RP-HPLC/MS/MS Analysis of the Phenolic Compounds, Antioxidant and Antimicrobial Activities of Salvia L. Species

The identification and quantification of the phenolic contents of methanolic extracts of three Salvia L. species namely S. brachyantha (Bordz.) Pobed, S. aethiopis L., and S. microstegia Boiss. and Bal. were evaluated using reverse phase high performance liquid chromatography, UV adsorption, and mass spectrometry (RP-HPLC/MS). In order to determine the antioxidant capacity of these species, cupric ions (Cu2+) reducing assay (CUPRAC) and ferric ions (Fe3+) reducing assay (FRAP) were performed to screen the reducing capacity and 1,1-diphenyl-2-picrylhydrazyl (DPPH) assay was employed for evaluation of the radical scavenging activity for both solvents. In further investigation, the antimicrobial activities of Salvia species were tested using the disc diffusion method against three Gram-positive and four Gram-negative microbial species, as well as three fungi species. The results showed that there is a total of 18 detectable phenols, the most abundant of which was kaempferol in S. microstegia and rosmarinic acids in S. brachyantha and S aethiopis. The other major phenols were found to be apigenin, luteolin, p-coumaric acid, and chlorogenic acid. All species tested showed moderate and lower antioxidant activity than standard antioxidants such as butylated hydroxyanisole (BHA), butylated hydroxytoluene (BHT), and ascorbic acid. The ethanolic extracts of Salvia species revealed a wide range of antimicrobial activity. S. brachyantha and S. microstegia showed the highest antimicrobial activities against B. subtilis, whereas S. aethiopis was more effective on Y. lipolytica. None of the extracts showed anti-fungal activity against S. cerevisiae. Thus these species could be valuable due to their bioactive compounds.


Introduction
Plants contain many bioactive phenolic compounds, which have biological activity including antioxidant and antimicrobial properties. They have many health benefits and can help in preventing performance liquid chromatography (RP-HPLC) method coupled with mass spectrometry was used, which, in the literature, is a commonly used technique for separation of phenols [28]. Since phenolic content could be connected to antioxidant and antimicrobial activity, we also measured the antioxidant and antimicrobial activities of these Salvia species. Additionally, in this study we evaluated antioxidant activity based on reducing power capacity via ferric ions (Fe 3+ ) reducing assay (FRAP) and cupric ions (Cu 2+ ) reducing assay (CUPRAC); and radical scavenging activity via 1,1-diphenyl-2-picrylhydrazyl (DPPH) scavenging activity.

Plant Materials
The environmental conditions of the studied samples are as follows: S. brachyantha grows in open, calcareous stony igneous slopes, rocky ledges, and steppes at an altitude of 1400-1600 m. The vegetation in these places is mostly formed by herbaceous and woody plants including Astragalus sp., Quercus sp., Verbascum sp. and Vicia sp. S. aethiopis grows in open Quercus forests, steppes, igneous and limestone slopes, and fallow fields, along with herbaceous plants, such S. multicaulis Vahl and S. trichoclada Benth., species of Trigonella, Astragalus, Alyssum, Galium, Vicia, and Bromus, at elevations between 1200 and 1500 m. S. microstegia grows on calcareous rocky limestone and igneous slopes, cliffs, screes, fieldsides, and among Quercus shrubs at an altitude of 1300-1600 m. The vegetation in this place is formed by herbaceous and woody plants including Quercus spp., Euphorbia spp.
All plant samples were collected from their natural habitats by Dr. Omer Kilic. The first plant material, S. brachyantha was collected from the vicinity of Saban village (Bingöl) from rocky slopes, on 18 May 2014, at an altitude of 1400-1500 m, collection number; 5681. The second plant material, S. aethiopis was collected south of Yelesen village, from stony areas, on 20 June 2013, at an altitude of 1600-1700 m, collection number; 4780. The third plant material, S. microstegia was collected from east of Dikme village, from steppe and moist areas, on 5 September 2013, at an altitude of 1750-1800 m, collection number; 5405. All plant samples were identified by Kilic with Flora of Turkey and the East Aegean Islands [29]. The voucher specimens were deposited at the Department of Park and Garden Plants, Bingol University and Yıldırımlı Herbarium from Ankara. The samples were taken from the aerial parts of whole plants. The aerial parts of the studied samples were dried in a shady and airy place, held for one week.

Plant Extraction
Plants extractions were described previously in detail [30]. An ethanol extract was prepared by grinding 100 g air-dried samples in a mill and sample powders were mixed with 300 mL ethanol on a magnetic stirrer for 24 h at room temperature. Ethanolic extracts were filtered and the mixtures were then placed on a rotary evaporator at 30 • C to remove ethanol (Yield: 15.6%). For water extract preparation, sample powders were mixed with 100 mL of distilled water, filtered, and lyophilized under 5 µm Hg pressure at −50 • C. Samples were then stored in a tightly caped plastic bottle at −20 • C until used for experimental studies.

