UHPLC-MS Characterization and Biological Insights of Different Solvent Extracts of Two Achillea Species (A. aleppica and A. santolinoides) from Turkey

In the current study, Achillea santolinoides and Achillea aleppica aeral parts and root were extracted with ethyl acetate, methanol, and water. Detailed phytochemical profiles were obtained using UHPLC-MS, yielding the identification of hydroxybenzoic and hydroxycinnamic acids, phenolic acid glycosides and sugar esters, acylquinic acids, O-glycosyl flavones and flavonols, and flavonoid aglycons, among others. The antioxidant properties and enzyme inhibitory activities of the extracts were assayed with in vitro tests. The phenolic content of the water extracts was significantly higher as compared to the ethyl acetate and methanol ones. A. aleppica aerial parts methanol extract possessed highest flavonoid content (49.18 mg rutin equivalent/g). Antioxidant properties assessment revealed that the methanol extract of A. santolinoides roots actively scavenged DPPH (54.11 mg TE/g) and ABTS radicals (112.53 mg TE/g) and possessed highest reducing potential (183.55 and 129.92 mg TE/g, for CUPRAC and FRAP, respectively). The ethyl acetate extracts of aerial parts and roots of both species showed highest inhibition against BuCHE (6.07–6.76 mg GALAE/g). The ethyl acetate extract of A. santolinoides aerial part showed highest inhibition against tyrosinase (73.00 mg KAE/g). These results showed that the tested Achillea species might represent novel phytotherapeutic avenues for the management of Alzheimer’s disease and epidermal hyperpigmentation conditions, which are both associated with oxidative stress. This paper could shed light into future potential industrial applications using the tested Achillea species.


Introduction
The Achillea genus, one of the most important genera of the Asteraceae family with ethnopharmacological significance, consists of approximately 85 species mainly distributed in Middle East regions, such as Iran, Turkey, and Serbia and Eastern regions of Europe [1]. Achillea species have been reported to possess highly bioactive compounds and were rich in flavones and other flavonoids [2], non-saturated carboxylic acids [3], phenolic glycosides [4], guaianolides [5], lignans [6], phthalate derivatives [7], piperidine amides, proazulenes [8], sesquiterpenes [9], sesquiterpene lactone-diol [10], sesquiterpene lactones [11], polyacetylenes [12], spirodepressolide [13], tannins [14], and triterpene alkamide [15]. An UHPLC-HRMS analysis was performed as described elsewhere . Briefly, the separation was carried out on a reversed phase column Waters Cortecs C18 (2.7 µm, 2.1 × 100 mm) column maintained at 40 • C. The binary mobile phase consisted of 0.1% formic acid in water (A) and B: 0.1% formic acid in acetonitrile (B). The gradient program began at 5% B for one min, gradually turned to 30% B over 19 min, increased gradually to 50% B over 5min, increased gradually to 70% B over 5 min, increased gradually to 95% over 3 min and finally the system was then turned to the initial condition of 5% B, and equilibrated over 4 min. The flow rate and the injection volume were set to 300 µL/min and 1 µL, respectively. Samples were prepared as follows: methanol and aqueous extracts were dissolved in methanol-water (1:1, v/v) by ultrasound (20 µg/mL), while for the etylacetate extracts methanol was used preparing sample at the same concentrations. The solutions were filtered thought syringe filters 0.22 µm (Filtratech, France) and injected into chromatographic system.
Mass spectrometry analyses were carried out on a Q Exactive Plus mass spectrometer (ThermoFisher Scientific, Inc., Waltham, MA, USA) equipped with a heated electrospray ionization (HESI-II) probe (ThermoScientific, Waltham, MA, USA). The tune parameters were as follows: spray voltage −2.5 kV; sheath gas flow rate 38; auxiliary gas flow rate 12; spare gas flow rate 0; capillary temperature 320 • C; probe heater temperature 320 • C and Slens RF level 50. Acquisition was acquired at Full scan MS and Data Dependent-MS 2 modes. Full scan spectra over the m/z range 100 to 1500 were acquired in negative ionization mode at a resolution of 70,000. Other instrument parameters for Full MS mode were set as follows: automatic gain control (AGC) target 3 × 106, maximum injection time (IT) 100 ms, number of scan ranges one. For DD-MS 2 mode, instrument parameters were as follows: microscans 1, resolution 17,500, AGC target 1 × 105, maximum IT 50 ms, MSX count 1, Top5, isolation window 2.0 m/z, stepped normalized collision energy (NCE) 10, 20, 60 eV. Data acquisition and processing were carried out with Xcalibur 4.0 software (ThermoScientific, Waltham, MA, USA). All chromatograms and MS/MS data for each identified compound including fragmentation patterns are given in Supplemental Materials (Figures S1-S9).

