Identification of Marker Compounds and In Vitro Toxicity Evaluation of Two Portuguese Asphodelus Leaf Extracts

The leaves of Asphodelus bento-rainhae subsp. bento-rainhae, an endemic Portuguese species, and Asphodelus macrocarpus subsp. macrocarpus have been used as food, and traditionally as medicine, for treating ulcers, urinary tract, and inflammatory disorders. The present study aims to establish the phytochemical profile of the main secondary metabolites, together with the antimicrobial, antioxidant and toxicity assessments of both Asphodelus leaf 70% ethanol extracts. Phytochemical screenings were conducted by the TLC and LC-UV/DAD-ESI/MS chromatographic technique, and quantification of the leading chemical classes was performed by spectrophotometric methods. Liquid-liquid partitions of crude extracts were obtained using ethyl ether, ethyl acetate, and water. For in vitro evaluations of antimicrobial activity, the broth microdilution method, and for the antioxidant activity, the FRAP and DPPH methods were used. Genotoxicity and cytotoxicity were assessed by Ames and MTT tests, respectively. Twelve known compounds including neochlorogenic acid, chlorogenic acid, caffeic acid, isoorientin, p-coumaric acid, isovitexin, ferulic acid, luteolin, aloe-emodin, diosmetin, chrysophanol, and β-sitosterol were identified as the main marker compounds, and terpenoids and condensed tannins were found to be the major class of secondary metabolites of both medicinal plants. The ethyl ether fractions demonstrated the highest antibacterial activity against all the Gram-positive microorganisms, (MIC value of 62 to 1000 µg/mL), with aloe-emodin as one of the main marker compounds highly active against Staphylococcus epidermidis (MIC value of 0.8 to 1.6 µg/mL). Ethyl acetate fractions exhibited the highest antioxidant activity (IC50 of 800 to 1200 µg/mL, respectively). No cytotoxicity (up to 1000 µg/mL) or genotoxicity/mutagenicity (up to 5 mg/plate, with/without metabolic activation) were detected. The obtained results contribute to the knowledge of the value and safety of the studied species as herbal medicines.


Introduction
Medicinal plants have been used as potential functional foods or resources to prevent various diseases worldwide in different traditional medicine systems. Medicinal plants and their respective phytochemicals, mainly secondary metabolites, are used not only to combat specific nutrient deficiencies, but to sustain secure food and primary healthcare medicines [1].
The species Asphodelus L. (Asphodelaceae) is consumed in large quantities in the cuisines (e.g., soups, pastries, etc.) of several countries and cultures. The leaves of Asphodelus aestivus Brot., for instance, are commonly consumed as a cooked vegetable dish in only to combat specific nutrient deficiencies, but to sustain secure food and primary healthcare medicines [1].
Asphodelus bento-rainhae subsp. bento-rainhae P. Silva is an endemic species from Serra da Gardunha and is considered as "vulnerable" on the Red List of Threatened Species of the International Union for the Conservation of Nature (IUCN), and co-exists with Asphodelus macrocarpus subsp. macrocarpus Parl. in the same geographical area. They are known by the common Portuguese name "abrotea" (Ancient Greek: Ἀβρότονον), and their leaf is used as fertilizer and fodder in Portugal [25]. To date, no data related to the phytochemical characterization, pre-clinical safety, and biological potential of Asphodelus bento-rainhae and Asphodelus macrocarpus leaves have been found in the literature. Therefore, the present study was conducted to identify the main chemical constituents, antimicrobial and antioxidant activities of leaf extracts of these species along with their in vitro toxicity assessments, using samples collected at different times of the year to determine the most appropriate period for the collection of material and to contribute to the knowledge of safety and their value as herbal medicinal products.

