Antioxidant and Antimicrobial Effects of Baby Leaves of Amaranthus tricolor L. Harvested as Vegetable in Correlation with Their Phytochemical Composition

Amaranth is used as a spinach replacement; therefore, it is sometimes called Chinese Spinach. So far, the activity of the plant has not been associated with the presence of specific compounds. Three cultivars of Amaranthus tricolor L. were investigated for their antioxidant and antimicrobial activities. The correlation between the bioactivity and metabolite profiles was investigated in order to indicate active compounds in A. tricolor. The phytochemical profile of a total of nine extracts was studied by HPLC-DAD-ESI/HRMS, revealing the presence of 52 compounds. The highest antioxidant activity was noticed in the Red cultivar (0.06 mmol TE/g DE (Trolox Equivalent/Dry Extract Weight) and was related to the presence of amino acids, flavonoids and phenolic acids, as well as individual compounds such as tuberonic acid hexoside. All studied extracts revealed antimicrobial activity. Gram-positive bacteria were more susceptible to N-(carboxyacetyl) phenylalanine, phenylalanine, tuberonic acid and succinic acid and Gram-negative bacteria to dopa, tryptophan, norleucine, tuberonic acid hexoside, quercetin-O-hexoside, luteolin-O-rhamnosylhexoside, luteolin-6-C-hexoside succinic acid, gallic acid-O-hexoside, dihydroxybenzoic acid and hydroxybenzoic acid. Maleic acid showed promising antifungal activity. In summary, A. tricolor is a good source of antioxidant and antimicrobial compounds.


Introduction
Amaranth is a well-known and popular plant. In Central America, it was the main food source along with corn and beans. Leaves, stems and seeds may be eaten raw and cooked. The leaves have a high nutritional value. It is one of the easiest plants to cultivate due to its resistance to environmental conditions. There are over 800 varieties of Amaranthus, and among them is Amaranthus tricolor L. [1][2][3], which have been used as a traditional medicine in Siddha and Ayurveda in the treatment of diarrhea, menorrhagia, intestinal hemorrhage, dysentery, hemorrhagic colitis, bronchitis and cough [4]. A. tricolor is used in the treatment of a number of diseases. Studies have shown that crude extracts of A. tricolor leaves have antioxidant, hepatoprotective, antimicrobial, anti-inflammatory and anticancer activities [1,5].
There has been an increase in interest in the use of compounds that prevent or reduce the effects of oxidative stress (OS) in living cells [5][6][7]. OS plays an important role in the aging process and the development of chronic and degenerative diseases [7]. Antioxidants have different mechanisms of action, as well as solubility, redox potential and mechanisms of action. They can quench reactive oxygen species (ROS) and reactive nitrogen species (RNS). It is possible that others may interfere with the oxidizing of metal ions or inhibit the activity of oxidative enzymes. The presence of antioxidants may increase the potential of other compounds [6]. The study of the antioxidant potential of compounds is carried out using various methods taking into account different reaction mechanisms [6,7].
The antimicrobial activity of the compounds may be related to their antioxidant properties. Microbial invasion can lead to inflammatory processes in the body with the release of a large variety of oxidants, cytokines and proteolytic enzymes by various cells of the immune system. Insufficient responses from the immune system will lead to elevated microbial colonization. Compounds with antimicrobial activities have been shown to be more effective in reducing the immune system responses [6].
There are studies of A. tricolor leaves (cv. Valentina). A total of 41 metabolites were indicated, including amino acids, organic acids, phenolic acids and fatty acids [8,9]. A. tricolor is a source of a rare group of highly bioactive compounds-betacyanins [10,11], which are not common in nature. Their occurrence is limited to a few families of the Caryophyllales order and some higher-order fungi. There is significant interest in these compounds due to their confirmed antioxidant, anti-inflammatory, antimicrobial, anticancer, neuroprotective and hepatoprotective activities [1,5].
The activity of A. tricolor has not been associated with specific compounds. The study reports correlations between the antimicrobial, antioxidant and phytochemical profiles of various edible A. tricolor Calaloo cv. Red, Passion and Green. Fresh leaves used as leafy vegetables were extracted in food-acceptable solvents: water, ethanol and acetone. Furthermore, the study of the phytochemical profile indicates the best source of active compounds (variety and extraction conditions). Research on the metabolites, as well as antioxidant and antimicrobial properties of A. tricolor Calaloo cultivars, were performed for the first time.