Determination of Antioxidant and Enzyme Inhibitory Effects
Antioxidant protocols included reducing power (cupric reducing antioxidant capacity (CUPRAC) and ferric reducing power (FRAP)), metal chelating, phosphomolybenum and free radical scavenging (2,2-diphenyl-1-picrylhydrazyl (DPPH) and 3-ethylbenzothiazoline-6-sulphonic acid (ABTS)) activities. Trolox and ethylenediaminetetraacetic acid (EDTA) were used as standards in the antioxidant assays and the results were expressed as the equivalents of these standards. Experimental details were given in our previous paper [30].
Inhibitory effects of the extracts were tested against different enzymes (tyrosinase, α-amylase, α-glucosidase and cholinesterases (AChE and BuChE). Several compounds were used as standards (galatamine for cholinesterases; kojic acid for tyrosinase; acarbose for αamylase and α-glucosidase) and the results were expressed as the equivalents of Antioxidants 2021, 10, 1180 4 of 29 these standards. The enzyme inhibitory assays were performed as done in our earlier paper [31].

Statistical Analysis
Relative quantitative data of extracts molecules obtained from UHPLC-MS analysis was submitted to principal component analysis and Clustered Image Maps successively, for viewing the differential expression of molecules among extracts. Afterward, for biological, One-way ANOVA following by Tukey's test were performed to determine any differences between the extracts of each studied species. p < 0.05 were assigned to be statistically significant. Then, for comparison both species extracts biological activities, principal component analysis (PCA) and Clustered Image Maps was subsequently achieved. For both realized Clustered Image Maps, "Wards" and "Euclidean" were use as linkage rule and similarity measure, respectively. The relationship between metabolites and biological activities was investigated using partial least squared regression analysis. The goodness of the model was measured through the estimation of the cumulative modeled variation in the metabolite matrix R 2 X(cum) and the cumulative modeled variation in the biological activities matrix R 2 Y(cum). All statistical procedures were performed using R software v. 3.6.1.

Bioinformatics Analysis
To investigate the genes targeted by the sesquiterpene lactones and derivatives and some phenolic compounds, the datasets for mRNA of DIGEP-Pred web-sever [32] was employed. The compounds were artabsin, dehydroleucodin, dihydrosantamarin, leucodin, matricin, tanaparthin peroxide, neochlorogenic acid, chlorogenic acid, homoorientin, vitexin and isovitexin. Only the genes with Pa (probability "to be active") higher than 0.5 were retained. Then for KEGG pathway analysis, the obtained up-regulated and down-regulated mRNA data were submitted to Enrichr websever [33].