Phytochemical Analysis
Thin-layer chromatography (TLC), followed by high-performance liquid chromatography (HPLC) coupled to a photodiode detector (UV/DAD), and electrospray ionization spectrometry (ESI/MS) techniques were applied for the rapid and reliable detection of several samples. The obtained chromatographic profiles of Asphodelus bento-rainhae and Asphodelus macrocarpus leaf extracts (AbLa, and AmLa, respectively) and their subsequent liquid-liquid partition with increasing polarity solvents, namely ethyl ether (AbLa-1, AmLa-1), ethyl acetate (AbLa-2, AmLa-2) and water (AbLa-3, AmLa-3), showed qualitative similarity in their chemical composition, characterized by the presence of terpenoids, phenolic acids, flavonoids, and anthracene derivatives. Based on both TLC and HPLC spectral analysis, using the authentic standards (cochromatography) and comparison with literature data (Figure 1), twelve known compounds, namely, neochlorogenic acid (a), chlorogenic acid (b), caffeic acid (c), isoorientin (d), p-coumaric acid (e), isovitexin (f), ferulic acid (g), luteolin (h), aloe-emodin (i), dios-βρóτoνoν), and their leaf is used as fertilizer and fodder in Portugal [25]. To date, no data related to the phytochemical characterization, pre-clinical safety, and biological potential of Asphodelus bento-rainhae and Asphodelus macrocarpus leaves have been found in the literature. Therefore, the present study was conducted to identify the main chemical constituents, antimicrobial and antioxidant activities of leaf extracts of these species along with their in vitro toxicity assessments, using samples collected at different times of the year to determine the most appropriate period for the collection of material and to contribute to the knowledge of safety and their value as herbal medicinal products.
Molecules 2023, 28, x FOR PEER REVIEW 3 of 20 metin (j), chrysophanol (k), and β-sitosterol (l) were identified as major marker com pounds of both species (Table 1, Figure 2).  Previously reported phytochemical studies of Asphodelus spp. revealed the presence of chlorogenic acid in the leaf and aerial part extracts of Asphodelus aestivus Brot. [8] and Asphodelus ramosus L. [26], while caffeic acid was only reported from the flower extract of A. ramosus [27].
Previously reported phytochemical studies of Asphodelus spp. revealed the presence of chlorogenic acid in the leaf and aerial part extracts of Asphodelus aestivus Brot. [8] and Asphodelus ramosus L. [26], while caffeic acid was only reported from the flower extract of A. ramosus [27].
Quantification results of the main chemical classes of secondary metabolites, namely total phenolics (TPC), total flavonoids (TFC), total anthraquinones (TAC), total condensed and hydrolysable tannins (TCTC and THTC, respectively), together with total terpenoids, (TTC) are presented in Table 2. Concerning the analysis between the different collection seasons for the same species, the results showed that the total content of TCTC and TFC in A. bento-rainhae (p-values: 0.034, 0.01, respectively) and THTC in A. macrocarpus leaf extracts were significantly higher in the first collection season (p-value: 0.01).
The analysis of the results between different species of the same collection season showed that TTC content in the first collection season and TFC content in the second collection season were significantly higher in A. macrocarpus when compared to those of A. bento-rainhae (p-values: 0.0021 and 0.01, respectively). However, TAC, TCTC, and TFC contents in the first collection season (p-value: 0.002, 0.007, and 0.006, respectively) and THTC content in the second collection season (p-value: 0.028) were significantly higher in A. bento-rainhae when compared to those of A. macrocarpus.
The obtained data showed the TCTC (180.96 ± 10.98 and 142.98 ± 6.71 mg CAE/g DW) and TTC (111.72 ± 22.77 and 165.47 ± 26.54 mg OAE/g DW) contents with the highest and TAC (1.07 ± 0.11 and 0.55 ± 0.07 mg RhE/g DW) with the lowest content in comparison to the other quantified chemical classes in both A. bento-rainhae and A. macrocarpus leaf extracts.
