastrointestinal tract itself, the total antioxidant capacity of non-treated herbal tea and herbal tea after gastric and intestinal phases from sagaan aya tea was also assessed.

Plant Materials and Chemicals
The information about samples of herb of A. frigida is listed in Table 1. The species were authenticated by Prof. T.A. Aseeva (IGEB SB RAS, Ulan-Ude, Russia). Plant material was dried and powdered before analysis.   [23] and Thymus baicalensis [24].

Total Extract and Herbal Tea Preparation
For preparation of the total extract of A. frigida herb the powdered sample (100 g) was extracted three times in a conical glass flask (2 L) with 70% methanol (1 L) with stirring and sonification for 60 min at 50 • C with ultrasound power of 100 W and the frequency 35 kHz. The resulted extracts were filtered through a cellulose filter, combined, evaporated in vacuo until dryness, and stored at 4 • C until further chemical composition analysis and bioactivity assays. The yield of total extract of A. frigida herb was 22.14 g, respectively.
For the preparation of herbal tea, an accurately weighted A. frigida herb (1 g) was mixed with 100 mL of distilled water, and then heated on a heater plate and boiled for 10 min. The mixture was left to stand at room temperature for 15 min, and then filtered under reduced pressure in volumetric flasks (100 mL). The final volume was reduced to initial sign and filtered through a 0.22 µm PTFE syringe filter before analysis. procedure was used for phenolic compounds profiling. Experiments were performed on an LCMS 8050 liquid chromatograph coupled with diode-array-detector and triple-quadrupole electrospray ionization detector (Shimadzu, Columbia, MD, USA), using a GLC Mastro C18 column (150 × 2.1 mm, Ø 3 µm; Shimadzu, Kyoto, Japan), column temperature was 35 • C. Eluent A was 0.5% formic acid in water and eluent B was 0.5% formic acid in acetonitrile. The injection volume was 1 µL, and elution flow was 100 µL/min. Gradient program: 0.0-1.0 min 5-21% B, 1.0-2.0 min 21-38% B, 2.0-2.7 min 38-55% B, 2.7-3.5 min 55-61% B, 3.5-5.0 min 61-94% B. The DAD acquisitions were performed in the range of 200-600 nm and chromatograms were integrated at 330 nm. MS detection was performed in negative ESI mode using the parameters as follows: temperature levels of ESI interface, desolvation line and heat block were 300 • C, 250 • C and 400 • C, respectively. The flow levels of nebulizing gas (N 2 ), heating gas (air) and collision-induced dissociation gas (Ar) were 3 L/min, 10 L/min and 0.3 mL/min, respectively. The MS and MS/MS spectra were both recorded in negative mode (−3 kV source voltage) by scanning in the range of m/z 100-1900 at the collision energy of 10-45 eV.

HPLC-DAD Quantification Condition
HPLC-DAD analysis was performed as described in Section 2.3 and chromatograms were recorded at 330 nm. To prepare the stock solutions of reference standards, 8 mg of 4-O-caffeoylquinic acid, 5-O-caffeoylquinic acid, 3,4-di-O-caffeoylquinic acid, 3,5-di-O-caffeoylquinic acid, 4,5-di-O-caffeoylquinic acid, 3,4,5-tri-O-caffeoylquinic acid, vicenin-2, isoorientin, cynaroside, apigenin, hispidulin, jaseosidine, luteolin-3 ,4 -dimethyl ester, eupatorin, acacetin, and cirsimaritin were accurately weighed and individually dissolved in methanol in volumetric flasks (1 mL). The external standard calibration curve was generated using eight data points, covering the concentration ranges 1-500 µg/mL. The calibration curves were created by plotting the concentration levels versus the peak area. All the analyses were carried out in triplicate and the data were expressed as mean value ± standard deviation (SD). For preparation of sample solution, an accurately weighted powdered plant (40 mg) was placed in an Eppendorf tube, 1 mL of 60% ethanol was added, and the mixture was weighted. Then the sample was extracted in an ultrasonic bath for 30 min at 50 • C. After cooling, the tube weight was reduced to initial sign, and the resultant extract was filtered through a 0.22 µm PTFE syringe filter before injection into the HPLC system for analysis.

