Non-Volatile Terpenoids and Lipophilic Flavonoids from Achillea erba-rotta Subsp. moschata (Wulfen) I. Richardson

Musk yarrow (Achillea erba-rotta subsp. moschata (Wulfen) I. Richardson) is endemic to the Central Alps, and is used to flavour alcoholic beverages. Despite its popularity as aromatizing agent and its alleged beneficial effects on digestion, the phytochemical profile of the plant is still largely unknown and undiscovered. As a consequence, its authentication in aromatized products is impossible beyond sensory analysis allowing forgery. To address these issues, we phytochemically characterized a sample of musk yarrow from the Italian Eastern Alps, identifying, in addition to widespread phytochemicals (taraxasterol, apigenin), the guaianolides 3, 8, 9; the seco-caryophyllane 6; and the polymethoxylated lipophilic flavonoids 1, 4, and 5. The flavonoid xanthomicrol 1, a major constituent of the plant, was cytotoxic to HeLa cells, but only modestly affected primary 3T3 fibroblasts. On account of their stability, detectability by UV absorption, and concentration, the oxygenated flavonoids qualify as markers to validate the supply chain of the plant growers to consumers.


Introduction
Musk yarrow (Achillea erba-rotta subsp. moschata (Wulfen) I. Richardson, syn. A. moschata Wulfen, Asteraceae) is a perennial herb endemic to the Central Alps, where it grows on siliceous soils at altitudes higher than 1800 m above sea level, on cliffs and moraines [1][2][3]. Because of its digestive properties and its pleasant aroma, the plant is used in folk medicine and to aromatize alcoholic and non-alcoholic beverages. Various studies by Vitalini and colleagues have confirmed the gastroprotective and digestive activity of Achillea erba-rotta subsp. moschata, only evidencing, however, the presence of widespread volatile terpenoids and phenolics, leaving its phytochemical profile only marginally investigated and mainly focused on antioxidant and antibacterial aspects [4][5][6][7]. The closely related A. erba-rotta subsp. erba-rotta is a prolific producer of sesquiterpene lactones and lipophilic flavonoids, two classes of compounds associated with the functionality of the gastrointestinal system and anti-ulcer activity [8][9][10]. Given the biological profile of the latter taxon, the importance of providing correct identification [11], and the lack in information in biomarkers to identify unique characteristics, we therefore investigated the occurrence of compounds from these two classes also in the subsp. moschata, with the twofold aim of identifying their active constituent and assessing their suitability for chemotaxonomic studies and the authentication of the plant in finished products.
Only xanthomicrol 1 could be isolated in sufficient amount to sustain bioac study. To this purpose, HeLa cells, a cancer cell line derived from a human cervica thelioid carcinoma previously used to assess the cytotoxicity of flavonoids [25,26], employed. Figure 2 shows the viability, expressed as % of the control, induced by in tion for 24 h with different amounts (5-200 μM) of xanthomicrol 1 in HeLa cells by assay. In accordance with previous reports on other malignant cell lines [27][28][29] thomicrol 1 reduced HeLa cell viability already with the lowest dosage investigated bility reduction of 25% at 5 μM), outperforming eupatilin 10 [25] and artemetin 11 Microscopic observation of xanthomicrol 1-treated cells ( Figure 3) before MTT ass lowed for evidence of changes in HeLa cell morphologies after 24 h incubation wi spect to control cells from dose of 5 μM. Control HeLa cells were small and closely l to each other (packed), while the xanthomicrol 1 treatment induced a remarkable inc in the number of apoptotic cells (rounded cells) in a concentration-dependent ma Moreover, the occurrence of clear apoptotic bodies and cell debris was observed highest xanthomicrol 1 concentrations.