Previously reported Asphodelus spp. leaf extracts quantified values of TPC, TFC, and TCTC indicated the important role of solvent selection for the extraction procedure. In fact, A. microcarpus ethanol extract showed a higher amount of TPC and TFC (54.44 ± 13.6 mg GAE/g of DW and 31.13 ± 1.96 mg QUE/g of DW, respectively) in comparison to the aqueous and methanol extracts [21]. However, in A. ramosus, the aqueous extract exhibited a higher amount of TPC (33.51 ± 0.33 mg GAE/g of DW) when compared to the methanol, methanol/water (50%), and ethyl acetate extracts [32]. A. aestivus acetone extract also showed an elevated amount of TFC (17.74 ± 0.46 mg CAE/g of DW) in comparison to the aqueous, ethanol and methanol extracts [2,19]. Contrary to the data mentioned above and that obtained by us, significantly higher amounts of TPC (183.7 ± 3.5, 128.5 ± 2.1 and 109.7 ± 1.5 mg GAE/g of DW) and lower amounts of TCTC (59.8 ± 0.6, 49.2 ± 0.5 and 41.4 ± 0.3 mg CAE/g of DW) were reported from A. tenuifolius methanol, ethanol, and petroleum ether extracts [33]. It was also observed that both TPC and TFC contents have increased with the increase of the extraction temperature in the experiments done with A. ramosus [32].

Determination of In Vitro Antioxidant Potential
In this study, the antioxidant activity was evaluated by two complementary methods, DPPH assay to determine the 50% inhibition of free radical scavenging activity, and FRAP, which evaluates the reducing potency of the antioxidants to react to the ferric tripyridyltriazine (Fe 3+ -TPTZ) complex.
Concerning the results shown in Table 3, overall, A. bento-rainhae exhibited stronger antioxidant activity when compared to A. macrocarpus extracts. Among all the tested extracts, ethyl acetate fractions (AbLa-2, AmLa-2) showed the highest antioxidant activity when compared to all the other fractions (IC 50 : 800 µg/mL and IC 50 : 1200 µg/mL, respectively). When comparing FRAP and DPPH, the obtained an r value of −0.975, showing a strong correlation between them, validating the results of both techniques, although the data of the FRAP test correlate better with the quantifications data. The classes of compounds that correlate better with the antioxidant power of the extracts are the flavonoids (TFC, r value of 0.943) and phenolic compounds (TPC, r value of 0.949), in which a higher content of these compounds is related to higher antioxidant power. In accordance with these results, phytochemical screenings of the crude extracts and their L-L partitions revealed the presence of homoorientin and chlorogenic acid as the main marker compounds of most active fractions (AbLa-2, AmLa-2). There is no report on the antioxidant activity of our studied Asphodelus species; however, the previously reported results of DPPH analyses of the other Asphodelus spp. showed that A. microcarpus leaf ethanol and methanol extracts exhibited the highest antioxidant activity (IC 50 : 55.9 µg/mL and IC 50 : 98 µg/mL, respectively) [21,23]. On the contrary, A. aestivus leaf methanol extracts noticeably showed a higher antioxidant activity when compared to ethanol extract (IC 50 : 160 µg/mL and IC 50 : 9540 µg/mL, respectively) [2,19]. A. tenuifolius leaf methanol extract exhibited the lowest IC 50 (18370 µg/mL) levels among the others, including our studied species [22].

Assessment of the Antibacterial Potential
The in vitro quantitative method of susceptibility testing (determination of MIC values) was used for the evaluation of the antimicrobial potential against both selected Grampositive and Gram-negative resistant pathogens in this study.