Chemical Profile
After the qualitative screening of metabolites profiles in the different extracts of the species, unsupervised principal component analysis (PCA) was carried out on the relative intensities of metabolites peak area obtained through UHPLC-MS analysis in order to screen the molecules variation between both species' samples. Before PCA processing, metabolites profiles were log transformed and autoscaled to ensure an equal contribution of variables in prediction outcomes. From the extracted principal components (PCs), only the first six showed eigenvalue above one. In addition, they displayed a cumulative proportion explained variance higher than 80%, therefore there were used as recommended by Kaiser [34]. The molecules strongly associated with each of them were summarized in Table S1. Overall, 18,9,5,5,10, and eight molecules had the highest contribution scores on PC1, PC2, PC3, PC4, PC5, and PC6, respectively.
Afterwards, looking at the different score plot displayed in Figure 1, a considerable difference between the samples was observed. On the other hand, despite some samples seemed have common characteristics, it was difficult to clearly identify the different samples. For this purpose, an additional analysis i.e., Clustered Image Maps was performed from the coordinates of the samples derived from PCA. The samples can be split into three main clusters, the cluster I and III comprised on five samples respectively and the cluster II was represented by two samples (Figure 2). Of these three clusters, the samples of the clusters I were remarkably rich in several molecules. Hence most of the molecules were occurred predominantly in the methanol and water extracts obtained from both species the aerial parts as well as the methanol extract of A. aleppica subsp. zederbaueri roots. This finding reflects the polar character of the molecules present in these two species.   The total phenolic and flavonoid contents were determined using Folin Ciocalteau and aluminum chloride colorimetric methods, respectively. In A. alleppica extracts, water extract of root possessed the highest level of total phenolic (43.24 mg GAE/g), while the methanol extract of root contained the highest amounts of total phenolic (52.07 mg GAE/g) in A. santolinoides extracts. On the other hand, methanol extract of A. aleppica aerial part and ethyl acetate extract of A. santolinoides aerial part were found to have the highest flavonoid content respectively (49.18 and 19.58 mg RE/g) ( Table 1). To identify the metabolites present in the studied extracts, non-targeted profiling was performed by ultra-high-performance liquid chromatography-quadrupol-Orbitrap high resolution mass spectrometry (UHPLC-HRMS). Under the conditions of Full scan-ddMS 2 /Top 5, the mass range for survey full scan was set at m/z 100-1200 and the MS/MS analyses were acquired by stepped higher energy collision-induced dissociation (hcd) at 10, 20, and 60 eV for data dependent (dd) MS 2 scans. The key points in the compounds annotation/dereplication were the accurate masses in Full MS and ddMS 2 , MS/MS fragmentation patterns, relative abundance of the precursor and fragment ions, elemental composition, matching with the simulated monoisotopic peak profiles, and consistence with the retentions times and fragmentation spectra of reference standards and literature data [35][36][37]. The chemical structures of main components are depicted in Figure 3.
A variety of metabolites were identified and tentatively elucidated in the assayed extracts, including, 14 hydroxybenzoic and hydroxycinnamic acids together with 12 phenolic acid glycosides and sugar esters, 18 acylquinic acids, 11 C-glycosyl flavones, 2 C, O-glycosyl flavones, 11 O-glycosyl flavones and flavonols, and 12 flavonoid aglycons, six sesquiterpene lactons, and five fatty acid amides (Table 2, Figure S1-S4). All compounds are reported for the first time in the studied Achillea sp.        ddMS 2 /Top 5, the mass range for survey full scan was set at m/z 100-1200 and the MS/MS analyses were acquired by stepped higher energy collision-induced dissociation (hcd) at 10, 20, and 60 eV for data dependent (dd) MS 2 scans. The key points in the compounds annotation/dereplication were the accurate masses in Full MS and ddMS 2 , MS/MS fragmentation patterns, relative abundance of the precursor and fragment ions, elemental composition, matching with the simulated monoisotopic peak profiles, and consistence with the retentions times and fragmentation spectra of reference standards and literature data [35][36][37]. The chemical structures of main components are depicted in Figure 3.

Figure 3.
The main components in the tested Achillea extracts (for the compound numbers see Table 2).
A variety of metabolites were identified and tentatively elucidated in the assayed extracts, including, 14 hydroxybenzoic and hydroxycinnamic acids together with 12 phenolic acid glycosides and sugar esters, 18 acylquinic acids, 11 C-glycosyl flavones, 2 C, O- The main components in the tested Achillea extracts (for the compound numbers see Table 2).
Among the compounds of the group, phenolic acid-hexosides 5, 15, and 16 were the major compounds in the aerial parts of both species ( Figure S1A,B), especially protocate-chuic acid-and 4-hydroxybenzoic acid-hexoside in A. santolinoides. Syringic acid-hexoside (6) was presented mainly in A. allepica roots (Table 2, Figure S1D). In addition, quinic acid was commonly found in all samples.

Sesquiterpene Lactones (STLs)
The dereplication of STLs was based on the fragmentation patterns and diagnostic ions in positive ion mode as more informative for this class of natural compounds [36,43]. Based on accurate masse in Full MS, MS/MS fragmentation patterns, relative abundance of precursor and fragment ions, and elemental composition, 6 STLs were tentatively annotated in Achillea extracts.