The decrease in HeLa cell numbers observed in the MTT assay on xanthomi could be the result of either cell cycle arrest or the induction of apoptosis. Therefor effect of xanthomicrol 1 on cell cycle progression of cancer HeLa cells was assess Scheme 1. The guaianolides biosynthesis through the ∆ 1,3 -diene [4 + 2] photocycloaddition of oxygen followed by rearrangement to a diepoxide leading to 8 and, after hydrolysis of one of the two epoxide functions, to 9.
Only xanthomicrol 1 could be isolated in sufficient amount to sustain bioactivity study. To this purpose, HeLa cells, a cancer cell line derived from a human cervical epithelioid carcinoma previously used to assess the cytotoxicity of flavonoids [25,26], were employed. Figure 2 shows the viability, expressed as % of the control, induced by incubation for 24 h with different amounts (5-200 µM) of xanthomicrol 1 in HeLa cells by MTT assay. In accordance with previous reports on other malignant cell lines [27][28][29], xanthomicrol 1 reduced HeLa cell viability already with the lowest dosage investigated (viability reduction of 25% at 5 µM), outperforming eupatilin 10 [25] and artemetin 11 [27]. Microscopic observation of xanthomicrol 1-treated cells ( Figure 3) before MTT assay allowed for evidence of changes in HeLa cell morphologies after 24 h incubation with respect to control cells from dose of 5 µM. Control HeLa cells were small and closely linked to each other (packed), while the xanthomicrol 1 treatment induced a remarkable increase in the number of apoptotic cells (rounded cells) in a concentration-dependent manner. Moreover, the occurrence of clear apoptotic bodies and cell debris was observed at the highest xanthomicrol 1 concentrations. G1 phase was observed in the 5 μM to 25 μM dosage range (Figure 4), diagnostic o increased number of apoptotic cells (sub-G1 population). The exposure of cells to x thomicrol 1 resulted in a dose-dependent accumulation of the proportion of cells in G2/M phase, and a clear cell cycle arrest at the G2/M phase was observed at 25 μM. Ta together, our results confirm that xanthomicrol 1 activates cell apoptosis and cell cy arrest, as previously observed in breast cancer cells [29].   thomicrol 1 resulted in a dose-dependent accumulation of the proportion of cells G2/M phase, and a clear cell cycle arrest at the G2/M phase was observed at 25 μM. together, our results confirm that xanthomicrol 1 activates cell apoptosis and cell arrest, as previously observed in breast cancer cells [29].    The decrease in HeLa cell numbers observed in the MTT assay on xanthomicrol 1 could be the result of either cell cycle arrest or the induction of apoptosis. Therefore, the effect of xanthomicrol 1 on cell cycle progression of cancer HeLa cells was assessed by flow cytometry. Dosages of 5, 10, and 2.5 µM were selected, due to their absence of toxicity to normal fibroblasts (Figure 2). In the event, an increased percentage of cells at the sub-G1 phase was observed in the 5 µM to 25 µM dosage range (Figure 4), diagnostic of an increased number of apoptotic cells (sub-G1 population). The exposure of cells to xanthomicrol 1 resulted in a dose-dependent accumulation of the proportion of cells in the G2/M phase, and a clear cell cycle arrest at the G2/M phase was observed at 25 µM. Taken together, our results confirm that xanthomicrol 1 activates cell apoptosis and cell cycle arrest, as previously observed in breast cancer cells [29].