Concerning the obtained results, leaf crude extracts (AbLa, AmLa), and their subsequent ethyl acetate (AbLa-2, AmLa-2) and aqueous (AbLa-3, AmLa-3) L-L partition fractions did not exhibit antimicrobial activity against both Gram-positive and Gramnegative microorganism pathogens at any of the concentrations tested (MIC > 2000 µg/mL). However, as shown in Table 4, only diethyl ether fractions (AbLa-1, AmLa-1) demonstrated considerable antibacterial activity against all the Gram-positive microorganisms, with MIC values ranging from 62 to 1000 µg/mL. In general, A. bento-rainhae exhibited higher activity when compared to A. macrocarpus, and no activity in the tested range of concentrations (MIC > 2000 µg/mL) was found against Gram-negative microorganisms (Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Acinetobacter baumannii).
Aloe-emodin (compound i, Table 1), identified as one of the main marker compounds of the diethyl ether fraction of both plant extracts, was also tested against the pathogen panel under the study. This compound was found to be highly active against all the Gram-positive strains, particularly against all Staphylococcus epidermidis strains with a MIC between 0.8 to 1.6 µg/mL. In accordance with our results, aloe-emodin was previously reported as a potential antimicrobial that was active against several Gram-positive bacteria [34]; however, in a recent study, aloe-emodin with MIC values of 4 to 32 µg/mL exhibited deformities in the morphology of S. epidermidis cells and the destruction of the selective permeability of the cell membranes [35]. Results of studies involving the determination of the antimicrobial activity of other Asphodelus spp. against a similar pathogen panel revealed their lower antimicrobial potential. For instance, a leaf ethanol extract of A. aestivus exhibited a MIC of 42,000 µg/mL against S. aureus, and of 60,000 µg/mL against Klebsiella pneumoniae [36]. The A. fistulosus leaf ethanolic and aqueous extracts showed activity against S. aureus (MIC 2200 µg/mL and 7600 µg/mL, respectively) [37]. A methanol extract of A. luteus aerial part showed an MIC between 1250 to 2500 µg/mL against methicillin-resistant S. aureus (MRSA) [17]. A methylene-chloride extract of the aerial part of A. tenuifolius was found to be more active against S. aureus (MIC = 1600 µg/mL), Enterococcus faecalis (MIC = 1000 µg/mL), and E. coli (MIC = 1800 µg/mL) in comparison to the n-butanol and ethyl acetate extracts of the same species [9]. Recently, an A. tenuifolius whole plant chloroform extract was shown to be active against S. epidermidis (MIC = 580 µg/mL) [38]. A. microcarpus leaf extracts also showed antimicrobial activity against several Gram-positive strains, with MIC values of 78 to 5000 µg/mL [7,15,17,39,40]. A. bento-rainhae and A. macrocarpus leaf extracts seem to be more active against the tested Gram-positive strains in comparison to the other tested Asphodelus spp. extracts. The antibacterial activity of A. fistulosus leaf aqueous extract against E. coli (MIC = 62 µg/mL) and of A. tenuifolius aerial part methylene-chloride extract against the same microorganism (E. coli, MIC = 1800 µg/mL) and also against P. aeruginosa (MIC = 150 µg/mL) are examples of the few studies relating the antimicrobial activity of Asphodelus spp. to Gram-negative strains.
Overall, the observed antimicrobial activity of both A. bento-rainhae and A. macrocarpus leaf crude extracts were similar to those obtained and reported form the other Asphodelus spp. tested against a similar panel of pathogens. However, the fractionation of crude extracts enabled the detection of significant antimicrobial activity in the diethyl ether L-L partition fractions, quantitatively the richest in 1,8-dihydroxyanthracene derivatives, a known chemical class of secondary metabolites with antimicrobial activity [34].

Pre-Clinical Safety Assessment
Following the guidelines of the genotoxicity by the Ames test, which is commonly used as an initial screen of genotoxicity, for a substance to be considered genotoxic in the test, the number of revertant colonies on the plates containing the test compounds/substance must be more than twice the number of colonies produced on the solvent control plates (i.e., a ratio above 2.0). In addition, a positive dose-response should be evident for the various concentrations of the tested mutagen [41,42]. Since the crude extracts obtained from the first collection season (AbLa, AmLa) exhibited higher contents of the main classes of secondary metabolites, they were subsequently selected for further safety examination.