Antioxidant Effects
The total antioxidant capacity of the extracts was determined using the phosphomolybdenum assay. As shown in Table 1, for both species, the aerial part ethyl acetate extracts (2.33 and 1.95 mmol TE/g) showed the highest activity. Further antioxidant assays, free radical scavenging (DPPH and ABTS), reducing power (FRAP and CUPRAC), and metal chelating were conducted in order to obtain a comprehensive understanding of the antioxidant potential of the extracts and results were presented in Table 3. The ability of the extracts to scavenge free radicals was summarized in Table 3. Methanol extracts of A. aleppica aerial parts (55.15 mg TE/g) and A. santolinoides roots (54.11 mg TE/g) showed highest scavenging activity against DPPH. In contrast A. aleppica roots water extract (101.88 mg TE/g) and A. santolinoides roots methanol extract (112.53 mg TE/g) were most potent in scavenging ABTS. Protocatechuic acid and its derivatives identified in the A. aleppica roots water extract, A. aleppica aerial parts methanol extract, and A. santolinoides roots methanol extract, has been reported to exhibit radical scavenging activity [47,48]. Neochlorogenic (3-caffeoylquinic) acid also identified in these extracts was previously reported to exhibit scavenging activity against DPPH [49]. The reducing capacity of the extracts to donate electron and thus act as reducing agents is commonly assessed using two widely used methods, namely FRAP (ferric ion) and CUPRAC (cupric ion) assays. Similar to the DPPH assay, methanol extracts of A. aleppica aerial parts and A. santolinoides roots showed highest reducing capabilities (Table 3). The chelating capacity of the extracts was also evaluated. The water extract of the aerial parts of A. aleppica (25.37 mg EDTAE/g) and ethyl acetate and water extract of the aerial parts of A. santolinoides (27.37 and 26.06 mg EDTAE/g), respectively possessed strong chelating ability. Caffeic acid, chlorogenic acid, and protocatechuic acid were identified in aerial parts of A. aleppica water and A. santolinoides ethyl acetate extracts. Interestingly, a study conducted by Andjelković, et al. [50] has assessed the metal chelating potential of these phenolic compounds and reported that caffeic acid and chlorogenic acid were the strongest metal chelators. It can also be suggested that the presence of these metal chelators created a synergistic effect, therefore enhancing the metal chelating properties of these extracts. The hydroalcoholic extract of A. santolinoides was previously reported to possess antioxidant effect on brain tissues in pentylenetetrazoleinduced seizures Wistar rat models [51]. The essential oil of A. santolinoides was also found to exhibit antioxidant potential against DPPH radical (IC 50 = 129-372 mg/mL) [52].

Enzyme Inhibitory Effects
The inhibitory ability of extracts prepared from the aerial parts and roots of the selected Achillea species against enzymes targeted in the management of diabetes mellitus type II, Alzheimer's disease, and skin hyperpigmentation problems was investigated. Alzheimer's disease has escalated to epidemic proportions and the need for complementary therapeutic agents to effectively manage this debilitating condition is of paramount importance. From Table 4, A. aleppica aerial parts ethyl acetate extract and A. santolinoides roots methanol exhibited highest inhibition against AChE. A previous molecular docking study confirmed the interaction of orientin with AChE which showed least binding energy and highest binding affinity [53]. Vitexin also identified in these extracts was previously reported to bind effectively with AChE through strong hydrogen bonding [54]. Acacetin was previously reported to exhibit moderate to potential AChE inhibitory properties [55]. However, in the present study, acacetin was not identified in extracts showing more potent inhibitory activity against AChE. Santin/eupatilin identified in the ethyl acetate extracts of A. aleppica roots and A. santolinoides aerial parts was previously reported to inhibit BuChE in an in silico study. On the other hand, the ethyl acetate extracts of A. aleppica aerial parts and roots (6.07 and 6.73 mg GALAE/g) and as well as that of A. santolinoides aerial parts (6.76 and 6.70 mg GALAE/g) were most active against BuChE. The inhibition of BuChE has been advocated in the later stage of Alzheimer's disease. During the progression of the disease, BuChE level increases, exacerbating the conditions of the patient [56]. The ability of the extracts to inhibit enzymes targeted in the management of diabetes type II, namely α-amylase and α-glucosidase, was presented in Table 4. A low inhibition against both enzymes was noted, suggesting that the different extracts of A. aleppica and A. santolinoides aerial parts and roots possessed weak anti-diabetic properties. Tyrosinase, a rate limiting enzyme responsible for the biosynthesis of melanin, is considered to be a key therapeutic strategy for the management of skin hyperpigmentation conditions. In the present study, methanol extracts of A. aleppica aerial parts and roots showed the highest inhibitory activity against tyrosinase. In other side, ethyl acetate and methanol extracts of both studied parts of A. santolinoides displayed strongest anti-tyrosinase activity. Hispidulin, isolated from Phyla nodiflora and identified in extracts which actively inhibited tyrosinase was previously reported to exhibit inhibitory action against tyrosinase with an IC 50 value of 146 µM [57].