Validation Analysis
The linearity of HPLC-DAD quantification method was studied by injecting five concentrations (1-500 µg/mL) of the 16 reference standards (4-O-caffeoylquinic acid, 5-O-caffeoylquinic acid, 3,4-di-O-caffeoylquinic acid, 3,5-di-O-caffeoylquinic acid, 4,5-di-O-caffeoylquinic acid, 3,4,5-tri-O-caffeoylquinic acid, vicenin-2, isoorientin, cynaroside, apigenin, hispidulin, jaseosidine, luteolin-3 , 4 -dimethyl ester, eupatorin, acacetin, cirsimaritin). Results from each analysis were averaged and subjected to regression analysis. Limits of detection (LOD) and quantification (LOQ) were determined using the following equations: LOD = (3.3 × S YX )/a; LOQ = (10 × S YX )/a, where S YX is a standard deviation of the response (Y intercept) and a is a slope of calibration curve. Scopoletin-7-O-neohesperidoside (100 µg/mL) spiked in reference standards mixture was used as internal standard. Intra-and inter-day variations, which are presented in terms of percent relative standard deviation (%RSD) of the analyte's peak area and variability assessed the precision of the HPLC-DAD quantification. For the intra-day variability test, the mixture solution containing 16 reference standards was analysed for five replicates within one day (50 µg/mL), while inter-day assay was analysed using the same concentration for intra-day precision on four different days (interval of 1 day). The repeatability test of the sample was performed on 7-fold experiments of the mixture solution contain 16 reference standards (100 µg/mL). The stability test was performed with one sample solution, which was stored at room temperature and analysed at regular intervals (0, 2, 4, 8, 12, 24 and 48 h.). For analysis of recovery data, the appropriate amounts of the powdered sample of 16 reference standards were weighted and spiked with a known amount of reference compound and then analysed five times.

Organoleptic Analysis and Crude Composition Analysis
Organoleptic parameters (colour, odour, taste) of A. frigida herbal tea was determined according to AHPA guidance on Organoleptic Analysis [25]. Extractives and ash were determined according to WHO recommendations [26]. The protein content was estimated by Bradford method using BSA as a reference substance [27]. The lipid content was determined by extracting a known volume of A. frigida herbal tea with chloroform-methanol mixture (4:1). Carbohydrate content was determined with spectrophotometric phenol-sulphuric acid method [28]. Free sugars, organic acids, amino acids and mineral content were determined using HPLC-UV assays described previously [29]. Macronutrients, free sugars, organic acids and amino acids content were expressed as mg per 100 mL of the beverage and mineral content as µg per 100 mL of the beverage.

DPPH • Radical Scavenging Assay
The DPPH • radical scavenging activity (DPPH • ) was assessed as described earlier [38]. 500 µL of a DPPH • methanol solution (freshly prepared, 100 µg/mL) was added to 500 µL of A. frigida herbal tea. After 15 min absorbance was measured at 520 nm. A 0.01% solution of trolox was used as a positive control (PC), and water was used as a negative control (NC). The ability to scavenge DPPH • radicals was calculated using the following equation: Scavenging ability (%) = ((A 520 NC − A 520 PC ) − (A 520 Sample − A 520 PC )/(A 520 NC − A 520 PC )) × 100, where A 520 NC is the absorbance of the negative control, A 520 PC is the absorbance of the positive control, and A 520 Sample is the absorbance of the sample solution. The IC 50 value is the effective concentration at which DPPH • radicals were scavenged by 50%. Values are expressed as mean obtained from five independent experiments.

DPPH • -HPLC-DAD Procedure
The DPPH • -HPLC-DAD procedure was realized as described previously [27]. Briefly, a sample of A. frigida herbal tea (100 µL) was added to DPPH • radical solution in methanol (250 µL, 20 mg/mL). The mixture was shaken for a few seconds and left to stand in the dark for 30 min at room temperature. Then, the sample was filtered through a 0.22 µm membrane filter. The untreated sample was prepared by adding a sample of A. frigida herbal tea (100 µL) to methanol (250 µL). HPLC analysis was performed as described in Section 2.3.