The obtained results of the Ames test for both AbLa and AmLa extracts are presented in Table 5. Neither extract induced an increase in the number of revertant colonies in any of the tested strains at any tested concentration, with (500, 1250, 2500, and 5000 µg/plate) and without (250, 625, 1250, 2500, 3750, and 5000 µg/plate) metabolic activation, when compared to the negative control. Moreover, cytotoxicity did not occur since there was neither a decrease in the number of spontaneous revertants nor a decrease on the background lawn of the plates at any of the concentrations tested. Therefore, under the conditions of this study, neither extract of the two species showed mutagenic activity.
Our cell viability assay ( Figure 3) concurrently indicated that none of the AbLa and AmLa extracts reduced HepG2 viability. The AbLa extract (50-500 µg/mL) enhanced HepG2 viability/proliferation up to~30% when compared to the 0 µg/mL concentration, whereas the same was observed for AmLa, i.e., it promoted HepG2 viability/proliferation by up to 40%, especially at higher concentrations (250-1000 µg/mL; p < 0.001 and p < 0.0001). Therefore, under the conditions of this study, the extracts of both species did not show mutagenic activity and in vitro cytotoxicity of HepG2, which is crucial to ensure their safety [42][43][44]. Table 5. Mutagenicity of A. bento-rainhae and A. macrocarpus leaf crude extracts in the bacterial reverse mutation test (Ames Test).

Plant Materials
The leaves of A. bento-rainhae (AbL) and A. macrocarpus (AmL) were collected from Serra da Gardunha, Portugal, first at the early flowering stage (AbLa, AmLa) in Spring, and then for the second time (AbLb, AmLb), during the Summer of 2019. All samples were dried in a well-ventilated dark space at room temperature. Corresponding voucher specimens were deposited in the Laboratory of Pharmacognosy, Department of Pharmacy, Pharmacology and Health Technologies, Faculty of Pharmacy, Universidade de Lisboa (Voucher specimens' number: OSilva_201901-A. bento rainhae and OSilva_201902-A. macrocarpus).

Preparation of Extract
Powder of the dried samples was obtained by grinding, and extraction was performed using the maceration method (with a mixture of ethanol/water 70:30) under agitation and filtration (3×, 24 h each). Hydroethanolic extracts were evaporated under reduced pressure at a temperature of less than 40 • C using a rotary evaporator and subsequently freeze-dried. The selected extracts (AbLa, AmLa) were then submitted to liquid-liquid partitioning (L-L), generating the diethyl ether (AbLa-1, AmLa-1), ethyl acetate (AbLa-2, AmLa-2), and aqueous (AbLa-3, AmLa-3) fractions.

Chromatographic Conditions
Silica gel 60 F 254 and 60 RP-18 F 254 pre-coated plates (Merck ® , Darmstadt, Germany) were used for TLC screenings. Different spray reagents, including anisaldehydesulfuric acid for the detection of terpenoids, natural product polyethylene glycol reagent (NP/PEG = NEU) for the detection of phenolics, and potassium hydroxide (KOH) 5% ethanolic solution for the detection of anthracene derivatives [45] were used.
A HPLC-UV/DAD analysis was performed using a Waters Alliance 2690 Separations Module (Waters Corporation, Milford, MA, USA) coupled with a Waters 996 photodiode array detector (UV/DAD) (Waters Corporation, MA, USA). An Atlantis T3 column, RP-18 end-capped (5 µm, 150 × 4.6 mm), connected to a pre-column with the same stationary phase was used. The injection volume was 25 µL with a flow rate of 1 mL/min. A mixture of water + 0.1% formic acid (solvent A) and acetonitrile (solvent B) was used as the mobile phase, and gradients (95% A and 5% B), 20 min (71% A and 29% B), 30 min (67% A and 33% B), 35 min (64% A and 36% B), 45 min (50% A and 50% B), 65 min (100% B) and 75 min (95% A and 5% B) were applied. Crude extracts (20 mg/mL) were solubilized in water and standard solutions were prepared in acetonitrile (1 mg/mL) and filtered through a polytetrafluoroethylene syringe filter (0.2 µm). Data were collected and analyzed using a Waters Millennium ® 32 Chromatography Manager (Waters Corporation, Milford, MA, USA). The chromatogram was monitored and registered on Maxplot wavelength (240-650 nm).
