LC-ESI/MS-Phytochemical Profiling with Antioxidant, Antibacterial, Antifungal, Antiviral and In Silico Pharmacological Properties of Algerian Asphodelus tenuifolius (Cav.) Organic Extracts

Asphodelus tenuifolius Cav. (A. tenuifolius) is a medicinal plant with a long history of traditional use to treat ailments. In this study, total phenolic and flavonoid content evaluation using LC-ESI/MS analysis and various biological activities (antioxidant, antibacterial, antifungal, antiviral and cytotoxicity) of organic extracts from the aerial parts of A. tenuifolius were analyzed. ADME tools were used to predict the potential of the identified compounds from the most potent extract as specific drugs. As shown, LC-ESI/MS results of chloroformic extract allowed the tentative identification of 12 compounds. Chloroformic extract was rich in polyphenols and flavonoids and exhibited the highest antioxidant activity given by DPPH (IC50 = 25 µg/mL) as compared to the BHT standard (11.5 µg/mL) and β-carotene bleaching assays (IC50 = 95.692 µg/mL). Antibacterial activity results showed that chloroformic extract has a highest activity against Gram-positive and -negative bacteria, especially against Salmonella Typhimurium DT104 (IZ = 19.3 mm, MIC = 18.75 mg/mL, MBC = 37.5 mg/mL). The MBC/MIC ratio was evaluated to interpret the activity that was bacteriostatic rather than bactericidal. Conversely, weaker antifungal activity was registered, and no antiviral activity was observed for all extracts against Herpes Simplex Virus type 2 and Coxsakievirus B-3 viruses. Cytotoxic activity on VERO cell line results revealed that butanol extract was not toxic, with CC50 value of 1430 µg/mL, while chloroformic extract showed moderate cytotoxicity. Additionally, in silico studies performed proved promising pharmacokinetic and drug-likeness properties of the main compounds from the chloroformic extract. Taken together, this work highlights the potent bioactivity and acceptable drug-likeness of this plant, which supports its further preclinical development.


Introduction
The Liliaceae family is a rich source of natural products displaying a vast range of structural diversity. It includes approximately 289 genera and 4000 grown species, with many important medicinal, edible and ornamental plants [1]. Among these genera, eighteen species of Asphodelus have been widely cultivated throughout the Mediterranean area, North Africa and Southeast Asia. It reaches its maximum diversity in the Iberian Peninsula and Northwest Africa [2]. Asphodelus species were found to contain several secondary metabolites such as flavonoids [3,4], anthraquinones [5,6], phenolic acids [7,8], triterpenes [9,10], fatty acids [11] and naphthalene derivatives [12]. Different ethnomedical uses were ascribed to Asphodelus species, including Asphodelus tenuifolius Cav., known as onion weed [13], native to the Mediterranean region, which is used traditionally not only as a vegetable but also for colds, hemorrhoids, rheumatic pain, diuretic agent and wound healing [14]. As recently overviewed [15], organic extracts from the whole plant, fruits, roots, seeds and leaves were studied for samples collected from different harvest regions in the world [13,[16][17][18][19][20][21] and also specifically from Algerian regions [14,22]. The diversity of results is mainly due to the harsh climatic conditions, which stimulate the biosynthesis of secondary metabolites, different plant growing conditions, collecting regions and different extraction procedures applied [23,24].
In the present study, we set out to provide more information on the A. tenuifolius phytochemical composition by LC-ESI/MS analysis. In addition, evaluation of total phenolic and flavonoids contents of chloroformic, ethyl acetate and butanol extracts was performed, investigating their antioxidant, antibacterial, antifungal, cytotoxic and antiviral activities. To the best of our knowledge, this work is the first report on chemical composition study based on RP-HPLC-ESI/MS measurements of A. tenuifolius plant extracts. Additionally, ADME/pharmacokinetics and drug-likeness properties of the identified metabolites were carried out through in silico SwissADME online program.

