Unveiling the Phytochemical Profile and Biological Potential of Five Artemisia Species

The Artemisia L. genus comprises over 500 species with important medicinal and economic attributes. Our study aimed at providing a comprehensive metabolite profiling and bioactivity assessment of five Artemisia species collected from northeastern Romania (A. absinthium L., A. annua L., A. austriaca Jacq., A. pontica L. and A. vulgaris L.). Liquid chromatography–tandem high-resolution mass spectrometry (LC-HRMS/MS) analysis of methanol and chloroform extracts obtained from the roots and aerial parts of the plants led to the identification of 15 phenolic acids (mostly hydroxycinnamic acid derivatives), 26 flavonoids (poly-hydroxylated/poly-methoxylated flavone derivatives, present only in the aerial parts), 14 sesquiterpene lactones, 3 coumarins, 1 lignan and 7 fatty acids. Clustered image map (CIM) analysis of the phytochemical profiles revealed that A. annua was similar to A. absinthium and that A. pontica was similar to A. austriaca, whereas A. vulgaris represented a cluster of its own. Correlated with their total phenolic contents, the methanol extracts from both parts of the plants showed the highest antioxidant effects, as assessed by the DPPH and ABTS radical scavenging, CUPRAC, FRAP and total antioxidant capacity methods. Artemisia extracts proved to be promising sources of enzyme inhibitory agents, with the methanol aerial part extracts being the most active samples against acetylcholinesterase and glucosidase. All Artemisia samples displayed good antibacterial effects against Mycobacterium tuberculosis H37Ra, with MIC values of 64–256 mg/L. In conclusion, the investigated Artemisia species proved to be rich sources of bioactives endowed with antioxidant, enzyme inhibitory and anti-mycobacterial properties.


Introduction
Artemisia L. is a genus of small herbs and shrubs belonging to the Asteraceae family which inhabit the northern temperate regions of Asia, Europe and North America [1]. The Artemisia genus comprises over 500 species with significant medicinal and economic attributes due to their biological and chemical diversity [2]. Artemisia species are recognized for their characteristic strong aromas and bitter tastes, which are assigned to the presence of terpenes and sesquiterpene lactones [3]. Nonetheless, other classes of phenolic unveiled other important pharmacological properties, including antibacterial [29,30], antifungal [24,27,29,31] and antimalarial activities [24].
A. vulgaris L., common mugwort, is a species growing in the temperate and coldtemperature regions of Asia, Europe and North America, and has been employed as a culinary and medicinal herb [2]. The aerial parts have been traditionally used as a bitter tonic and anti-flatulent in treating gastrointestinal disorders and to alleviate gynecological ailments, such as amenorrhea or dysmenorrhea [38]. The main phytochemicals found in A. vulgaris include essential oil (up to 0.3%, comprising 1,8-cineole, sabinene, βthujone and caryophyllene oxide as the main constituents), sesquiterpene lactones (vulgarin, psilostachyin and psilostachyin C), flavonoids (derivatives of kaempferol and quercetin), coumarins (coumarin, esculin, scopoletin and umbelliferone), phenolic acids (caffeic and chlorogenic acids), sterols, carotenoids and polyacetylenes [4,9,17]. To date, studies on A. vulgaris have confirmed its known traditional uses and revealed novel significant biological properties, e.g., antioxidant, spasmolytic, antibacterial, antifungal, antinociceptive, hepatoprotective, estrogenic and cytotoxic effects [2,17].
Our study aimed at promoting interest in the Romanian Artemisia species by providing novel insights into their metabolite profiles and bioactivities. To date, most studies have focused on the valorization of the aboveground parts with respect to their analgesic [35], anti-inflammatory [35,39], antimicrobial [40][41][42], antioxidant and cytotoxic properties [39,43,44]. To the best of our knowledge, we report herein for the first time a comprehensive phytochemical characterization of both the roots and aerial parts of five Artemisia spp. (A. absinthium, A. annua, A. austriaca, A. pontica and A. vulgaris) from the spontaneous flora of northeastern Romania by means of liquid chromatography-tandem high-resolution mass spectrometry (LC-HRMS/MS). The biological profile screening was achieved by in vitro testing of antioxidant (free radical scavenging, metal chelating and reducing power, and total antioxidant capacity), enzyme inhibitory (anti-cholinesterase, anti-tyrosinase, anti-amylase and anti-glucosidase) and anti-Mycobacterium activities.