An HPLC-MS/ESI analysis was carried out using an HPLC (Waters Alliance 2695), with an autosampler and photodiode array detector (Waters PDA 2996) in tandem with a triple quadrupole mass spectrometer (Micromass ® Quatro Micro TM API, Waters ® , Drinagh, Ireland) using an electrospray ionization source (ESI) operating in negative mode. A LiChrospher 100 RP-18 (5 µm) 250 × 4 mm column with respective pre-column (Merck, Darmstadt, Germany) was used. A mixture of water + 0.1% formic acid (solvent A) and acetonitrile (solvent B) was used as the mobile phase. Data were acquired and analyzed using MassLynx™ V4.1 software (Waters ® , Drinagh, Ireland).
Peaks assignment and the identification of compounds were based on a co-chromatography technique with the comparison of retention times, UV-DAD, and mass spectral data with those of standards and published data.
All of the above-mentioned colorimetric techniques were assessed in triplicate for method validation, and a UV-Vis spectrophotometer (Hitachi, U-2000, Tokyo, Japan) was used. Values were obtained using standard equations (where X was the concentration of standard equivalents expressed as milligrams per gram of dried extract and Y was the measured absorbance). All of the obtained data were treated statistically by a one-way analysis of variance (ANOVA) with the Asphodelus species as the source of variance. Once both of the Asphodelus species were collected in two different seasons, the obtained data were also analyzed by ANOVA, with the season as the source of variance. The significant value was set for a p-value < 0.05.

In Vitro Antimicrobial Activity
The antibacterial assay was carried out by the broth microdilution method [51] in 96well tissue culture plates (VWR®, Radnor, PA, USA) to determine the activities by testing minimum inhibitory concentrations (MIC) of extracts against twelve reference (ATCC, LGC Standards S.L.U., Barcelona, Spain) and clinical strains (INSA clinical strains collection) of both Gram-positive (Staphylococcus aureus, S. epidermidis, S. saprophyticus, S. haemolyticus) ( Table 6) and Gram-negative (Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, Acinetobacter baumannii) bacteria representing the antimicrobial resistance status. Samples to be tested were initially prepared in water or DMSO 10% and were screened at the concentration of 2-2000 µg/mL for crude extracts or L-L partitions and 0.2-200 for pure compounds. Serial dilutions were performed in a Mueller-Hinton medium and were distributed (50 µL) in each of the microplate wells using a microplate liquid handler (Precision TM BioTek, Winooski, VT, USA). For the preparation of inoculum from a pure bacterial culture on agar, a suspension in Mueller-Hinton medium (10 8 CFU/mL) with a turbidity of 0.5 for Gram-negative and 0.25 for Gram-positive bacteria on the McFarland scale (Grant Bio™ DEN-1B, Cambridgeshire, UK) were prepared and stored at 4 • C until use. For MIC determination, the prepared suspensions were diluted at a ratio of 1:10, and from this dilution, 50 µL was added to all the wells. Two controls were included for each extract, fraction or compound, one plate in the absence of the extract solution and the other in the presence of the solvent (DMSO), to verify the absence of contamination and to check the validity of the inoculum. After incubation at 37 • C for 18 h, the plates were read in a lighted place, and the MIC was determined. All experiments were carried out in triplicate, as previously described, to obtain consistent values.