Plant Material and Extraction Procedure
The whole plant of A. tenuifolius Cav. was collected in May 2012 from southwest Algeria and identified by an expert botanist M. Mohamed Ben Abd-elhakem (Ex-Director of the National Agency of Preservation of Natural Resources, Bechar, Algeria). An authenticated voucher specimen with the identification number (AS10TEN) was deposited at the herbarium of the VARENBIOMOL research unit, University Mentouri Constantine. A total of 1250 g of the whole plant was dried, powdered and extracted with 80% ethanol aqueous solution at room temperature (each extraction lasting about 48 h). After filtration and concentration under vacuum at about 40 • C, the combined concentrated ethanol extract was suspended in distilled water. Each resulting solution was extracted successively using chloroform, ethyl acetate and butanol. The organic phases were filtered and concentrated in a vacuum at 38 • C to obtain dry extracts: chloroform (CHE, 3.60 g), ethyl acetate (EAE, 4.05 g) and butanol (BE, 6.40 g).

Total Phenolic Content (TPC)
Total phenolic content (TPC) was determined using Folin-Ciocalteu method according to Dewanto et al. [25] using Gallic acid as a standard. The reference range was prepared with Gallic acid in different concentrations from 50 to 500 µg/mL. Total phenolic contents are expressed as milligram Gallic acid equivalents per gram of dry residue (mg GAE/g DR). All samples were analyzed in three replicates.

Total Flavonoids Content (TFC)
Total flavonoids content (TFC) was measured using the colorimetric method introduced by Dewanto et al. [25] with few modifications. A calibration curve was constructed using catechin standard solution in different concentrations from 50 to 500 µg/mL. Total flavonoids contents were calculated as catechin equivalents per gram of plant dry residue (mg CE/g DR). Measurements were performed at least in triplicate.

LC-ESI/MS Analysis
LC-ESI/MS profiles were performed using a Hewlett-Packard (Palo Alto, CA, USA) Model 1100 Series liquid chromatography coupled to a Photo Diode Array detector (Agilent, Palo Alto, CA, USA) 1100 Series, and to an Esquire LC-ion trap mass spectrometer (Bruker Daltonics, Billerica, MA, USA) equipped with an electrospray ionization (ESI) interface. Separation was achieved on a Phenomenex Luna C18 analytical column (250 mm × 4.6 mm i.d.; 5 µm particle diameter, end-capped). The mobile phase consisted of water (eluent A) and acetonitrile (eluent B) at a flow rate of 1 mL/min. The injection volume was 2 µL. Gradient elution was carried out using the following timetable: 20% A and 80% B in 30 min, then 100% B in 40 min. The resulting total run time was 70 min. The Photo Diode Array detector was set at a range of 200-700 nm for all peaks. The chromatogram was recorded at 215 nm, 254 nm, 300 nm and 330 nm.
(2) Bruker-Avance 400 spectrometer by using a 5 mm BBI probe. NMR spectra were acquired in CDCl 3 , DMSO-d 6 , CD 3 COCD 3 and CD 3 OD in the phase-sensitive mode with the transmitter set at the solvent resonance and time-proportional phase increment (TPPI) used to achieve frequency discrimination in the ω 1 dimension. The standard pulse sequence and phase cycling were used for HSQC and HMBC experiments.

Evaluation of Antioxidant Activity DPPH Radical-Scavenging Activity and β-Carotene Bleaching Assay
The effect of the various tested extracts on DPPH-degradation was estimated according to the method described by Espín et al. [26] while β-carotene linoleic acid bleaching assay was done according to Condelli et al. [27].
where A 0 is the absorbance of the blank, and A 1 is the absorbance of the sample. Results were expressed as percentage of β-carotene bleaching inhibition (AA%) and calculated as follows (Equation (2)): AA% = (A β-carotene after 180 min/A initial β-carotene) × 100; where A 0 and A 1 have the same meaning as in Equation (1). The results are expressed as IC 50 values (µg/mL).