Betacyanins
The Amaranthus genus has been reported as a very good source of amaranthin and isoamaranthin, but there is no detailed report on the betacyanin profile in edible leaves of young A. tricolor cv. Callaloo. Extensive research in the genus Amaranthus L. demonstrated that amaranthin and isoamaranthin are dominant compounds. The authors also indicate the presence of betanin/isobetanin, celosianin I/isocelosianin I and celosianin II/isocelosianin II in selected species [10].

Fatty Acids
Nine fatty acids were tentatively identified (Figures 1 and 2 [30]. Fatty acids were previously identified in the genus Amaranthus [9] but solely palmitic acid, stearic acid, oleic acid and linoleic acid. Here, the fatty acids were mainly extracted from A. tricolor Calaloo Passion and Green. Furthermore, tuberonic acid hexoside (17) and trihydroxyoctadecadienoic acid (19) were the dominant compounds.  Table 1). These compounds extracted mainly with ethanol and acetone proved the MS/MS spectral data similar to those from Gök et al. [19].

Phenolic Acids
Phenolic acids were the most abundant constituents of the three species of A. tricolor. Performed on the basis of HRMS, MS/MS data and comparison with fragmentation paths described in the literature [15,17,19,21,25,26], 14 compounds were tentatively described (Figures 1 and 2; Table 1).
The MS/MS spectrum provided fragments obtained by losing the CO 2 group (−44 Da) from the molecule, which is characteristic for phenolic acids. This regularity was noticed for both phenolic acids and glycosylated conjugates [19].

Antioxidant Assays
The antioxidative capacities of the H 2 O/EtOH/Ac extracts of three A. tricolor Calaloo cultivars: Red, Passion and Green (Figures 3 and 4) were assessed by employing three in vitro cell-free assays: ABTS, FRAP and DPPH, and the results are shown in Table 2 expressed as mmol Trolox Equivalents per grams of dry extract weight (DE).   Table 2. Antioxidant properties of three cultivars of A. tricolor Calaloo (Red, Passion and Green) assessed by the ABTS, FRAP and DPPH assays, and correlation coefficients between identified metabolites (absolute peak areas) and antioxidant activity (mmol TE/g DE). Statistical significance is marked by font: boldface means 95% significance and very strong correlation (R = 0.7-1.0), and italic font means 95% significance and strong correlation (R = 0.5-0.7).  In the ABTS assay, the values were in the range of 0.015 to 0.060 mmol TE/g DE, which represents a variation 4-fold. In this test, Red-H 2 O showed the highest antioxidant activity (0.060 mmol TE/g DE) compared to the other varieties, while the lowest activity showed Green-Ac (0.015 mmol TE/g DE).
In the DPPH assay, the values were in the range of 0.009 to 0.021 mmol TE/g DE, which represents a variation of approximately 2-fold. In this assay, Red-H 2 O also possessed the highest antioxidant activity (0.021 mmol TE/g DE), while the lowest activity also showed Green-Ac (0.009 mmol TE/g DE).
The FRAP values varied from 0.009 to 0.053 mmol TE/g DE, which represents a higher variation than in the ABTS and DPPH assays of 6-fold. In this test, Red-H 2 O also had the highest antioxidant activity (0.053 mmol TE/g DE). Green-Ac possessed the lowest antioxidant potential (0.009 mmol TE/g DE) for the ABTS and DPPH tests.
The extrahent had a great influence on the antioxidant activity, which can be seen in Table 2. According to the three ABTS, FRAP and DPPH tests used within a given variety of A. tricolor, the activity decreased depending on the solvent used in the sequence water > ethanol> acetone.
Ascorbic acid is very popular for its antioxidant properties, so it was used as a reference compound. The investigated extracts of three A. tricolor Calaloo cultivars (Red, Passion and Green) showed much lower antioxidant activity than pure ascorbic acid. However, it is important to remember that the plant material matrix contains a wide range of different compounds, possibly including those that can inhibit free radical scavenging. Therefore, comparisons of the results for pure compounds with the results obtained on complex extracts should be treated with approximation.
The obtained results of the antioxidant activity of ascorbic acid measured by the ABTS and DPPH methods are similar (7.8 and 8.1 mmol TE/g DE, respectively). However, the result obtained by the FRAP method (4.6 mmol TE/g DE) differs significantly from the others, which may be due to the fact that the presence of iron (III) ions in the FRAP reagent can significantly intensify the oxidation of ascorbic acid [31].