Data Mining
Subsequent to comparison of the bioactivities of the samples of each species, principal component analysis (PCA) was used in order to uncover the similarities/differences among the extracts of both species, in light of assessed antioxidant and enzyme inhibitory activities. The results of PCA were displayed in Figure 4. 88% variability of the data were captured by the first three Principal components (PCs) which each exhibited eigenvalue greater than 1. Therefore, these PCs were retained according to the method outlined by Kaiser [34]. By Referring to Sup 2, the first PC had higher correlation with more bioactivities, notably ABTS, DPPH, FRAP, CUPRAC, BuChE, amylase and glucosidase. The second PC was predominated by MCA, AChE and tyrosinase while the third PC was dominated by PBD and MCA. From the three score plots summarized in Figure 4A, a tendency to differentiate certain groups was noted. Hence, in PC1 vs. PC3 and PC2 vs PC3, extracts from A. aleppica roots EA and MeOH and A. santolinoides roots EA were grouped together. Similarly, in PC1 vs PC2 and PC1 vs. PC3, A. A. santolinoides roots MeOH and A. aleppica aerial parts MeOH were close together. Following PCA, a hierarchical classification was done to obtain a clearer picture of the different group. Based on the scores of samples on the three PCs, the hierarchical analysis revealed two principal clusters, each of which was divided into two sub-clusters ( Figure 4B). The samples of the first cluster (A. aleppica roots water, A. aleppica aerial parts water, A. santolinoides roots water, A. santolinoides aerial parts water, A. santolinoides roots MeOH and A. aleppica aerial parts MeOH) were characterized by higher antioxidant activity while samples of the second cluster (A. aleppica roots MeOH, A. A. santolinoides roots EA, A. wilhelmsii, A. aleppica roots EA, A. A. santolinoides aerial parts MeOH, A. A. santolinoides aerial parts EA and A. aleppica aerial parts EA) were marked by stronger enzyme inhibitory activity.

KEGG Analysis
After the phytochemical screening and in vitro evaluation of biological properties of the samples, we have been engaged in the investigation of KEGG pathway enrichment analysis of identified sesquiterpene lactones and derivatives and five of the main phenolics of Achillea species. In respect of the genes modulation, 73, 122, 280, 122, 113, 122, 57, 57, 254, 287 and 272 mRNA were found to be up-regulated and down-regulated by artabsin, dehydroleucodin, dihydrosantamarin, leucodin, matricin, tanaparthin peroxide, neochlorogenic acid, chlorogenic acid, homoorientin, vitexin and isovitexin respectively (Table S3). As regards the first enriched pathway, "hypertrophic cardiomyopathy", "longevity regulating pathway", "steroid hormone biosynthesis", "AMPK signaling pathway", "IL-17 signaling pathway" and "pathways in cancer" were found to be modulated by the mRNA targeted by artabsin, dehydroleucodin, dihydrosantamarin, leucodin, matricin, tanaparthin peroxide, neochlorogenic acid, chlorogenic acid, homoorientin, vitexin and isovitexin respectively ( Figure 6). Structure of these compounds are reported in Figure 3. Moreover, it is worth noting that "AMPK signaling pathway" was predicted to be regulated by nine compounds except neochlorogenic acid and chlorogenic acid. AMP-activated protein kinase (AMPK) is one of the central mediators of cellular and organismal metabolism. It has key roles in promoting catabolic pathways to produce more ATP and in inhibiting anabolic pathways [58]. Once activated, AMPK leads to a concomitant activation of ATP-producing catabolic pathways, such as glycolysis and fatty acid oxidation and inhibition of energyconsuming biosynthetic pathways, such as fatty acid, protein, and glycogen synthesis. Otherwise, AMPK is a known target for treating type-2 diabetes and metabolic syndrome and for reducing the incidence of cancer [59]. Sesquiterpenes lactones have been reported to induce an anticancer actions through an impact on multiple signaling pathways as well as a changes in the redox cell balance [60]. These effects lead to the increase in apoptotic factors and the reduction of metastasis, cellular invasion and anti-apoptotic factors. Illustratively, earlier study demonstrated the potentiality of matricin to significantly exert anti-proliferative and apoptosis-inducing effects in non-small cell lung cancer cells via activation of MAPK pathway [61]. Additionally, expression of anti-apoptotic proptein Bcl-2 was significantly decreased while the level of pro-apoptosis protein Bax as well as the activity of apoptosis marker enzymes caspase-9, caspase-8 and caspase-3 were significantly increased. Similarly, in a study of the anti-alcoholic liver disease activity of leucodin isolated from Artemisia capillaries, it has been demonstrated that leucodin dose dependently enhances phosphorylation of AMPK in alcohol-exposed HepG2 cells [62]. Furthermore, homoorientin has been demonstrated to have anti-pancreatic cancer activity via the AMPK signaling pathways [63]. While the literature has reported multiple biological mechanism, notably the regulation of AMPK pathway, to explain the pharmacological activities of vitexin and isovitexin [64]. This finding demonstrated that homoorientin, vitexin, isovitexin and both sesquiterpene lactones compounds can modulate AMPK signaling pathway. Hence, the nine compounds present in the different parts of both studied species could serve as AMPK activators and could be a promising candidate for the prevention and treatment of cancer. However, further studies on purified compounds will be necessary to confirm the conclusions of the present bioinformatics study.