Disk Diffusion Assay
The antibacterial and antifungal activities of A. tenuifolius organic extracts were evaluated by the agar disk diffusion method described by Rios and Recio [28] and Snoussi et al. [29]. Eight strains generally recognized as the most important pathogens affecting food dishes (Escherichia coli ATCC 35218, Vibrio parahaemolyticus ATCC 17802, Staphylococcus aureus ATCC 25923, Salmonella typhimurium DT 104, Staphylococcus epidermidis CIP 106510, Salmonella typhimurium ATCC 1408, Bacillus cereus ATCC 11778, Listeria monocytogenes ATCC 19115) were tested in the present study. On other hand, the antifungal activity effect was tested against four Candida strains (C. tropicalis 06-85, C. parapsilosis ATCC 22019, C. krusei ATCC 6258 and C. albicans ATCC 2019). The same technique was used to evaluate the antifungal activity. Ampicillin (10 mg/mL) and Amphotericin B (10 mg/mL) were used as a positive control [30]. The antibacterial activities were evaluated by measuring the diameter of the growth inhibition zone (IZ) around the discs using a flat rule. All tests were performed in triplicate, and the mean diameter of IZ was calculated. The results were expressed in terms of IZ of growth around each disc in millimeters, considered as low activity (IZ from 1 mm to 6 mm), moderate activity (7 mm to 10 mm), high activity (11 mm to 15 mm) and very high activity (16 mm to 20 mm) [31].

Cytotoxic and Antiviral Activities
The cytotoxic activity was evaluated according to the method described by Snoussi et al. [30] on VERO (African green monkey kidney) cells line. The 50% Cytotoxic Concentration (CC 50 ), defined as the concentration of the extract able to reduce of 50% the cell viability, was determined by regression analysis in comparison to negative control. The extracts that demonstrated activity at or below 100 µg/mL were categorized as having strong cytotoxic activity. Consequently, CC 50 values between 100 µg/mL and 500 µg/mL were categorized as having moderate cytotoxicity, the extracts that had CC 50 values between 500 µg/mL and 1000 µg/mL were considered to have weak cytotoxic activity [34] and the extracts had CC 50 values more than 1000 µg/mL were considered to be nontoxic [35]. Antiviral activity was also evaluated according to the method reported by Snoussi et al. [30] on two viruses, Herpes Simplex Virus type 2 (HSV-2) and Coxsakievirus B-3 (CVB-3).

In Silico ADME Profiles
The pharmacokinetics and drug-likeness properties of identified compounds from CHE of A. tenuifolius were estimated using ADME (absorption, distribution, metabolism and excretion) descriptors through SwissADME online server (http://www.swissadme.ch/ (accessed on May 2020)) by entering chemical structure followed by SMILES [36,37].

Statistical Analysis
The results were given as the average ± SD for three replicates. The IC 50 of DPPH, the CC 50 and the antiviral IC 50 were calculated by linear regression analysis. The β-carotene bleaching method values, the total secondary metabolite contents and the inhibition zone determination were performed using Microsoft Excel. The data were subjected to Duncan's multiple range tests. The statistical analyses were determined with the SPSS statistical software program (SPSS v.16), and p values < 0.05 were regarded as significant.

Phytochemical Contents
The chloroformic extract was rich in phenolic constituents (40.99 mg GAE/g DR) compared to the rest of the studied A. tenuifolius extracts (Table 1). This can be attributed to the higher solubility of constituents containing phenolic rings in this extract, whereas butanol extract was the one with the lowest phenolic concentration (10.54 mg GAE/g DR), suggesting their weaker solubility in this extract. In fact, Mahboub et al. [22] reported different results obtained from lyophilized samples of A. tenuifolius harvested from septentrional Algerian Sahara (101.82 µg GAE/g DR). In the same way, Munir et al. [38] and Al-Laith et al. [39] reported other data on A. tenuifolius from Pakistan and Bahrain, respectively, with total phenolic contents ranging within 53.40 to 76.23 mg GAE/g DR and 139.66 to 442.44 mg GAE/g DR, respectively, for various solvent extracts and harvested sites. Table 1. Total phenolic and flavonoid contents expressed as gallic acid and catechin equivalents, in mg/g dry sample, respectively, DPPH radical-scavenging activity and β-carotene bleaching capacity.

Extracts
Total Phenolic Content (mg GAE/g DR) Results of total flavonoid contents are shown in Table 1. Chloroformic extract displayed also the highest flavonoid contents (213.07 mg CE/g DR), followed by EAE (202.89 mg CE/g DR). Previously, Mahboub et al. [22] found a good level of flavonoid contents marked in lyophilized dried extracts with a value of 16.10 µg QE/g DR. Similarly to our results, Munir and colleagues [38] also found flavonoid contents in the range of 165.82 to 312.12 mg QE/g DR for all studied extracts. The presence of phenolics and flavonoids is very important to assess the antioxidant potency, especially due to their chemical structure affecting redox properties, which play a vital role in absorbing and neutralizing free radicals [40,41].
In summary, the lowest polar solvent (CHE) displayed the highest phenolic and flavonoid contents, and the higher polar solvent (BE) showed the lowest amount, indicating the richness of A. tenuifolius extract in low polar phenolic compounds.