Correlation between Antioxidant Activity and Phytochemical Composition
The antioxidant activity results (mmol TE/g DE) of the A. tricolor Calaloo extracts were correlated with their metabolite composition (absolute peak area of each assigned peak from the chromatograms) and are presented in Table 2.
There was no significant correlation between the antioxidant activity and betacyanins or organic acids (p <0.05), which clearly shows that other compounds might be responsible for the antioxidant potential of the extracts tested. However, the antioxidant potential of individual, pure betacyanins and organic acids cannot be excluded, which has been confirmed previously in numerous times in the literature [10,11,32,33]. It should be noted that Zhang et al. [34] also investigated the correlation between organic acids and antioxidant activity measured by the three DPPH, ABTS and FRAP assays. These results indicate that organic acids exhibited a very low contribution to the total antioxidant activity [34].
Our experiments revealed that three compounds from the group of amino acids (glutamic acid (5), dopa (8) and norleucine (12) showed a strong positive correlation (R = 0.5-0.7). Dopa had a significant and strong correlation with the antioxidant activity measured by the ABTS, FRAP and DPPH methods (R = 0.550, 0.690 and 0.640, respectively). The glutamic acid correlations shown with the FRAP and DPPH assays were R = 0.539 and 0.524, respectively; however, norleucine had a significant and strong correlation assessed only by the FRAP method (R = 0.595).
It has been confirmed in the literature that glutamic acid markedly increases the total phenol concentration and antioxidant activity [35]. Dopa, used as a drug in Parkinson's disease, has been found to be an effective antioxidant in different in vitro assays, including anti-lipid peroxidation; reductive ability; ABTS, DPPH and superoxide anion radical scavenging activities; hydrogen peroxide scavenging and metal chelating activities with efficiency compared to the standard antioxidant compounds, such as α-tocopherol and Trolox. Furthermore, Dopa oxidation products prevented H 2 O 2 -induced oxidative damage to cellular DNA in cultured tissue cells [36]. Norleucine was also confirmed to have unusually strong antioxidant activity [37]. Here, glutamic acid (5), dopa (8) and norleucine (12) were not dominant amino acids in the A. tricolor Calaloo varieties. A higher content of these compounds was noticed in A. tricolor Calaloo Passion and Green.
Almost all fatty acids showed negative or weak correlations (R = 0-0.3) with the antioxidant activity. An exception to this was tuberonic acid hexoside (17), which showed a very strong (R = 0.7-1.0) and significant correlation with the antioxidant potential evaluated by the FRAP assay (R = 0.755) and a strong and significant correlation (R = 0.5-0.7) with the activity evaluated by the DPPH assay (R = 0.678). A. tricolor Calaloo seems to be a rich source of tuberonic acid hexoside (17). Trihydroxyoctadecadienoic acid (19) present in high concentrations in A. tricolor also showed correlations with the antioxidant potential assessed by the FRAP (R = 0.068) and the DPPH (R = 0.239) assays.
The oxidation fatty acids rate depends on the number of double bonds in the carbon chain. Therefore, the susceptibility to oxidation increases exponentially in proportion to the number of unsaturated bonds in fatty acids. Numerous studies have confirmed the antioxidant activity of fatty acids [38][39][40].
Almost all compounds from the flavonoids group showed a positive correlation, while only luteolin-6-C-hexoside (28) had significant and very strong correlations with the an-tioxidant activity measured by the ABTS, FRAP and DPPH assays (R = 0.765, 0.885 and 0.703, respectively), which the literature confirmed [12,21]. It is worth noting that quercetin O-hexoside (26) showed very strong and significant correlation with the antioxidant potential evaluated only by the FRAP method (R = 0.775) and a strong correlation against activity assessed by the ABTS and DPPH methods (R = 0.590 and 0.500, respectively). Both compounds are present in high concentrations in A. tricolor extracts.
Compounds belonging to the phenolic acid group showed both positive and negative correlations. Only hydroxybenzoic acid (45) had very strong and significant correlations with the antioxidant potential assessed by the ABTS and FRAP assays (R = 0.708 and 0.772, respectively) and a strong correlation with activity evaluated only by the DPPH method (R = 0.546). Gallic acid (39) showed strong and significant correlations with activity assessed by the methods ABTS, FRAP and DPPH (R = 0.629, 0.649 and 0.541, respectively). However, three compounds from this group (vanillic acid (42), dihydroxybenzoic acid (44) and p-coumaric acid (48) showed a strong correlation with the antioxidant potential assessed by only one assay.
In the past decade, phenolic acids have been demonstrated to possess potent antioxidant activities, which mainly depend on the number and arrangement of hydroxyl groups. In addition, they are indicated as antimicrobials.