Oxygen Radical Absorbance Capacity (ORAC) Assay
Lipophilic and hydrophilic oxygen radical absorbance capacity (L-ORAC, H-ORAC) were determined using fluorimetric microplate assay [39,40]. To prepare the sample, 100 mL of A. frigida herbal tea was extracted with 3 × 20 mL of hexane. The hexane fractions were combined and evaporated in vacuo (20 • C) to dryness. The residue was dissolved in 2.5 mL of acetone and then diluted with 7.5 mL of 7% randomly methylated β-cyclodextrin (RMCD; solution in acetone-water mixture 1:1, v/v) (L-ORAC sample). The residual A. frigida herbal tea after hexane treatment was extracted with 25 mL of acetone-water-acetic acid (70.0:29.5:0.5, v/v) by vortexing for 1 min and followed sonication for 5 min (40 • C). The sample was centrifuged at 4000× g for 15 min and organic supernatant was transferred to a volumetric flask (25 mL) and diluted to 25 mL with acetone (H-ORAC sample). An aliquot of L-ORAC or H-ORAC (40 µL) was to the 48-well plate, followed to 400 µL of fluorescein (0.3 mg/mL in 0.075 M phosphate buffer) and 150 µL of 2,2 -azobis(2-amidinopropane) (17.2 mg/mL in 0.075 M phosphate buffer), and readings of fluorimetric microplate reader (excitation wavelength 485 nm, emission wavelength 520 nm) were initiated immediately. The ORAC value was calculated using regression between trolox concentration (µM) and the net area under fluorescence decay curve. For the standard trolox assay, the dilution of trolox in 0.075 M phosphate buffer 6.25-50.00 µM was used. Data were expressed as trolox equivalents (µM) per 100 mL of the beverage. All the analyses were carried out in triplicate and the data were expressed as mean value ± standard deviation (SD).

Simulated Gastrointestinal Digestion Assay
The simulated gastrointestinal digestion was realized as described previously [41]. For the simulated gastric digestion phase, the sample of A. frigida herbal tea (25 mL) was incubated with freshly prepared simulated gastric fluid (25 mL, pH 2.0) in a 50 mL Erlenmeyer flask for 60 min at 37 • C in a shaking water bath (167 rpm). The gastric digestion phase was terminated by inactivating pepsin by raising the pH of the solution to 7.0 with the addition of 1 M NaOH. For the simulated intestinal digestion phase, after gastric digestion (pH 7.0) the whole sample was transferred to the dialysis bag used as the simulated small intestinal compartment. One mL of bile solution and 4 mL of simulated intestinal fluid were added to the dialysis bag, and digestion was continued for 4 h with continuous stirring. The dialysis bag was immersed in a vessel containing buffer solution (similar in composition to simulated intestinal fluid without pancreatin addiction, 1000 mL, pH 7.0) and maintained at 37 • C while mixing. This vessel was connected to a buffer feeding reservoir (at 37 • C) and a receiving flask. The buffer solution in which the dialysis bag was immersed was constantly replenished from the feeding reservoir at a transfer rate of 1.6 mL/min using a peristaltic pump. Samples (50 µL) were collected at 60 min of gastric digestion and at 240 min of intestinal digestion from retentate. HPLC samples were neutralized (if needed), and then freeze-dried. Samples were filtered through 0.2 µm syringe filters before injection into the HPLC system for analysis. HPLC-DAD conditions were similar to those in Section 2.3. To

Statistical and Multivariative Analysis
Statistical analyses were performed using a one-way analysis of variance (ANOVA), and the significance of the mean difference was determined by Duncan's multiple range test. Differences at p < 0.05 were considered statistically significant. The results are presented as mean values ± SD (standard deviations) of the three replicates. Advanced Grapher 2.2 (Alentum Software Inc., Ramat-Gan, Israel) was used to perform linear regression analysis and to generate graphs. Principal component analysis (PCA) based on a data matrix (16 markers × 21 samples) was performed using Graphs 2.0 utility for Microsoft Excel (Komi NTc URO RAN, Syktyvkar, Russia) to generate an overview for groups clustering.