Total Phenolic and Flavonoid Content
Total phenolic content (TPC) and total flavonoid content (TFC) were determined as previously described [45,46], with the data provided as mg gallic acid equivalents (GAE)/g extract (TPC) and mg rutin equivalents (RE)/g extract (TFC), respectively.

Inoculum Preparation
Mycobacterium tuberculosis H37Ra (ATCC 25177) was grown for two weeks on Löwenstein-Jensen slopes. The collected bacteria were transferred to 7H9-S medium (Middlebrook 7H9 broth supplemented with 10% ADC (albumin-dextrose-catalase) and 0.2% glycerol and vortexed with glass beads (1 mm diameter) for three minutes. After 30 min of room temperature incubation for larger clump sedimentation, the upper phase was transferred to a sterile tube and left for the second sedimentation for 15 min. Next, planktonic bacteria from above the sediment were placed in a fresh tube, with the turbidity adjusted to 0.5 McFarland standard with 7H9-S broth.

MIC Determination
Artemisia extracts were tested in a concentration range of 256 to 16 mg/L. Serial twofold dilutions were prepared in dimethyl sulfoxide (DMSO) using a 7H9-S medium as dilution. The final DMSO concentration did not exceed 1% (v/v) and did not influence the growth of the tested strain. Ethambutol, rifampicin and streptomycin were used as reference standards. Stock solutions were prepared according to the manufacturer's instructions. Final twofold dilutions from 16 to 0.001 mg/L were prepared in 7H9-S broth. The round bottom micro-well plates were prepared as follows: 50 µL of inoculum and 50 µL of tested substances were added to each well. The sterility, growth and 1% DMSO controls were included. The final density of the inoculum in each well was approximately 5 × 10 5 CFU/mL. The plates were closed with sealing foil to prevent liquid evaporation and incubated for 8 days at 37 • C. Next, 10 µL of resazurin (Alamar Blue) solution was added to each well, followed by incubation for 48 h at 37 • C and assessment for color development. The minimum inhibitory concentration (MIC) was defined as the lowest drug concentration that prevented a blue to pink color change. The MIC determination was repeated twice. The obtained results were identical.

Data Analysis
A biological activities dataset was scaled, centered and submitted to principal component analysis (PCA) and hierarchical clustering analysis (HCA). For both PCA and HCA, "Ward's rule" and "Euclidean distance" were employed for clustering. Afterwards, the biomolecules dataset was logarithm-transformed, scaled, centered and submitted to clustered image maps (CIMs). All multivariate analyses were performed using R v 4.1.2 software (R Foundation for Statistical Computing, Vienna, Austria). The Pearson correlation test was used to examine the relationship between phytoconstituents in tested extracts and biological activities. GraphPad. 9.0 (GraphPad Software, San Diego, CA, USA) was used for the correlation analysis.