Antimicrobial Activity
The The finest antimicrobial activity was characterized by acetone extracts (with the exception of Passion). The different values of the obtained results may result from the fact that medium-polar acetone is better extrahent of organic compounds present in plant material than polar water or ethanol, and thus, it can extract compounds that either increase the antibacterial activity or inhibit the action of bioactive compounds.
The extracts of edible leaves of A. tricolor turned out to be a much better source of bioactive compounds, showing significant antimicrobial activity (MIC = 8-16 mg/mL) compared to the results obtained by Abdoulaye et al. [41] for the methanol extract from A. cruentus leaves against stains of S. aureus ATCC 25923, E. coli ATCC 25922 and P. aeruginosa ATCC 27853 with MIC values greater than 30 mg/mL [41].
MIC for the reference antimicrobial substances were the following: MIC of vancomycin for S. aureus ATCC 29213 was 1 µg/mL, MIC of ciprofloxacin for E. coli ATCC 25922 was 0.5 µg/mL and MIC of fluconazole for C. albicans ATCC 10231 was 1 µg/mL. Vancomycin is a glycopeptide antibacterial developed as an alternative penicillin to treat strains of almost all Gram-positive bacteria, such as Staphylococcus aureus [42].
Ciprofloxacin is a broad spectrum fluoroquinolone antibacterial agent and is effective in the treatment of a wide variety of infections, particularly those caused by Gram-negative pathogens [43], while fluconazole is a triazole antifungal agent that is now an established part of therapy in patients with immune deficiencies [44].
The reference compounds show significantly higher antimicrobial activity than the tested extracts of A. tricolor, but it should be noted that the extracts have complex compositions; therefore, comparisons of pure compounds with the complex matrix should only be indicative (Table 3). Table 3. Antimicrobial activity of three cultivars of A. tricolor Calaloo (Red, Passion and Green) assessed as the MIC (minimum inhibitory concentration), MBC (minimum bactericidal concentration), MFC (minimum fungicidal concentration) and correlation coefficients between identified metabolites (absolute peak areas) and microbial activity (MIC values).

Gram-Positive Bacteria
Gram
Reference strains of bacteria and yeast from the American Type Culture Collection (ATCC, LGC Standards, Teddington, UK) were used in the study.

ABTS Radical Scavenging Assay
The nine extracts of A. tricolor were assessed by the 2,2 -Azino-bis(3-ethylbenzothiazoline-6-sulfonic) acid (ABTS) radical scavenging activity. During the reaction of ABTS with sodium persulfate in the dark for 16 h, radical cation (ABTS +• ) was obtained. 40 µL ABTS +• (1 mM aqueous solution) was mixed with various volumes of aqueous extracts with a starting concentration of 10 mg/mL for the Red, Passion and Green varieties extracted with acetone (Red-Ac, Passion-Ac, Green-Ac) and for Green extracted with ethanol (Green-EtOH). The starting concentration of 5 mg/mL was for the Red variety, the extractant of which was ethanol (Red-EtOH) and for the water-extracted Passion (Passion-H 2 O), while the other extracts had starting concentration of 2.5 mg/mL. The absorbance decreases in the range of 10-90% of its initial intensity was obtained by adjusting the volume of the aqueous extracts. The final concentration of the extracts ranged from 0 to 3.5 mg/mL (for the initial concentration of 10 mg/mL), 0 to 1.3 mg/mL (for the initial concentration of 5 mg/mL) and 0 to 0.9 mg/mL (for the initial concentration of 2.5 mg/mL) in 200 µL of the total volume of each sample. In the same way, reference compounds such as Trolox (0.025 mg/mL) and ascorbic acid (0.020 mg/mL) were prepared. After 30 min of reaction kept in the dark, the absorbance of the mixture was read on a microplate reader (Infinite M200, Tecan, Austria) at λ 734 nm at 20 • C. All experiments were repeated three times. Water plus plant extracts solution was used as a blank, while 40 µL 1 mM ABTS solution plus 160 µL water was used as a negative control. The positive control was 40 µL 1 mM ABTS solution plus 0.1 mM ascorbic acid. The results were reported as mmol Trolox Equivalent per gram of dry extract (mmol TE/g DE) of each sample following the equation for Trolox y = −20.364x + 0.943 (R2 = 0.9994), which indicates how many times the given extract potential is higher or lower than the standard.