Caffeoylquinic Acids and Flavonoids of Artemisia frigida Herb: HPLC-DAD-ESI-QQQ-MS Profile, Organ Distribution and Chemotaxonomy
The phenolic profile of A. frigida was investigated using HPLC with both diode array detection (DAD) and electrospray ionization triple quadrupole mass spectrometry (ESI-QQQ-MS) ( Figure 2). Detected compounds were identified by their retention times (t R ) and ultraviolet spectral (UV) and mass spectrometric data ( Figure S1) by comparison with reference standards ( Figure S2) and literature data. The method gave reproducible detection of 59 phenolics in A. frigida with various structures like caffeoylquinic acids, flavonoid glycosides and flavonoid aglycones (Table 2, Figure S3). Such a study of A. frigida, using HPLC-DAD-ESI-QQQ-MS, has not been realized previously.  Table 2.  Compounds are numbered as listed in Table 2.
Compound 39 had close to 33 spectral values but a greater retention time (t R 2.875 min versus 2.714). On the UV spectrum, a hypochromic shift of band II (λ max , nm 267→263) was observed, which is typical for 5-O-glucosides of flavones [45].  [43], is known. None of these compounds have been previously found in A. frigida and the genus Artemisia generally.
Flavone glycoside 3, which gave a deprotonated ion with m/z 695, was the only flavonoid that showed the loss of hexuronic acid residues in its mass spectrum (m/z 695→519, 519→343). The fragmentation character and UV spectrum indicated that the most likely identity of the compound was pentahydroxyflavone trimethyl ester di-O-hexuronide. The flavone with the closest structure found in A. frigida is 5,7-dihydroxy-3 ,4 ,5 -trimethoxyflavone-7-O-(2 -O-glucuronyl)glucuronide or friginoside B [15,16]. Diosmetin-7-O-glucoside (20), identified after comparison of t R , UV, and ESI-MS data with the reference standard, was found in A. frigida for the first time. Three isomeric flavones 25, 35 and 40 were luteolin-methyl ester-O-hexoside due to their UV and ESI-MS patterns [14]. Their structures await further investigation.
Flavonol aglycones and glycosides common to other sections of the genus Artemisia [58] are trace or non-detectable components in the Absinthium section. The chemosystematic significance of caffeoylquinic acids is currently not obvious, since they are probably obligatory components of the genus Artemisia generally [21].

Variation of Phenolics in A. frigida Herb: HPLC Quantification and Principal Component Analysis Data of Twenty One Siberian Populations
The known HPLC-DAD method of caffeoylquinic acid and flavonoid separation of Artemisia components [21] was used for quantitative analysis of A. frigida herbs. Sixteen non-trace compounds were selected as quantifiable markers, including six caffeoylquinic acids  Figure S6). LOD (0.12-0.96 µg/mL) and LOQ (0.36-2.89 µg/mL) values were appropriate for quantitative analysis. Values of intra-and inter-day precision were 1.23-2.44% and 1.09-2.31%, respectively (Table S4). The repeatability of the method was 1.12-2.44% and the stability values varied from 1.44% to 2.37%.
(a) (b)  [59] and wild species [28,38,[60][61][62]. These changes had to be taken into consideration due to their direct impact on bioactivity [63].  (Figure 4). A possible reason for the differences between A. frigida populations could be the value of the climate continentality or extremeness. Due to the very large area of the Siberian territory, there are various ecological regions located there, resulting in variation in the chemical profiles of plants. Phenolic compounds are the early chemical markers reflecting plant relationships with the environment, and similar changes in the phenolic profile have already been demonstrated for some cultivated plants [59] and wild species [28,38,[60][61][62]. These changes had to be taken into consideration due to their direct impact on bioactivity [63].