Mass Spectroscopy (MS) Instrumentation
MS detection was performed using a Shimadzu LC-MS 8040 model triple quadrupole mass spectrometer (Shimadzu, Kyoto, Japan) equipped with an electrospray ionization (ESI) source operating in both positive and negative ionization modes. LC-MS/MS data were collected and processed by Lab Solutions software (Shimadzu, Kyoto, Japan). The multiple reaction monitoring (MRM) mode was used to quantify the analyses: the assay of investigated compounds was performed following two or three transitions per compound, the first one for quantitative purposes and the second and/or the third one for confirmation. The optimum ESI conditions were determined as DL temperature; 250 • C, heat block temperature; 400 • C, nebulizing gas flow (nitrogen); 3 L/min and drying gas flow (nitrogen); 15 L/min [32].

CUPRAC Assay
The reducing capacity of extracts was measured by the CUPRAC method [33]. In this assay, 0.25 mL CuCl 2 solution (0.01 M), 0.25 mL of ethanolic neocuproine solution (7.5 × 10 −3 M) and 0.25 mL of CH 3 COONH 4 buffer solution (1.0 M) were mixed and sample extract at different concentrations (10-30 µg/mL) was added to this mixture. The final volumes were adjusted to 2 mL with distilled water and 30 min later absorbances of the samples were measured at 450 nm. Increased absorbance was interpreted as increased reducing capacity.

FRAP Assay
This assay was used to measure the Fe 3+ ion's reducing power [34]. The samples at different concentrations in distilled water were mixed with phosphate buffer (2.5 mL, 0.2 M, pH 6.6) and potassium ferricyanide [K 3 Fe(CN) 6 ] (2.5 mL, 1%) and incubated at 50 • C for 20 min. Then, 2.5 mL trichloroacetic acid (10%) and 0.5 mL of FeCl 3 (0.1%) were added to the reaction mixture. The increases in the absorbance were spectrophotometrically measured at 700 nm as an indication of reducing capacity.

DPPH Assay
Hydrogen or electron donating abilities of the samples were measured using DPPH assay [35]. Accordingly, purple colored DPPH solution (1 mM) prepared in ethanol was added to samples at different concentrations (10-30 µg/mL). The mixture was incubated at room temperature for 30 min. and the radical scavenging activity of the samples was measured spectrophotometrically at 517 nm against a blank. Decreased absorbance of the sample indicated the DPPH free radical scavenging capability.

Microbiological Assay
The antimicrobial activities of extracts were detected by the disc diffusion method. 30, 60, and 90 µL of each extract was absorbed onto sterile discs, of 8 mm diameter. To inoculate the media for assay, a 1% rate of each microorganism from 10 6 to 10 7 CFU/mL suspensions was added to 15 mL sterile media (for bacteria Muller-Hintone agar, for yeast Sabouraud 2% glucose agar). Each of these inoculated mediums was poured into a Petri dish (9 cm) and left at +4 • C for 1 h. Subsequently discs prepared from samples were added to these inoculated medias and left again at +4 • C for 1 h. Four antibiotic standard discs were used as the positive controls. Sensitivity was deduced by comparing the inhibition zone diameter produced by the erythromycin (E-15), ampicillin/sulbactam (SAM-20), amikacin (AK-30) and rifampicin (RD-5). The Petri dishes were incubated at 35 • C for 18-24 h, except for C. albicans ATCC 10231, Y. lipolytica, and S. cerevisiae which were incubated at 27 • C. Inhibition zones were measured, calibrated, and recorded as the mean diameter of three replications in mm.

Identification of Phenols
For centuries, plant extracts have been used to treat many diseases and their mode of action may well have been based on the phenolic compound content [36,37]. Phenolic compounds are the most widely occurring groups of phytochemicals and are of considerable physiological and morphological importance in plants as well as having strong antioxidant properties [38]. Numerous studies have connected the antioxidant, anti-inflammatory, anti-cancer, and antimicrobial activities of many plants, herbs, and species to their phenolic content. For example, the anti-inflammatory, antioxidant, and anticancer activities of Suaeda fruticosa were explained by its appreciable level of phenolic compounds (31.8 mg gallic acid equivalent (GAE)/g dried weight (DW)) [39]. Therefore identification and quantification of phenols from different sources is becoming increasingly important due to their potential application for treating diseases.
The commonly used sample preparation technique for phenols is extraction with organic solvents, while spectrophotometric and chromatographic techniques are the main methods utilized in their identification and quantification [36,40]. In this study, we identified the phenolic acids of three Salvia species from Turkey, using the RP-HPLC/MS technique. According to our present knowledge, no studies have been reported regarding phenolic compounds from S. microstegia, S. aethiopis and S. brachyantha in the open literature. However, there are other studies reporting phenolic compounds of different Salvia taxa [11,17,41,42].
The typical chromatograms of the samples are given in Figure 1. The identification of each peak was possible via retention time as well as MS spectra of samples and authentic standards. The characteristics of each peak from HPLC/MS analysis of authentic standards/samples are provided in Table 1. With 18 phenolic acids, S. brachyantha was found to be richest in terms of the number and amount of phenolic compounds.