DPPH Radical Scavenging Assay
The studied extracts were also assessed by 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical scavenging assay. The references were the same as before in ABTS assay. 60 µL DPPH • (1 mM ethanolic solution) was mixed with aqueous extracts with starting concentration of 30 mg/mL for Green-Ac, 20 mg/mL for Passion-H 2 O and Green-H 2 O, while the other extracts had starting concentration of 10 mg/mL. A decrease in radical absorbance in the range of 10-90% of its initial intensity was obtained by adjusting the volume of the extracts. The final concentration of the extracts ranged from 0 to 12.0 mg/mL (for the initial concentration of 30 mg/mL), 0 to 6.0 mg/mL (for the initial concentration of 20 mg/mL) and 0 to 4.0 mg/mL (for the initial concentration of 10 mg/mL) in 200 µL of the total volume of each sample. The references were prepared in the same way. After 30 min of reaction kept in the dark, spectrophotometric measurements were performed on microplate reader (Infinite M200, Tecan, Austria) at the wavelength 515 nm at 20 • C. All experiments were repeated three times. Ethanol plus plant extracts solution were used as a blank, while 60 µL 1 mM DPPH solution plus 140 µL ethanol was used as a negative control. The positive control was a 60 µL 1 mM DPPH solution plus 0.1 mM ascorbic acid. The results were reported as mmol Trolox Equivalent per gram of dry extract (mmol TE/g DE) of each sample following the equation for Trolox y = −31.873x + 1.2759 (R2 = 0.9993), which indicates how many times the given extract potential is higher or lower than the standard.

FRAP-Ferric Reducing Antioxidant Power Assay
The ferric reducing antioxidant power method of FRAP was used to measure antioxidant activity. As the references, Trolox (0.05 mg/mL) and ascorbic acid (0.02 mg/mL) were used. The FRAP reagent was freshly prepared by mixing 300 mM buffer acetate pH 3.6 with 20 mM of FeCl 3 × 6 H 2 O and 10 mM of TPTZ dissolved in 40 mM of hydrochloric in ratios of 10:1:1 (v/v/v), respectively. The 133 µL of freshly prepared FRAP reagent is added to an appropriate volume of the aqueous extract selected to ensure that the absorbance of the sample is within the range of the Trolox standard curve for a total volume of 200 µL. The starting concentration for the aqueous solution of the Green-Ac and Red-Ac varieties was 10 mg/mL, for the Green-EtOH and Red-EtOH varieties and Passion-H 2 O or Passion-Ac, the initial concentration was 5 mg/mL, while for the remaining extracts it was 2.5 mg/mL. After 10 min of reaction kept in the dark, a microplate reader (Infinite M200, Tecan, Austria) was used to spectrophotometric measurements at λ 593 nm at 20 • C. All experiments were repeated three times. Water plus the solution of plant extracts were used as a blank, while 133 µL FRAP solution plus 67 µL water was used as a negative control. The positive control was 133 µL FRAP solution plus 0.1 mM ascorbic acid. The results were reported as mmol Trolox Equivalent per gram of dry extract (mmol TE/g DE) of each sample following the equation for Trolox y = 25.739x + 0.0554 (R2 = 0.999), which indicates how many times the given extract potential is higher or lower than the standard.