Total Phenolic and Flavonoid Content
The TPCs and TFCs of the Artemisia extracts were determined using colorimetric methods. The results are given in Table 1. Apparently, methanol extracts contained more phenolics and flavonoids than chloroform in both plant parts. In addition, with the exception of A. pontica, the extracts of the aerial parts were richer than those of the roots in terms of total phenolics. The highest level of phenolics were determined in the methanol aerial part extracts of A. vulgaris, with 106.34 mg GAE/g. After that, the methanol aerial part extracts of A. pontica and A. annua contained significant levels of total phenolics (>50 mg GAE/g). Among the root extracts, the methanol extract from A. annua reached the highest value with 76.34 mg GAE/g, followed by the methanol extracts from A. pontica (65.65 mg GAE/g) and A. austriaca (41.68 mg GAE/g). With regard to the TFCs, the highest content was determined in the methanol extract of A. annua aerial parts with 47.74 mg RE/g, followed by the methanol extracts of A. austriaca (40.30 mg RE/g) and A. vulgaris (39.39 mg RE/g) aerial parts. The lowest level of total flavonoids was found in the chloroform extract of A. absinthium (0.37 mg RE/g). A number of studies have shown a comparable total content of phenolics and flavonoids in the Artemisia genus. For example, in a previous study, Ali et al. [48] found that the total phenolic content in A. absinthium extract was 3.61 mg GAE/g extract, which was lower than our values. In addition, Guo et al. [49] showed that the TPC and TFC in the aqueous extract of A. annua were 39.58 mg GAE/g and 7.04 mg RE/g, respectively. Our findings are also comparable to the results in the literature for other Artemisia species, such as A. copa (155.6 mg GAE/g dry plant in infused extract, reported by Larrazábal-Fuentes et al. [50]), A. vulgaris (117.14 mg GAE/g extract in methanol extract, reported by Jakovljevic et al. [51]), A. alba (110.20 mg GAE/g extract in methanol extract, reported by Jakovljevic et al. [51]) and A. argy (108.56 mg GAE/g extract in methanol extract, reported by Xiao et al. [52]). Although the spectrophotometric methods are widely used in phytochemical studies, recently, most phytochemists have been more concerned with colorimetric methods for assessing bioactive components [53]. This could be explained by the complex nature of phytochemicals and the fact that only specific compounds do not reduce the reagents used in the relevant assays. Given these facts, further chromatographic techniques are needed to evaluate the chemical profiles of plant extracts.

LC-HRMS/MS Analysis
The methanol and chloroform extracts obtained from the roots and aerial parts of the five Artemisia species were subsequently subjected to an in-depth LC-HRMS/MS analysis. The assignment of the peaks observed in the base peak chromatograms (BPCs) of the extract samples was performed by comparing the spectrometric data with the relevant literature [54][55][56][57][58][59][60][61] or online databases (KNApSacK; METLIN; NIST Chemistry WebBook). The metabolite profiling allowed the annotation of 73 compounds belonging to different phytochemical classes, such as phenolic acids, flavonoids, sesquiterpenes, organic acids, sugars, coumarins, triterpenes, lignans and fatty acids ( Table 2, Table S1). In the following sub-sections, a brief description of these categories will be provided, whereas the intra-and interspecies differences will be thoroughly detailed in the Multivariate Analysis Section.

Antioxidant Activity
Antioxidant compounds are of increasing interest in the pharmaceutical and nutraceutical fields. These compounds provide powerful shields against free radicals, and a negative correlation between their consumption and the prevalence of chronic and degenerative diseases has been reported. In this sense, phytochemicals are considered a great treasure trove of antioxidants, and many compounds found in plants have been identified as natural and safe antioxidants. In the light of these facts, attempts were made to determine whether the tested Artemisia species are a source of natural antioxidants. Various chemical assays were performed, including radical quenching (ABTS and DPPH), reducing power (CUPRAC and FRAP), metal chelation and phosphomolybdenum assays. The results are shown in Table 3. Non-biological radicals, such as DPPH and ABTS, are commonly used in in vitro experiments to assess the abilities of plant extracts to scavenge radicals. From Table 3, the methanol extracts showed stronger radical scavenging abilities than the chloroform extracts in both parts. The best radical scavenging ability was