HPLC-DAD-ESI/MS Analysis
A qualitative analysis of constituents present in chloroformic extract was performed by LC-ESI/MS. As shown in Figure 1, the full scan using negative electrospray ionization mode revealed the presence of several phenolic compounds, based on the over 17 major peaks detected within 70 min of elution. Table 2 summarizes retention time (Rt), m/z, molecular weight and formula of compounds identified or deduced based on data reported in the literature. All tentatively identified compounds are already known from A. tenuifolius plant and Asphodelus genus, while some minor metabolites are still unidentified.  The analysis of the chloroformic extract leads to the tentative identification (   [48]. Finally, (17) asphodelin (m/z 505.0) reported in many studies from A. acaulis, A. albus, A. fistulosus and A. microcarpus [12,43,45,49]. However, compounds corresponding to peaks 2, 4, 9, 11 and 16 are still unidentified.

Antioxidant Activities
Antioxidant activity of different Algerian A. tenuifolius extracts was evaluated using two different tests: DDPH radical scavenging activity and β-carotene bleaching. The results presented in Table 1 were expressed as IC 50 values and compared to the positive control BHT. As shown, all extracts displayed potent free radical scavenging activity on DPPH with IC 50 values ranging from 25 to 92 µg/mL, clearly less important than thatpositive control BHT (11.5 ± 0.01 µg/mL). CHE had the strongest DPPH radical scavenging activity (25 ± 4.36 µg/mL) followed by EAE (45 ± 2.88 µg/mL), and BE (92 ± 4.05 µg/mL). In our results, the extract rich in flavonoids and phenolics (CHE) was found to be the most significant scavenger of DPPH radical, which is supported by the good correlation between phenolic and flavonoid contents and DPPH outcomes. Polyphenolic compounds are usually the major antioxidants in plant extracts [50]. Based on LC-ESI/MS results, chloroformic extract showed an important richness, especially in apigenin-7-O-glucoside, apigenin, luteolin, anthraquinones and their derivatives formerly known for their antioxidant roles [51][52][53][54][55].
Previous studies on methanol, ethanol and petroleum ether extracts of Algerian A. tenuifolius displayed an important antioxidant activity, with IC 50 values ranging between 28.34 and 75.91 µg/mL [14]. Similarly, Al-Laith et al. [39] reported that, for extracts of the plant collected from two different harvested sites from Bahrain, IC 50 values of antioxidant and antiradical activity varied between 18.37 and 37.24 mg/mL. Kalim et al. [18], also reported DDPH radical scavenging results from Indian A. tenuifolius with an IC 50 = 2.00 µg/mL of 50% methanolic extract. This difference could also be due to the variable plant growing conditions, collection regions and extraction procedures [56,57]. Table 1 depicts the inhibition of β-carotene bleaching by A. tenuifolius extracts. It is possible to note that EAE had a significant activity, with an IC 50 value of 73.581 µg/mL, close to BHTreported a study on a whole Indian A. tenuifolius methanol extract against two Gram-positive bacteria (S. aureus and B. cereus), with IZ diameters of 9 and 13 mm, respectively, while acetone extract showed only activity against a Gram-negative bacteria K. pneumoniae (IZ = 17 mm). Additionally, Panghal et al. [20] outlined an antibacterial study of Indian A. tenuifolius fruits using six organic solvents for extraction, in which no activities were observed for petroleum ether and aqueous extracts against all tested bacterial strains, specifically against E. coli for the aqueous extract (IZ = 13.67 ± 0.5). The same study revealed that benzene extract exhibited very good susceptibility to K. pneumoniae, P. aeruginosa (IZ = 10.33 ± 0.5) and A. fumigatus (IZ = 10.66 ± 0.5). In the same way, all other extracts displayed antibacterial activity against overall tested bacterial strains. Moreover, a good antibacterial activity was observed by Dangi et al. and Menghani et al. [16,19] against the number of selected bacteria. From Indian A. tenuifolius, various