Antimicrobial Activity
According to the recommendations of the European Committee on Antimicrobial Susceptibility Testing (EUCAST) [45], 96-well microtitrate plates were used to perform in vitro antimicrobial activity in extracts of A. tricolor. MIC (minimum inhibitory concentration), MBC (minimum bactericidal concentration) and MFC (minimum fungicidal concentration) were assessed. The reference strains of bacteria were subcultured on Mueller-Hinton Broth (MHB), while the yeast strains were on RPMI-1640 Medium (RPMI) and incubated for 18 to 24 h at 35 • C in ambient air. The suspension of microbial colonies was prepared according to the standard of turbidity of the bacterial suspension of 0.5 units of density according to McFarland, corresponding to 1.5 × 10 8 CFU (colony forming units)/mL for bacteria and 5 × 10 6 CFU/mL for yeast.
The extracts of A. tricolor (100 mg/mL) were dissolved in sterile distilled water. The final extracts concentrations diluted in MHB or MHB2% were ranged from 32 to 0.125 mg/mL. 2 µL of each inoculum was added to wells containing 200 µL of the serial dilution of the extracts and the plates were incubated for 18 to 24 h at 35 • C in ambient air. The Absorbance Microplate Reader EL × 800 (BioTek Instruments, Inc., Winooski, VT, USA) set at λ 580 nm was used to verify the MIC of the extracts, which showed complete inhibition of bacterial or yeast growth. The lowest extracts concentrations with no visible bacterial or yeast growth were assessed as MBC or MFC, indicating killing of 99.9% of the inoculum. MBC or MFC were evaluated by spreading 5 µL of microorganisms from each well of the microtitrate plates used for the evaluation of MIC, showing no growth onto appropriate culture medium (MHB for bacteria or RPMI for yeast). In ambient air, the plates for 18 to 24 h at 35 • C were incubated. The most frequently recurring representative value from the three measurement series was selected as the final result in determining the MIC, MBC, and MFC.
There is no statistical analysis involved in the development of the assays because they are from susceptibility tests. Vancomycin (0.06-16 µg/mL), ciprofloxacin (0.014-16 µg/mL), and fluconazole (0.06-16 µg/mL) were included as a reference antimicrobial compound.

Statistical Analysis
One-way analysis of variance (ANOVA) of the means of nine extracts of three A. tricolor Calaloo cultivars-Red, Passion and Green-was performed with Statistica, version 7.1 (StatSoft, TIBCO Software Inc. Palo Alto, CA, USA). The results were subjected to ANOVA and the differences between means were located using Fisher's test. Significance was assessed at α level 0.05 to find out how many and which cultivars have different contents. Data were reported as the mean ± standard deviation (SD) of three measurements.

Conclusions
The high antioxidant activity was observed in the aqueous extracts, where the Red variety of A. tricolor showed the highest activity in all the assays. A strong bacteriostatic, bactericidal and fungicidal effect against selected Gram-positive bacteria strains S. aureus, B. subtilis, M. luteus and B. cereus causing food poisoning in humans, as well as Gramnegative bacterium B. bronchiseptica responsible for mild forms of respiratory diseases in humans and yeast strain C. krusei was noticeable in the Red acetone extract of A. tri-color. Selected metabolites, such as N-(carboxyacetyl) phenylalanine, dopa, norleucine, tryptophan, quercetin-O-hexoside, luteolin-O-rhamnosylhexoside, luteolin-6-C-hexoside, succinic acid, gallic acid-O-hexoside, dihydroxybenzoic acid and hydroxybenzoic acid show antioxidant and antibacterial activities. The highest content of these compounds was identified in the Green variety of A. tricolor extracted with acetone. The effectiveness of these compounds may be higher against microbes, as the presence of antioxidants may reduce the inflammatory processes that accompany microbial diseases. The greatest effect on the antifungal activity against fungal exhibited maleic acid, for which no antioxidant activity was noticed. The Passion/Green-EtOH were the richest sources of maleic acid.
Research on the activity of pure compounds requires a considerable amount of time to separate compounds from a complex plant matrix. This study provides a tool for fast screening of plant material in search of biologically active compounds. Furthermore, this study indicates the richest source of active compounds. The activity of the compounds may not be related to their content in the extract, because the matrix components can negatively or positively affect their activities. Therefore, the measure of extract activity will not indicate the potential of the plant.