A. frigida Herbal Tea Phenolics: General Characteristics, HPLC Profile, in vitro Digestion Stability and Antioxidant Capacity
Tea made from A. frigida is an herbal beverage with a specific camphoraceous, eucaluptus-like, fresh and balsamic aroma, with a slightly bitter and pleasant taste. The concentrations of macronutrients, free sugars, organic acids, amino acids and minerals have valid values and were appropriate for herbal teas [27,36] (Table S6). The sensory evaluation showed that A. frigida herbal tea has good levels of colour, flavour and overall preferences, demonstrating its potential as a new herbal tea (Table S7, S8, Figure S7). The basic phytochemical components of A. frigida herbal tea were phenolic compounds like phenylpropanoids (42.18 mg/100 mL) and flavonoids (4.63 mg/100 mL) as well as water-soluble polysaccharides (24.11 mg/100 mL) (Table S9). Coumarins, anthocyanidins, tannins, iridoids, essential oils and alkaloids were the trace or undetectable components of A. frigida herbal tea.
The total content of caffeoylquinic acids and flavonoid glycosides in A. frigida herbal tea was 36.58 and 3.75 mg per 100 mL. Flavonoid aglycones were present in trace or undetectable concentrations. The two dominant phenolics of A. frigida herbal tea, 5-O-caffeoylquinic acid and 3,5di-O-caffeoylquinic acid, were present at levels of 16.09 and 16.35 mg/100 mL of decoction, respectively. Together, these two compounds accounted for more than 88% of the total caffeoylquinic acid content and about 80% of the total phenolic content. A high caffeoylquinic acid content was previously found in decoctions of other Artemisia plants, such as A. capillaris Thunb., which had 15.72-27.69 mg/100 mL of 5-O-caffeoylquinic acid and 4.84-10.18 mg/100 mL of 3,5-di-Ocaffeoylquinic acid [64], and A. campestris subsp. maritima, demonstrating a total caffeoylquinic acid content of 22.10 mg/100 mL [53]. At present, the known plant source of phenolic compounds in the everyday diet is coffee: the levels of total caffeoylquinic acids in soluble and instant coffee beverages range from 14.7 to 44.0 mg/100 mL [65]. We therefore conclude that A. frigida herbal tea solution can also be a good source of caffeoylquinic acids.