found in the methanol extract of A. annua roots (DPPH: 237.03 mg TE/g; ABTS: 240.78 mg TE/g), followed by the methanol extracts of A. pontica roots (DPPH: 179.63 mg TE/g; ABTS: 176.12 mg TE/g) and A. vulgaris aerial parts (DPPH: 139.56 mg TE/g; ABTS: 173.86 mg TE/g) in both assays. The weakest radical scavenging ability was recorded in the chloroform extract of A. absinthium roots (DPPH: 5.11 mg TE/g; ABTS: 7.54 mg TE/g). The term "reducing power" refers to the ability of antioxidant compounds to donate electrons. For this purpose, CUPRAC and FRAP assays involving the conversion of Cu +2 to Cu + and Fe +3 to Fe +2 , respectively, were performed. In both plant parts, the methanol extracts had higher reducing potentials than the chloroform extracts. The methanol extracts of A. vulgaris aerial parts (498.32 mg TE/g), A. annua roots (438.43 mg TE/g) and A. pontica aerial parts (290.14 mg TE/g) exhibited the highest CUPRAC activities. With one small exception, the methanol extracts of A. annua roots (294.52 mg TE/g), A. vulgaris aerial parts (198.51 mg TE/g) and A. pontica roots (165.55 mg TE/g) demonstrated the highest levels of capability in the FRAP assay. The obtained results from the free radical scavenging and reducing power assays are almost consistent with the total phenolic results for the extracts. In this sense, the phenolic components in the extracts can be considered as the main contributors to the free radical scavenging and reducing abilities. Similar to our findings, several researches [62,63] reported a strong correlation between total phenolic content and antioxidant properties. However, we observed different results for metal chelation abilities. The chelating abilities of plant extracts may reflect the inhibition of hydroxyl radicals' production in the Fenton reaction. The best metal chelating ability (MCA) was found in the methanol extract of A. pontica roots with 22.93 mg EDTAE/g extract. Intriguingly, in three Artemisia species tested, the chloroform extracts from the aerial parts had higher potentials than the methanol extracts (A. annua, A. pontica and A. vulgaris). In addition, two chloroform extracts of the roots (A. annua and A. vulgaris) showed no metal chelating ability. Non-phenolic chelators, such as polysaccharides, peptides or sulfides, may be responsible for the conflicting results. In support of our findings, several investigators reported a weak correlation between total phenolic content and metal chelating ability [64,65]. Some researchers also pointed out that the chelating ability of phenolics contributes only in a small extent to the antioxidant properties of plant extracts [66]. The phosphomolybdenum assay is related to the reduction of Mo (VI) to (Mo (V) by antioxidant compounds and is considered one of the total antioxidant capacity assays. As can be seen in Table 3, we observed different results for each species. For example, the aerial part extracts from two Artemisia species (A. absinthium and A. vulgaris) exhibited greater potentials than root extracts. In addition, the root extracts from two species (A. annua and A. pontica) showed stronger activity as compared to aerial parts. In the literature, several researchers have reported that the results from phosphomolybdenum assays exhibited weak correlations with total phenolic content [67,68]. This fact could be explained by the presence of non-phenolic antioxidants, such as tocopherol, ascorbic acid and terpenoids. Artemisia members have been previously found to possess interesting antioxidant properties. For example, in a recent study by Minda et al. [69], the DPPH radical scavenging abilities of three Artemisia species (A. absinthium, A. dracunculus and A. annua) were investigated, the materials exhibiting more than 90% scavenging ability at a concentration of 1000 µg/mL. In another study conducted by Kamarauskaite et al. [70], the fractions of A. absinthium and A. ludoviciana were assessed by ABTS and FRAP assays and their values were found to be 367-1693 µM TE/g and 5385-6952 µM TE/g, respectively. Ferrante et al. [71] also investigated the antioxidant properties of A. santonicum methanol extract (DPPH: 278.57 mg TE/g; ABTS: 217.60 mg TE/g; CUPRAC: 515.30 mg TE/g; FRAP: 255.35 mg TE/g; metal chelating: 21.96 mg EDTAE/g and phosphomolybdenum: 2.20 mmol TE/g). Other examples of Artemisia species whose antioxidant capacities have been determined in the literature include A. lactiflora [47], A. indica [72], A. santolinifolia [73] and A. monosperma [74]. Based on the solvents, plant parts, and species triangle, our results may provide new information on the antioxidant properties of Artemisia species.