extracts were found to be active against almost all the tested bacteria. Eddine et al. and Mahboub et al. [14,22] reported studies from Algerian A. tenuifolius, collected from the south-east and septentrional Sahara of Algeria, respectively. In detail, the results on the antibacterial effect of methanol, ethanol and petroleum ether extracts reported by Eddine et al. [14] showed a marked activity against S. aureus and P. putida (IZ = 11 mm), B. cereus, P. aerigunosa, A. tumefaciens and E. coli (IZ = 10 mm), and S. Arizona (IZ = 9 mm). Mahboub et al. [22] found a remarkable inhibition against E. coli and P. aeruginosa for the lyophilized sample, unlike our results on chloroformic, ethyl acetate and butanol extracts. Table 3 shows that antibacterial effect of the extracts was more important than the antifungal one, suggesting that yeast strains are more resistant to bacteria. These results are consistent with the ones previously reported, indicating that the inhibitory activity is pathogen-specific and depends on a number of factors, including the solvents, concentration of the crude drug, temperature, plant parts used for the extraction of secondary metabolites and rate of diffusion [58]. Both MICs and MBCs of A. tenuifolius organic extracts are summarized in Table 3. MICs ranged from 0.58 to 37.5 mg/mL, from 0.39 to 25 mg/mL, from 0.78 to 50 mg/mL for CHE, EAE and BE, respectively, for all bacterial and fungal strains tested. Regarding MBCs, high concentrations were needed to eliminate the growth of all tested bacterial and fungal strains, with values ranging from 2.34 to >100 mg/mL. Results indicated that CHE was the most effective compared to data obtained for EAE and BE. Our values of MICs and MBCs parameters are different from those obtained in numerous studies, specifically by Soliman et al. [59], in which MICs ranged from 25 to 50 µg/mL; Faidi et al. [17], found values from 0.15 to 4.1 mg/mL; Dangi et al. [16], with MICs results in range from 8 to 32 µg/mL; and Panghal et al. [20], reporting values in the range from 31 to 500 µg/mL. This behavior could be related mainly to the difference in the phenolic composition of each studied extract, plant origin, tested microorganisms and the size of the inoculum [60]. The inhibitory effect of A. tenuifolius extracts was evaluated against four yeasts. The results revealed weaker antifungal potency as compared with the standard, and Amphotericin B. EAE seems to be the most effective, especially against Candida albicans, followed by BE and CHE. In fact, C. albicans ATCC 2019 was found to be the most sensitive to A. tenuifolius extracts, while C. krusei ATCC 6258 was the most resistant yeast. In addition, the MIC values related to the extraction solvents on the four Candida species are similar (ranging from 12.5 to 37.5 mg/mL). Moreover, concentrations ranging from 50 to >75 mg/mL for almost all tested Candida strains were sufficient to reproduce a fungicidal effect. Amphotericin B was more efficient on the four tested Candida species in comparison with the three types of extract tested with low MIC values, ranging from 0.024 to 0.195 mg/mL, and low MFC values (0.39 to 6.25 mg/mL). Previous results have shown a moderate antifungal activity of A. ramosus L., and A. tenuifolius L. tested on C. albicans, C. dubliniensis, C. glabrata and C. krusei, with a diameter of growth inhibition zone ranging between 10 and 16 mm [61]. Recently, Soliman et al. [59] reported that the ethanolic extract of A. tenuifolius (100 µg/mL) inhibits the growth of C. albicans on Lauria Bertani agar with GIZ = 16 ± 0.5 mm. Additionally, Salhi et al. [62] reported that the A. tenuifolius aqueous extract (at 20%) was able to inhibit mycelial growth in Fusarium graminearum with a percentage of about (60.34%).
According to reported data [33,63], an extract has bacteriostatic effect when the ratio MBC/MIC is more than 4 and a bactericidal effect if the ratio MBC/MIC is less than 4. In our study, and based on the obtained results, the effect of CHE was bacteriostatic rather than bactericidal.