In Vitro Antioxidant Activity
The antioxidant potential was determined by two methods, initially started by a modified free radical scavenging activity (DPPH method) [52], followed by the ferric reducing antioxidant power test (FRAP assay). DPPH solution (3.9 mL, 6 × 10 −5 M in methanol) was mixed with 100 µL of diluted extracts or standard (ascorbic acid). After 30 min of incubation at room temperature, the absorbance of samples and standard solutions was measured at 517 nm. The percentage of DPPH free radical scavenging activity was calculated using the following formula: % scavenging = [absorbance of control−absorbance of test sample/absorbance of control] × 100. Results were expressed as mean ± standard deviation and presented in inhibitory concentration (IC 50 value), representing the sample concentration required to scavenge 50% of the DPPH free radicals.
For the Frap assay [53], 100 µL of plant extracts (1000-5000 µg) were mixed with 3 mL of working FRAP reagent (300 mM acetate buffer pH 3.6, 10 mM TPTZ in 40 mM HCl and 20 mM FeCl 3 . 6H 2 O in the ratio of 10:1:1 at the time of use); thereafter, samples were placed in the water bath at 37 • C. The reduction of ferric tripyridyl triazine (Fe III TPTZ) complex to ferrous form (which has an intense blue color) can be monitored by measuring the change in absorption at 593 nm, measured after 4 min. Ascorbic acid concentrations (25-175 µg/mL) were used to obtain a standard curve with an equation of Y = 0.616X−1.1702, R 2 = 0.9989. The FRAP reagent was used as a blank, and results were expressed as mmol ascorbic acid/g dry extract. Values were obtained in three sets of experiments and assessed in triplicate for method validation.
To ascertain if both methods were equally valid in measuring the antioxidant activity, they were correlated through the Pearson coefficient index (−1 < r < 1). A Pearson coefficient absolute value higher than 0.9 shows a strong correlation between the two methods. The Pearson index was also used to correlate the data of antioxidant activity with the quantification of the several chemical classes of compounds to ascertain their relationship with antioxidant power. Once both Asphodelus species were collected in two different seasons, the obtained data were also analyzed by ANOVA, with the season as the source of variance. The significance value was set for a p-value < 0.05.

In Vitro Genotoxicity/Mutagenicity Evaluation by Ames Test
A bacterial reverse mutation test (Ames test), commonly employed as an initial screening of the genotoxicity potential of herbal substances/preparations, was used to detect relevant genetic changes and genotoxic carcinogens [54]. The assessment of mutagenicity was performed according to the OECD No. 471 [55], the ICH S2 (R1) [56] guidelines, and following the published protocols [44], using five Salmonella enterica serovar Typhimurium tester strains (TA98, TA100, TA102, TA1535, and TA1537) in a direct plate incorporation method with and without metabolic activation. TA100, TA98, TA102 and TA1535 were kindly provided by the Genetic Department of the Nova Medical School of the Universidade NOVA de Lisboa (Portugal), having received them from Professor B.N. Ames (Berkeley, CA, USA). TA1537 was from ATCC, NUMBER: 29630™, LOT: 7405375. The strains were inoculated in nutrient broth medium and incubated for 12-16 h, at 37 • C in the dark, shaking at 210 rpm in an orbital incubator, and kept at 4 • C until use.