Enzyme Inhibitory Activity
Enzyme inhibition is a concept that is currently gaining traction in the treatment of various global health problems, such as type 2 diabetes, obesity and Alzheimer's disease. This phenomenon demonstrates that the inhibition of specific enzymes can be a highly effective therapeutic strategy to alleviate disease symptoms [75]. Amylase and glucosidase, for example, are thought to be important players in the management of blood glucose levels in diabetics [76]. Furthermore, acetylcholinesterase inhibitors may improve memory function in Alzheimer's patients by increasing acetylcholine levels in the synapses [77]. Enzyme inhibitors, therefore, are being sought as a safe and effective way to treat the diseases listed above. In this sense, plants are considered excellent treasures [78]. Given these facts, we looked into the tested Artemisia species' enzyme inhibitory properties. The results are summarized in Table 4. The best AChE inhibition was determined in the methanol extract of A. absinthium with 3.02 mg GALAE/g, followed by the chloroform extracts of A. absinthium (2.50 mg GALAE/g) and A. annua (2.36 mg GALAE/g) aerial parts. With regard to BChE inhibition, the chloroform root extracts (A. annua, A. austriaca and A. absinthium) were recorded as the strongest extracts. With the exception of A. vulgaris roots, the chloroform extracts were more active against BChE than the methanol extracts in all tested Artemisia species. Two methanol root extracts (A. austriaca and A. pontica) were not active on BChE. As can be seen from Table 4, tyrosinase inhibitory effects were higher in the methanol extracts as compared to the chloroform extracts, except for A. vulgaris aerial parts. The most active methanol extracts were A. annua (49.42 mg KAE/g), A. austriaca (47.27 mg KAE/g) and A. pontica (44.91 mg KAE/g). The weakest tyrosinase inhibition potential was found in the chloroform extract of A. austriaca roots with 13.16 mg KAE/g. For all parts and species, the chloroform extracts had stronger amylase inhibitory effects than the methanol extracts. The best amylase inhibitory effects were recorded in the chloroform extracts of A. austriaca parts (root: 0.57 mmol ACAE/g; aerial parts: 0.54 mmol ACAE/g). As for glucosidase inhibitory activity, the aerial parts of all Artemisia species showed stronger abilities than the root extracts, and the best ability was obtained by the methanol extract of A. vulgaris aerial parts (11.32 mmol ACAE/g). The methanol root extract of A. austriaca had the weakest glucosidase inhibitory effect (0.16 mmol ACAE/g). To the best of our knowledge, scientific information on enzyme inhibitory properties of the members of the genus Artemisia is scarce [56,71,[79][80][81][82][83]. In this sense, the present work could provide further insights into the future application of Artemisia species as natural sources of enzyme inhibitory agents. Data are presented as mean ± standard deviation (SD) of three determinations; different superscript letters within columns indicate significant differences in the tested extracts for the same parts (p < 0.05). ACAE, acarbose equivalents; AChE, acetylcholinesterase; BChE, butyrylcholinesterase; GALAE, galanthamine equivalents; KAE, kojic acid equivalents; n.a., not active.

Anti-Mycobacterium Activity
The emergence of multidrug-resistant Mycobacterium tuberculosis strains represents a major barrier to tuberculosis eradication, leading to longer treatment regimens, higher toxicity and even treatment failure [84]. Thus, there is an urgent demand to explore novel drugs and combinations to improve tuberculosis therapy. Recent work has presented the antimalarial drug artemisinin as a promising antitubercular agent [85,86]. Moreover, Martini et al. found that dichloromethane extracts from leaves of A. annua L. and A. afra Jacq. ex Willd. displayed even higher anti-mycobacterial effects than sesquiterpene artemisinin [87]. In the present study, M. tuberculosis H37Ra was exposed to chloroform and methanol extracts obtained from the roots and aerial parts of five Artemisia species. Within the tested concentration