Discussion
The phytochemical profile of Achillea erba-rotta subsp. moschata is different from the one of the subsp. erba-rotta [8]. Thus, the subsp. moschata contains sesquiterpene lactone of the guaiane-type, while the subsp. erba-rotta contains germacrane derivatives, and this profile was retained in all samples investigated, validating the chemotaxonomic value of this finding.
Interestingly, only the lipophilic flavonoid eupatilin 10 was isolated from A. erbarotta subsp. erba-rotta, which also contained significant amounts of artemetin 11 as well as the coumarin scopoletin. None of these compounds could be detected in the subsp. moschata, which contained the lipophilic flavonoids xanthomicrol 1, tanetin 4, and penduletin 5.
Sesquiterpene lactones are a major class of bitter compounds of dietary relevance. Their bitterness is associated to the activation of receptor hTAS2R46, a broadly-tuned taste receptor also targeted by the alkaloid strychnine [30], while their gastroprotective properties are critically associated to the presence of the electrophilic exomethylene-γ-lactone moiety [10]. Since the sesquiterpene lactones from the subsp. moschata are both bitter and electrophilic, it is tempting to associate the eupeptic properties of the plant with their presence. In addition, the lipophilic 5-hydroxylated flavonoid eupatilin 10, a compound

Discussion
The phytochemical profile of Achillea erba-rotta subsp. moschata is different from the one of the subsp. erba-rotta [8]. Thus, the subsp. moschata contains sesquiterpene lactone of the guaiane-type, while the subsp. erba-rotta contains germacrane derivatives, and this profile was retained in all samples investigated, validating the chemotaxonomic value of this finding.
Interestingly, only the lipophilic flavonoid eupatilin 10 was isolated from A. erba-rotta subsp. erba-rotta, which also contained significant amounts of artemetin 11 as well as the coumarin scopoletin. None of these compounds could be detected in the subsp. moschata, which contained the lipophilic flavonoids xanthomicrol 1, tanetin 4, and penduletin 5.
Sesquiterpene lactones are a major class of bitter compounds of dietary relevance. Their bitterness is associated to the activation of receptor hTAS2R46, a broadly-tuned taste receptor also targeted by the alkaloid strychnine [30], while their gastroprotective properties are critically associated to the presence of the electrophilic exomethylene-γlactone moiety [10]. Since the sesquiterpene lactones from the subsp. moschata are both bitter and electrophilic, it is tempting to associate the eupeptic properties of the plant with their presence. In addition, the lipophilic 5-hydroxylated flavonoid eupatilin 10, a compound structurally closely related to 1, 4, and 5, shows clinically relevant antiulcer activity [9,10], making possible a synergistic interaction between the two major classes of secondary metabolites from the plant.
None of these compounds were obtained in amounts sufficient to sustain an in vivo animal study of gastroprotection, but we have, nevertheless, investigated the cytotoxicity of the major polymethoxylated flavonoid (xanthomicrol, 1), since the related compounds artemetin 11 and eupatilin 10 showed significant cytotoxicity against malignant cells. The biological results showed that, while not qualifying as a significantly potent cytotoxic agent, xanthomicrol 1 nevertheless showed selectivity for cancer vs. primary cells, since normal fibroblasts (3T3 murine line) were significantly less sensitive to its activity. Polymethoxylated flavonoids are widespread in plants belonging to the genus Citrus L., and lipophilic citrus flavonoids have taken the lion's share of studies on this class of compounds. Citrus polymethoxylated flavonoids lack a free 5-hydroxyl, an element conversely present in asteraceous lipophilic flavonoids and critical in terms of pharmacodynamic and pharmacokinetic properties. Due to the hydrolytic and oxidative stability of polymethoxylated flavonoids, their point-like distribution in plants, and their strong UV-absorption properties, these compounds fully become qualified as markers for the presence of musk yarrow in alcoholic beverages, a first important step to providing commercial support to ongoing cultivation efforts.
Cell cultures: Human adenocarcinoma HeLa cell line and mouse 3T3 fibroblasts were obtained from the American Type Culture Collection (ATCC, Rockville, MD). Cells were grown in Dulbecco's modified Eagle's medium (DMEM) with high glucose, supplemented with 2 mM L-glutamine, penicillin (100 units/mL)-streptomycin (100 mg/mL), and foetal calf serum (FCS) (10% v/v), at 37 • C in a 5% CO 2 incubator. Subcultures of 3T3 and HeLa cells were grown in T-75 culture flasks and passaged with a trypsin-EDTA solution. Cell culture materials were purchased from Invitrogen (Milan, Italy).

Plant Material
Musk yarrow flowering tops were collected during summer of 2015 in the territory of the province of Trento, both from natural populations and cultivated plants, in both cases harvested at full blooming stage. A voucher specimen of the plant (AM-2015) is stored at Novara laboratories.

Statistical Analyses
Evaluation of the statistical significance of differences was performed using Graph Pad INSTAT software (GraphPad software, San Diego, CA, USA). Results were expressed as mean ± standard deviation (SD), and statistically significant differences were evaluated with p < 0.05 as a minimal level of significance. All data were preliminary assessed for normal distribution with Graph Pad INSTAT software. Multiple comparison of means groups was assessed by one-way analysis of variance (one-way ANOVA) followed by the Bonferroni multiple comparisons test to substantiate statistical differences between groups, whereas comparison of means between two groups was assessed by Student's unpaired t-test with Welch's correction, which does not require the assumption of equal variance between populations.

Conclusions
A heterogeneous sample composed of wild and cultivated flowering tops of musk yarrow from the Italian Eastern Alps was phytochemically profiled, identifying, in addition to widespread phytochemicals (taraxasterol, apigenin), the guaianolides 3, 8, 9; the secocaryophyllane 6; and the polymethoxylated lipophilic flavonoids 1, 4, and 5. The eupeptic properties of the plant could result from the synergistic activity of these compounds on sensory and inflammatory targets, while the stability and point-like distribution of the lipophilic flavonoids qualify them as ideal markers for validating the supply chain of the plant growers to consumers and can be applied as chemosystematic markers for taxonomical study and quality controls in finished products.

Funding:
We are grateful to MIUR for financial support to the groups in Novara and Naples (PRIN2017, Project 2017WN73PL, Bioactivity-directed exploration of the phytocannabinoid chemical space). Data Availability Statement: Not applicable.