The extracts (25 mg/mL) were dissolved in DMSO (up to 30%), which also served as the negative control. An amount of 200 µL of extract dilutions were mixed with 500 µL sodium phosphate buffer (0.1 M, pH 7.4) (assay without metabolic activation) or S9 mix (assay with metabolic activation), 100 µL of the bacterial culture, and 2 mL of melted top-agar, supplemented with 0.05 mM biotin and histidine, at 45 • C. This mixture was then vortexed and plated on Petri dishes with Vogel-Bonner agar medium and supplemented with 2% glucose. After a 48-h incubation at 37 • C, manual counting of His+ revertant colonies for each concentration was performed. All assays were performed in triplicate. The results were expressed as the mean number of revertant colonies with the standard deviation (mean ± SD). The positive controls were sodium azide (SA, 1.5 µg/plate for TA100 and TA1535), 2-nitrofluorene (2-NF, 5 µg/plate for TA98), 9-aminoacridine (9-AA, 100 µg/plate for TA1537), and tert-butyl hydroperoxide (tBHP, 50 µg/plate for TA102) in the assay without metabolic activation, and 2-aminoathracene (2-AA, 2 µg/plate for TA98 and 10 µg/plate for TA102, TA1535 and TA1537) and benzo(a)pyrene (BaP, 5 µg/plate for TA100) in the assay with metabolic activation.

In Vitro Cytotoxicity Evaluation by MTT Assay
Cytotoxicity was evaluated by the methylthiazolyldiphenyl-tetrazolium bromide (MTT) reduction assay [57] on a human liver cell line HepG2 (ATCC Cat. No. HB-8065, Middlesex, UK). HepG2 were seeded in 96-well plates at a density of 8.5 × 10 4 cells/cm 2 in α-MEM (Sigma-Aldrich ® , St. Louis, MO, USA) with 1 mM sodium pyruvate (PAN Biotech, Aidenbach, Germany) and 1% non-essential amino acids (NEAA, PAN Biotech, Aidenbach, Germany) supplemented with 10% fetal bovine serum (FBS, Gibco ® -Thermo Fisher Scientific TM (Waltham, MA, USA), in a humidified chamber at 37 • C in a 5% CO 2 atmosphere. After 48-h incubation, the cell culture medium was replaced by fresh medium with AbLa and AmLa extracts (9:1) at final concentrations of 50, 125, 250, 500, and 1000 µg/mL. Cells were also incubated with a complete cell culture medium, DMSO 1% and DMSO 20% in α-MEM as a positive, solvent, and negative control, respectively. After 48 h, the cells were carefully washed with 100 µL PBS, and 200 µL 0.5 mg/mL MTT (Sigma-Aldrich ® ) in a cell culture medium was added. HepG2 were incubated for 3 h in a humidified chamber at 37 • C in a 5% CO 2 atmosphere. The purple crystals were solubilized with 200 µL DMSO and measured at 570 nm using a microplate spectrophotometer (SPECTROstar Omega; BMG LabTech, Ortengerg, Germany). The results were expressed as a percentage relative to the solvent control. Four wells were used for each sample, and at least two independent experiments were performed.

Conclusions
The weak antimicrobial activity verified with our crude leaf extracts of Asphodelus bento-rainhae and Asphodelus macrocarpus is consistent with the results obtained when testing other Asphodelus spp. against a similar panel of pathogens [4]. However, fractionation of these extracts enabled the detection of significant antimicrobial activity in the diethyl ether L-L partition fractions, quantitatively the richest in 1,8-dihydroxyanthracene derivatives, a known chemical class of secondary metabolites with antimicrobial activity [34]. Furthermore, the well-known antibacterial agent aloe-emodin was identified as the main compound responsible for this activity. Although the in vitro cytotoxicity and mutagenicity of this compound has been reported by others, no cytotoxicity or mutagenic activity was observed in the corresponding extracts and fractions that we tested.
On the other hand, the ethyl acetate L-L partition fractions are quantitatively the richest in phenolic acids and flavonoid derivatives, and showed the highest antioxidant activity, confirming the major role of the different classes of the identified phenolic compounds in the activity of Asphodelus bento-rainhae and Asphodelus macrocarpus leaves as medicinal plants. Moreover, the negative results of the Ames and MTT tests indicate that the hydroethanolic leaf extracts of both species are safe in terms of toxicity, and these data together with the phytochemical profiles will provide appropriate information for inclusion in the future quality monograph of these medicinal plants.