range, most of the samples were active against the mycobacterial strain, except for the methanol root extracts of A. absinthium, A. austriaca and A. vulgaris ( Table 5). The chloroform extract of A. austriaca aerial parts showed the highest anti-mycobacterial effect (MIC = 64 mg/L), while the other active extracts displayed similar degrees of potency, with MICs of 128-256 mg/L. Considering that plant extracts can be categorized as having strong activity when their MIC is within the range of 50-500 mg/L, moderate activity with an MIC of 500-1500 mg/L and weak activity with an MIC above 1500 mg/L [88], it can be stated that the Artemisia samples possess strong anti-Mycobacterium effects. To the best of our knowledge, previous studies referred only to the anti-mycobacterial potential of Artemisia herbal extracts. We report herein for the first time on the anti-mycobacterial activity of Artemisia root extracts and that their potency was found to be similar to that of the aerial part extracts. Our results are in agreement with the study of Bhowmick et al. [89] which showed the inhibitory effects of an hexane extract from A. annua aerial parts against Mycobacterium smegmatis (MIC range: 250-1000 mg/L). Further, the bioactivity-linked fractionation of the extract revealed that its inhibitory effects were due to the sesquiterpenes deoxyartemisinin and artemisinic acid [89]. Regarding our Artemisia species, the LC-MS/MS analysis showed that chrysartemin A and B were present only in the most active sample-A. austriaca herbal extract (Table S1). These two compounds, which were previously reported in other Artemisia species (A. mexicana and A. klotzchiana) [90], belong to the guaianolide-type sesquiterpenes known for their inhibitory effects against Mycobacterium strains [90,91]. Therefore, we can conclude that sesquiterpenes found in Artemisia extracts may contribute to their overall anti-mycobacterial activity; moreover, it cannot be excluded that the synergistic effects among different classes of constituents identified in our samples (e.g., sesquiterpenes, phenolic acids, flavonoids, coumarins, fatty acids) might explain the observed effects. Table 5. Anti-Mycobacterium tuberculosis H37Ra activity of Artemisia spp. extracts.

MIC (mg/L)
in the metal chelation assay. In the phosphomolybdenum assay, artabsinolide A (17) and absinthin derivative II (71) were found to be the main contributors. With regard to the enzyme inhibition assays, different compounds acted as inhibiting agents in each assay. Tuberonic acid-O-hexoside (10) was moderately correlated with cholinesterases (AChE and BChE). For tyrosinase inhibition, dicaffeoylquinic acid derivatives (30 and 32) had low correlation values (R < 0.4). Dihydroxytrimethoxyflavone (52) and dihydroxytetramethoxyflavone (61) were closely associated with glucosidase inhibition assay (R > 0.7). Two sesquiterpenes (chrysartemin A (13) and B (15)) showed stronger correlations with anti-Mycobacterium ability compared to other compounds. Taken together, the tested extracts have great potential as natural sources of bioactive agents and could therefore be considered as valuable raw materials with pharmaceutical and nutraceutical applications.
assay, artabsinolide A (17) and absinthin derivative II (71) were found to be the main contributors. With regard to the enzyme inhibition assays, different compounds acted as inhibiting agents in each assay. Tuberonic acid-O-hexoside (10) was moderately correlated with cholinesterases (AChE and BChE). For tyrosinase inhibition, dicaffeoylquinic acid derivatives (30 and 32) had low correlation values (R < 0.4). Dihydroxytrimethoxyflavone (52) and dihydroxytetramethoxyflavone (61) were closely associated with glucosidase inhibition assay (R > 0.7). Two sesquiterpenes (chrysartemin A (13) and B (15)) showed stronger correlations with anti-Mycobacterium ability compared to other compounds. Taken together, the tested extracts have great potential as natural sources of bioactive agents and could therefore be considered as valuable raw materials with pharmaceutical and nutraceutical applications.  Table 2.