Cytotoxic and Antiviral Activities
Our results showed that only BE was safe and non-toxic, with CC 50 value of about 1340 µg/mL, while chloroformic extract and ethyl acetate extract exhibited moderate cytotoxicity on VERO cell line, with CC 50 values of 400 and 333 µg/mL, respectively, indicating the presence of some cytotoxic compounds responsible for the observed toxicological activity [47]. Consistent with our results, Soliman et al. [59], testing different A. tenuifolius extracts at different concentrations by using fresh human erythrocytes, reported that all extracts are safe and not toxic.
The antiviral activity was evaluated for the first time on A. tenuifolius extracts against Herpes Simplex Virus type 2 (HSV-2) and Coxsackievirus B-3 (CVB-3), but results showed that no extracts were active, despite the high number of polyphenols reported in the literature to possess inhibitory activity against viruses [64].

ADME Predictions
Pharmacokinetic and drug-likeness properties were evaluated using the SwissADME online program to predict whether the identified bioactive molecules become a starting scaffold or lead compounds toward future synthetic drug discovery. As shown ( Table 4), out of the present compounds, only 3-6 and 8 were estimated to have high absorption in the gastrointestinal tract, which makes them a successful drug. Another advantage is that they are not P-gp substrate (except 1), which makes them a good candidate against multidrug-resistant cancer cells, overexpressing this drug transporter. They were predicted to not be blood-brain-barrier (BBB)-permeant, meaning that they are not able to cross the blood-brain barrier into the brain, where it binds to specific receptors. The prediction of putative drug-drug interaction through Cytochromes P450 (CYPs) inhibition, which affects the metabolism of numerous xenobiotics, demonstrates that only compounds 5, 6 and 7 were expected to be inhibitors of CYP1A2. The compound 17 affected CYP2C19; those 7 and 10, 12-15 and 17 affected CYP2C9; and 3, 5 and 6 affected CYP2D6. These results suggest that all compounds may be metabolized by more than one enzyme, which in turn can minimize the risk of drug-drug interaction. In addition, eight of the twelve compounds were not inhibitors of the isoenzyme CYP3A4, which is largely implicated in the metabolism and elimination of the majority of clinically used drugs. As for the skin sensitivity prediction given by their logKp values, all tested compounds displayed negative values ranging from −8.95 to −5.24 cm/s, meaning that they are not accessible through the skin. Regarding their drug-likeness properties, more than half of the compounds considered were expected to have a good bioavailability score (0. 55 The druglikeness of the identified molecules can be estimated also through a visualization of their bioavailability radar ( Figure 3A), with the pink area representing the optimal range for each property (lipophilicity, size, polarity, solubility, unsaturation and flexibility). As shown, all compounds fall entirely in the pink area (except for unsaturation fraction), suggesting their better drug-like properties. In contrast, their pharmacokinetic properties may be also predicted via the passive gastrointestinal absorption (HIA) and brain penetration (BBB) of the top bioavailable compounds as a function of the position of the molecules in the WLOGP-versus-TPSA referential. The depicted results of the BOILED-Egg model ( Figure 3B) indicate clearly that only compounds 5, 6, 8 and to lesser degree 3 with a red point in the white ellipse have a high probability of being passively absorbed by the gastrointestinal tract and are non-substrate of the P-gp.

Conclusions
The aerial parts of Algerian A. tenuifolius were subjected to successive solvent fractionation, and chloroform extract was targeted. To the best of our knowledge, this work represents the first attempt to study the chemical composition by LC-ESI/MS analysis which led to the tentative identification of 12 compounds out of 17 detected, as well to study the biological activities, especially antibacterial, cytotoxic and antiviral, of the various solvent extracts. In this context, A. tenuifolius chloroformic extract gave interesting results in terms of both antioxidant and antibacterial activities. Our findings confirm the interesting potential of this plant as a valuable source of natural bioactive molecules that can be used in the food industry. The ADME of some isolated compounds from chloroformic extract demonstrate their good bioavailability and drug-likeness properties, especially tamgermanetin, luteolin, apigenin and aloe-emodin, which suggest, in parallel with in vivo and preclinical assays, a future new drug candidate.  Data Availability Statement: All data generated or analyzed during this study are included in this article.

Acknowledgments:
The authors are grateful to the Algerian Ministry of Higher Education and Scientific Research (MESRS) for the financial support. The authors would like also to thank all of the colleagues who contributed to this study, in particular, Adriano Sterni, from the University of Trento for the LC-MS and mass spectra recording.

Conflicts of Interest:
The authors declare no conflict of interest.