A. frigida Herbal Tea Phenolics: General Characteristics, HPLC Profile, in vitro Digestion Stability and Antioxidant Capacity
Tea made from A. frigida is an herbal beverage with a specific camphoraceous, eucaluptus-like, fresh and balsamic aroma, with a slightly bitter and pleasant taste. The concentrations of macronutrients, free sugars, organic acids, amino acids and minerals have valid values and were appropriate for herbal teas [27,36] (Table S6). The sensory evaluation showed that A. frigida herbal tea has good levels of colour, flavour and overall preferences, demonstrating its potential as a new herbal tea (Tables S7 and S8, Figure S7). The basic phytochemical components of A. frigida herbal tea were phenolic compounds like phenylpropanoids (42.18 mg/100 mL) and flavonoids (4.63 mg/100 mL) as well as water-soluble polysaccharides (24.11 mg/100 mL) (Table S9). Coumarins, anthocyanidins, tannins, iridoids, essential oils and alkaloids were the trace or undetectable components of A. frigida herbal tea.
The phenolic profile of A. frigida herbal tea included nineteen compounds predominantly hydrophilic in nature, such as caffeoylquinic acids (1,2,4,19,21,24,26,21,26), flavonoid glycosides (7-9, 13, 14, 27, 29) and flavonoid aglycones (45,51,55) (Figure S8). Most lipophilic components were not extracted due to the hydrophilic nature of tea extractant, water. For the same reason, the quantifiable compounds in sagaan aya tea were six caffeoylquinic acids  (Table 3). The total content of caffeoylquinic acids and flavonoid glycosides in A. frigida herbal tea was 36.58 and 3.75 mg per 100 mL. Flavonoid aglycones were present in trace or undetectable concentrations. The two dominant phenolics of A. frigida herbal tea, 5-O-caffeoylquinic acid and 3,5-di-O-caffeoylquinic acid, were present at levels of 16.09 and 16.35 mg/100 mL of decoction, respectively. Together, these two compounds accounted for more than 88% of the total caffeoylquinic acid content and about 80% of the total phenolic content. A high caffeoylquinic acid content was previously found in decoctions of other Artemisia plants, such as A. capillaris Thunb., which had 15.72-27.69 mg/100 mL of 5-O-caffeoylquinic acid and 4.84-10.18 mg/100 mL of 3,5-di-O-caffeoylquinic acid [64], and A. campestris subsp. maritima, demonstrating a total caffeoylquinic acid content of 22.10 mg/100 mL [53]. At present, the known plant source of phenolic compounds in the everyday diet is coffee: the levels of total caffeoylquinic acids in soluble and instant coffee beverages range from 14.7 to 44.0 mg/100 mL [65]. We therefore conclude that A. frigida herbal tea solution can also be a good source of caffeoylquinic acids.
Caffeoylquinic acids are known powerful antioxidants with a wide range of activity [66]. Plant sources of caffeoylquinic acids, such as various Artemisia plant extracts, always provide a high level of antioxidant protection [21]. In order to confirm the antioxidant potential of A. frigida herbal tea, we used an HPLC-based bioactivity profiling assay consisting of a prechromatographic reaction of the plant sample with 2,2-diphenyl-1-picrylhydrazyl free radical (DPPH • ) ( Figure 5). The resulting two-dimensional chromatogram demonstrated the high radical scavenging potency of the selected compounds as peaks of reduced area. sources of caffeoylquinic acids, such as various Artemisia plant extracts, always provide a high level of antioxidant protection [21]. In order to confirm the antioxidant potential of A. frigida herbal tea, we used an HPLC-based bioactivity profiling assay consisting of a prechromatographic reaction of the plant sample with 2,2-diphenyl-1-picrylhydrazyl free radical (DPPH • ) ( Figure 5). The resulting twodimensional chromatogram demonstrated the high radical scavenging potency of the selected compounds as peaks of reduced area.  Table 1. In Figure 5. HPLC-DAD chromatograms (330 nm) of A. frigida herbal tea (a) before and (b) after prechromatographic reaction with DPPH • radicals. Compounds are numbered as listed in Table 1. In (b) red circles show the compounds with the highest scavenging capacity, and numbers demonstrate the percentage peak area decrease compared with the peak area in (a).
Of particular interest are the changes in phenolics that occur during the process of digestion carried out in different digestion media with various compositions, enzyme contents and pH levels [67]. Phenolic components of herbal beverages are generally considered to be stable in gastric and intestinal juices, but caffeoylquinic acids are known to be less resistant than flavonoid compounds [68]. Both classes of phytocomponents are known scavengers of free radicals [63], making it possible to use caffeoylquinic acids and flavonoids and their source, Artemisia plants, as prospective antioxidant agents [69]. There are no available data on the bioactivity of A. frigida and its preparations in connection with the digestive process transformation, so we studied the chemical and bioactivity changes that occurred in A. frigida herbal tea during in vitro digestion.
HPLC data on A. frigida herbal tea after gastric and intestinal media treatment demonstrated a 10.6% (32.67 mg/100 mL) and 35.2% (23.69 mg/100 mL) reduction of total caffeoylquinic acid content respectively, from initial levels ( Table 3). Flavonoid glycosides were more stable, as their concentrations decreased by 1.6-10.7%. Despite this reduction, however, the total phenolic content in A. frigida herbal tea after the intestinal phase of digestion was still high (27.04 mg/100 mL).
The total antioxidant capacity (TAC) of non-treated A. frigida herbal tea calculated as the sum of the capacities of hydrophilic (H-ORAC) and lipophilic (L-ORAC) antioxidants was 2918.77 µmol Trolox equivalents (TE) per 100 mL of beverage ( Figure 6). The capacity of hydrophilic antioxidants (2826.14 µmol TE/100 mL) was higher than that of lipophilic antioxidants (92.63 µmol TE/100 mL) due to the hydrophilic nature of water as the extractant. The process of digestive media treatment resulted in a reduction of the content of antioxidant phenolics and finally to a decrease in antioxidant capacity. the capacities of hydrophilic (H-ORAC) and lipophilic (L-ORAC) antioxidants was 2918.77 μmol Trolox equivalents (TE) per 100 mL of beverage ( Figure 6). The capacity of hydrophilic antioxidants (2826.14 μmol TE/100 mL) was higher than that of lipophilic antioxidants (92.63 μmol TE/100 mL) due to the hydrophilic nature of water as the extractant. The process of digestive media treatment resulted in a reduction of the content of antioxidant phenolics and finally to a decrease in antioxidant capacity. The index of TAC of gastric medium-treated A. frigida herbal tea was 11% less (2593.56 μmol TE/100 mL) than that of the untreated sample, but intestinal medium treatment resulted in a further decrease of up to 2090.14 μmol TE/100 mL. The reduction of hydrophilic antioxidant content in the beverage after digestive media treatment played a key role in the loss of general antioxidant activity. Despite the reducing effect of digestion phases on the phenolic content and activity of A. frigida herbal tea, the final levels of antioxidant activity remained high because of the appropriate content of bioactive components. Earlier data about in vitro digestion of Artemisia gorgonum Webb infusion also The index of TAC of gastric medium-treated A. frigida herbal tea was 11% less (2593.56 µmol TE/100 mL) than that of the untreated sample, but intestinal medium treatment resulted in a further decrease of up to 2090.14 µmol TE/100 mL. The reduction of hydrophilic antioxidant content in the beverage after digestive media treatment played a key role in the loss of general antioxidant activity. Despite the reducing effect of digestion phases on the phenolic content and activity of A. frigida herbal tea, the final levels of antioxidant activity remained high because of the appropriate content of bioactive components. Earlier data about in vitro digestion of Artemisia gorgonum Webb infusion also demonstrated a reduction of the phenolic compound content, but FRAP, CUPRAC and other antioxidant parameters were high after gastric and intestinal media treatment [70]. Despite the limited data on the digestive stability of Artemisia, one can assume that the antioxidant capacity of Artemisia extracts remains high because of their high hydroxycinnamate content.