The results of the PCA for the antioxidant and enzyme inhibitory activities of Artemisia species are presented in Figure 2. Firstly, the screening of the eigenvalues suggested that the first three principal components (PCs) were sufficient to synthesize most of the data variation. Indeed, these components manifested a variance of 45%, 21% and 11%, respectively. The first PC represented the variation in the antioxidant activities, since it was predominantly and negatively linked with both radical scavenging (ABTS and DPPH) and reducing power (FRAP and CUPRAC) (Figure 2A). The second PC discriminated the samples based on their anti-glucosidase, phosphomolybdenum, antiamylase and anti-BChE activities ( Figure 2B), as it was significantly and positively bound to the mentioned bioactivities. The third PC separated the samples in terms of their metal chelating capacity and anti-tyrosinase activities ( Figure 2C). Examination of the scatter plots (PC1 vs. PC2, PC1 vs. PC3 and PC2 vs. PC3) reported in Figure 2D-F evidenced considerable variability among the samples. Three groups seem to emerge in the first scatter plot (PC1 vs. PC2) ( Figure 2D); the same trend is noticed in the second scatter plot (PC1 vs. PC3); however, the samples representing the three groups were different from those obtained previously ( Figure 2E). In the third scatter plot (PC2 vs. PC3), no clear clusters were identified ( Figure 2F). For a better identification of the different groups, HCA was applied by using the coordinates of the samples on the three dimensions of the PCA. By using Ward's method and Euclidean distance, we obtained three main clusters (Figure 3). Among the clusters, the samples representing cluster 1 (i.e., the methanol root extracts of A. annua, A. pontica and aerial part extracts of A. vulgaris) were characterized by the highest radical scavenging (ABTS and DPPH) and reducing power (FRAP and CUPRAC) activities.  Table 2.
The results of the PCA for the antioxidant and enzyme inhibitory activities of Artemisia species are presented in Figure 2. Firstly, the screening of the eigenvalues suggested that the first three principal components (PCs) were sufficient to synthesize most of the data variation. Indeed, these components manifested a variance of 45%, 21% and 11%, respectively. The first PC represented the variation in the antioxidant activities, since it was predominantly and negatively linked with both radical scavenging (ABTS and DPPH) and reducing power (FRAP and CUPRAC) (Figure 2A). The second PC discriminated the samples based on their anti-glucosidase, phosphomolybdenum, anti-amylase and anti-BChE activities ( Figure 2B), as it was significantly and positively bound to the mentioned bioactivities. The third PC separated the samples in terms of their metal chelating capacity and anti-tyrosinase activities ( Figure 2C). Examination of the scatter plots (PC1 vs. PC2, PC1 vs. PC3 and PC2 vs. PC3) reported in Figure 2D-F evidenced considerable variability among the samples. Three groups seem to emerge in the first scatter plot (PC1 vs. PC2) ( Figure 2D); the same trend is noticed in the second scatter plot (PC1 vs. PC3); however, the samples representing the three groups were different from those obtained previously ( Figure 2E). In the third scatter plot (PC2 vs. PC3), no clear clusters were identified ( Figure 2F). For a better identification of the different groups, HCA was applied by using the coordinates of the samples on the three dimensions of the PCA. By using Ward's method and Euclidean distance, we obtained three main clusters (Figure 3). Among the clusters, the samples representing cluster 1 (i.e., the methanol root extracts of A. annua, A. pontica and aerial part extracts of A. vulgaris) were characterized by the highest radical scavenging (ABTS and DPPH) and reducing power (FRAP and CUPRAC) activities.  Next, to determine the phytochemical differences between the studied Artemisia species, CIM analysis with respect to the phytochemical compounds dataset was carried out. The extracts were separated into two large clusters, namely, the roots on the one   Next, to determine the phytochemical differences between the studied Artemisia species, CIM analysis with respect to the phytochemical compounds dataset was carried out. The extracts were separated into two large clusters, namely, the roots on the one Next, to determine the phytochemical differences between the studied Artemisia species, CIM analysis with respect to the phytochemical compounds dataset was carried out. The extracts were separated into two large clusters, namely, the roots on the one hand and the aerial parts on the other (Figure 4). Overall, several compounds were more abundant in the extracts from the aerial parts than in the extracts from the roots. Furthermore, in cluster 2, which represented all the aerial part samples, the methanol and chloroform extracts of all species were very similar. On the other hand, in cluster 2, A. annua extracts were similar to A. absinthium extracts and A. pontica extracts were similar to A. austriaca extracts, whereas the A. vulgaris aerial part extracts represented a distinct cluster. Similarly, in cluster 1, both A. vulgaris root extracts were found to be clearly different as compared to the other extracts. These findings suggest that A. vulgaris distinguished itself from the other four Artemisia species investigated in the current work. Moreover, some compounds of A. vulgaris (i.e., caffeic acid-O-pentoside, esculetin-O-hexoside I, coumaroylquinic acid, coumaric acid-Opentoside, luteolin-O-deoxyheoside-O-hexoside and coumaroylcaffeoylquinic acid) could be used as potential markers for this species, due to their abundance in the aerial parts of A. vulgaris. hand and the aerial parts on the other (Figure 4). Overall, several compounds were more abundant in the extracts from the aerial parts than in the extracts from the roots.