Conclusions
This study allowed obtaining a detailed phytochemical fingerprint of A. aleppica and A. santolinoides roots and aerial part. Chlorogenic acid was the main derivative in aerial parts of both the species. 3,5-diCQA was the most important diCQA derivative in A. aleppica while 1,3diCQA was the most significant in A santolinoides. Sesquiterpene lactone and

Conclusions
This study allowed obtaining a detailed phytochemical fingerprint of A. aleppica and A. santolinoides roots and aerial part. Chlorogenic acid was the main derivative in aerial parts of both the species. 3,5-diCQA was the most important diCQA derivative in A. aleppica while 1,3diCQA was the most significant in A santolinoides. Sesquiterpene lactone and fatty acid amides have been also detected showing large chemical diversity in the constituents of the plant. The extraction with ethyl acetate, methanol and water allowed to prepare samples with different composition that were used to assess their in vitro bioactivity on several antioxidant and enzyme inhibition assays. The methanol extract of A. santolinoides roots possessed significant antioxidant activities. The ethyl acetate extracts of the aerial parts and roots of both Achillea species showed significant inhibition against butyrylcholinesterase while the ethyl acetate extract of A. santolinoides aerial part actively inhibited tyrosinase. The detailed phytochemical investigation, the evaluation of in vitro bioactivity, of the two Achillea species indicate these plants as valuable starting point for potential future studies and possible applications extracts in cosmetic, pharmaceuticals and nutraceuticals products. KEGG mapping using some of the phenolics and sesquiterpenes of the plants allowed to predict some of the possible molecular targets for significant bioactivities. This information opens new opportunities of research and application for A. aleppica and A. santolinoides extracts and isolated compounds.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/antiox10081180/s1, Table S1. Percentages of explained variances, eigenvalues and contribution of metabolites on the first six components of PCA. Table S2. Percentages of explained variances, eigenvalues and contribution of biological activities on the first three components of PCA. Table  S3. Upregulated and downreagulated mRNA by the six sesquiterpene lactones and some phenolic compounds. Figure S1: Extracted ion chromatogram of hydroxybenzoic and hydroxycinamic acids and derivatives in negative ion mode of methanolic extracts from Achillea wilhemsii aerial parts (A), A. allepica aerial parts (B), A. wilhemsii roots (C), A. allepica roots (D) (for the compound numbers see Table 2). Figure S2: Extracted ion chromatogram of acylquinic acids in negative ion mode of methanolic extracts from Achillea wilhemsii aerial parts (A), A. allepica aerial parts (B), A. wilhemsii roots (C), A. allepica roots (D) (for the compound numbers see Table 2). Figure S3: Extracted ion chromatogram of flavonoid glycosides in negative ion mode of methanolic extracts from Achillea wilhemsii aerial parts (A), A. allepica aerial parts (B), A. wilhemsii roots (C), A. allepica roots (D) (for the compound numbers see Table 2). Figure S4: Extracted ion chromatogram of flavonoid aglycones in negative ion mode of methanolic extracts from Achillea wilhemsii aerial parts (A), A. allepica aerial parts (B), A. wilhemsii roots (C), A. allepica roots (D) (for the compound numbers see Table 2). Figure S5: MS/MS spectra of hydroxybenzoic, hydroxycinnamic acids and their glycosides, and sugar esters (for the compound numbers see Table 2). Figure S6: MS/MS spectra of acylquinic acids (for the compound numbers see Table 2). Figure S7: MS/MS spectra of C-, C,Oand O-flavonoid glycosides(for the compound numbers see Table 2). Figure S8: MS/MS spectra of 6 methoxyflavonoids (aglycones and glycosides) (for the compound numbers see Table 2). Figure S9: MS/MS spectra of sesquiterpene lactones (for the compound numbers see Table 2).