Conclusions
In this work, Artemisia frigida as a new herbal source of hydroxycinnamates and flavonoids was studied, and its phenolic metabolic profile was analysed, focusing on caffeoylquinic acids, flavonoid glycosides and aglycones. The wide spectrum of phenolics was characterized, most of them being newly found in A. frigida and the Artemisia genus. The HPLC profile data on selected phenolics in A. frigida of Siberia origin exhibited high variability in the chemical patterns of the species, which requires consideration of how it might be reflected in the biological activity and practical use of a raw material. The new herbal beverage, A. frigida tea, was also investigated chemically, and most of the phenolics therein demonstrated high stability after in vitro digestion. The high antioxidant capacity of A. frigida herbal tea before and after digestive media treatment makes it a good candidate for use as a prophylactic and therapeutic remedy for various redox imbalances. Satisfactory sensory attributes of A. frigida tea and global interest in functional herbal beverages opens up new commercial avenues for Artemisia tea.
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-3921/8/8/307/s1, Table S1. Ethnopharmacological use of Artemisia frigida (fringed sagewort) by the various nomadic people; Table S2. Phenolic compounds of A. frigida (literature data); Table S3. Regression equations, correlation coefficients (r 2 ), standard deviation (S YX ), limits of detection (LOD), limits of quantification (LOQ) and linear ranges for 16 compounds; Table S4. Intra-and inter-day precision, repeatability, stability and recovery for 16 compounds; Table S5. Content of selected caffeoylquinic acids and flavonoids in the 21 samples of A. frigida herb; Table S6. Macronutrients, free sugars, organic acids, amino acids and mineral composition of A. frigida herbal tea; Table  S7. Demographic information of participants of tea sensory evaluation; Table S8. Sensory evaluation data of A. frigida herbal tea, green tea, black tea and Artemisia absinthium tea; Table S9. Phytochemical composition of A. frigida herbal tea; Figure S1. Mass spectra of compounds 1-59 found in A. frigida; Figure S2. Structures of reference standards used in present work; Figure S3. A series of three successive chromatograms of A. frigida extract demonstrating good reproducibility of the method used; Figure S4. HPLC-DAD chromatograms of A. frigida extract of leaves, flowers, stems and roots; Figure S5. HPLC-DAD chromatogram of the reference mixture of 16 pure compounds and internal standard; Figure S6. Calibration curves for 16 reference compounds used for quantitative HPLC-DAD assay; Figure S7. Sensory profiles of A. frigida herbal tea, green tea, black tea and Artemisia absinthium tea according to 30 participants estimation; Figure S8. HPLC-DAD chromatogram of A. frigida herbal tea.