Conclusions
In this work, five Artemisia species collected from the spontaneous flora of northeastern Romania, namely, A. absinthium, A. annua, A. austriaca, A. pontica and A. vulgaris, were comprehensively investigated with respect to their phytochemical profiles and multi-biological potential (antioxidant, enzyme inhibitory and anti-mycobacterial). The LC-HRMS/MS-based metabolite profiling allowed the annotation of 73 different compounds, of which 15 were phenolic acids (i.e., chlorogenic, neochlorogenic, dicaffeoylquinic, feruloylquinic, coumaroylquinic acids), 26 were flavonoids (i.e., as poly-hydroxylated/poly-methoxylated flavones) and 14 were sesquiterpenes (i.e., artemisinin, pseudosantonin, arteannuin B). CIM analysis of the phytochemical profile revealed three main clusters, the first comprising A. annua together with A. absinthium, the second A. pontica together with A. austriaca and the third A. vulgaris. The antioxidant activity analysis of the five species revealed the superior antioxidant activity of the aerial part extracts as compared to the root extracts, as well as the better antioxidant activity of the methanol extracts as compared to the chloroform extracts. Furthermore, PCA and HCA allowed us to differentiate the samples into three main clusters with respect to antioxidant and enzyme inhibitory potential, with one cluster (cluster 1-the methanol

Conclusions
In this work, five Artemisia species collected from the spontaneous flora of northeastern Romania, namely, A. absinthium, A. annua, A. austriaca, A. pontica and A. vulgaris, were comprehensively investigated with respect to their phytochemical profiles and multi-biological potential (antioxidant, enzyme inhibitory and anti-mycobacterial). The LC-HRMS/MSbased metabolite profiling allowed the annotation of 73 different compounds, of which 15 were phenolic acids (i.e., chlorogenic, neochlorogenic, dicaffeoylquinic, feruloylquinic, coumaroylquinic acids), 26 were flavonoids (i.e., as poly-hydroxylated/poly-methoxylated flavones) and 14 were sesquiterpenes (i.e., artemisinin, pseudosantonin, arteannuin B). CIM analysis of the phytochemical profile revealed three main clusters, the first comprising A. annua together with A. absinthium, the second A. pontica together with A. austriaca and the third A. vulgaris. The antioxidant activity analysis of the five species revealed the superior antioxidant activity of the aerial part extracts as compared to the root extracts, as well as the better antioxidant activity of the methanol extracts as compared to the chloroform extracts. Furthermore, PCA and HCA allowed us to differentiate the samples into three main clusters with respect to antioxidant and enzyme inhibitory potential, with one cluster (cluster 1-the methanol root extracts of A. annua and A. pontica and the aerial part extracts of A. vulgaris) being characterized by the highest radical scavenging (ABTS and DPPH) and reducing power (FRAP and CUPRAC) activities. In addition, the chloroform extract of A. austriaca aerial parts showed the highest antibacterial effects against M. tuberculosum H37Ra (MIC = 64 mg/L), while other extracts displayed MIC values of 128-256 mg/L. Aside from the chemotaxonomic importance, the current study makes significant contributions to knowledge of the chemical and versatile biological profile of the investigated Artemisia ssp. collected from Romanian flora. Overall, our research could open prospects for the large-scale exploitation of Artemisia species (both roots and aerial parts) as rich sources of bioactive metabolites endowed with interesting antioxidant, enzyme inhibitory and anti-mycobacterial properties.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/antiox11051017/s1, Table S1. Full spectro-chromatographic data of compounds identified in the Artemisia root and aerial parts extracts by LC-HRMS/MS.