Quantification of Phenols
In order to establish the relationship between peak area and concentration, linear regression analysis was carried out for the investigated standards listed in Table 1. We constructed calibration curves in three different ranges (250-10,000 µg/mL, 25-1000 µg/mL, and 100-4000 µg/mL) for different standards. The linearity, which shows the correlation between the peak area and concentration, was expressed as correlation coefficient (r 2 ) and higher than 0.999 for all standards except for hyperoside (r 2 : 0.950). We also calculated limit of detection (LODs) and limit of quantification (LOQs), which ranged from 0.05-25.8 to 0.17-74.5, respectively (Table 1) [43].
With the aid of the established calibration curves of the standard compounds, we then quantified the phenolic compounds based on the area of individual peaks from the HPLC chromatogram profile of the samples and compared them with the areas of the standards at known concentrations. The amount of each compound was expressed as µg/kg dried Salvia extracts. The main phenolic acids were kaempferol in S. microstegia with 1954.28 µg/kg, rosmarinic acid in S. brachyantha with 7619.58 µg/kg, and S. aethiopis with 1904.53 µg/kg, respectively. By comparing Figure 2B,C and considering Table 2, it can be seen that, in terms of phenolic acid content, S. microstegia and S. brachyantha gave similar results. The less abundant phenolics were apigenin, p-coumaric acid, and luteolin for S. microstegia; kaempferol and apigenin for S. aethiopis; and chlorogenic acid and rutin for S. brachyantha. Differences in the variety and amount of phenolic compound from different Salvia species are expected since it is possible that the harvesting season and geographic origin affect the phenolic contents [44]. In addition, each species has a unique phenolic composition peculiar to that plant. For example, six different Artemisia species exhibited significant dissimilarities in terms of antioxidant capacities and phenolic compounds [45]. Furthermore, a study where eight Salvia species, namely S. aethiopis, S. candidissima, Salvia chionantha Vahl., S. limbata, S. microstegia, S. nemorosa L., S. pachystachya Trautv., S. verticillata L., and S. virgata Jacq., collected from Turkey were investigated for their amount of phenolic compounds, revealed that the amount of total phenolic content varies from 50.3 to 167.1 mg GAE/g DW among species [46]. Moreover, the extraction solution could be also affecting phenolic contents. For instance, the separate usage of water, 80% methanol, and 70% ethanol extracts of Moringa oleifera Lam. leaves resulted in 7.43%, 12.33%, and 11.04% in total phenolics, respectively [47].

Quantification of Phenols
In order to establish the relationship between peak area and concentration, linear regression analysis was carried out for the investigated standards listed in Table 1. We constructed calibration curves in three different ranges (250-10,000 μg/mL, 25-1000 μg/mL, and 100-4000 μg/mL) for different standards. The linearity, which shows the correlation between the peak area and concentration, was expressed as correlation coefficient (r 2 ) and higher than 0.999 for all standards except for hyperoside (r 2 : 0.950). We also calculated limit of detection (LODs) and limit of quantification (LOQs), which ranged from 0.05-25.8 to 0.17-74.5, respectively (Table 1) [43].

Antioxidant Activity
An antioxidant molecule is a substance that, even at low concentration, can hinder or delay the oxidation of a substrate [48,49]. Hence there are multiple ways of measuring the antioxidant effect of a source to have complete understanding of the mechanism of action. Furthermore, an extract showing low antioxidant activity could not be labelled as a poor source of antioxidant, since an extract is composed of chemicals with different functional groups and polarities and may behave differently depending on the reaction mixture [50]. We used ethanol as well as water as extraction solvents to dismiss the possibility that some of the polyphenols are not extractable in water due to poor solubility [51]. In this study, the antioxidant capacity of three Salvia species were measured using FRAP, CUPRAC, and DPPH assays.
The CUPRAC assay is widely used for measurement of the antioxidant capacity of plant extracts, due to the requirements of standard equipment, as well as fast and reproducible results [50]. This assay is based on the measurement of the absorbance of a complex that results from the reaction of antioxidant with Cu 2+ -neocuproine reagent [52]. The results showed that the cupric ions (Cu 2+ ) reducing capacity of the ethanol extract of S. brachyantha was higher than the other samples tested, but lower than synthetic antioxidants butylated hydroxyanisole (BHA) and butylated hydroxytoluene (BHT) (Figure 2A). The water extract of S. aethiopis and the ethanol extract of S. microstegia showed an activity similar to ascorbic acid (Figure 2A,D). The lowest activity was observed in the water extract of S. microstegia (Figure 2A). It is worth noting that the ethanol extract of S. microstegia showed higher activity than the water extract, which might indicate that the ethanol extract has a higher phenolic content than the water extract. Similar findings were reported for Salvia species. For instance a previous study revealed that Salvia chionantha possessed lower CUPRAC values than BHA and α-tocopherol [53]. As mentioned above, another factor affecting the antioxidant capacity is the selection of extraction solvent, for example, ethyl acetate extract of S. cadmica indicated a CUPRAC value that was almost a third of the one where methanol was used as extraction solvent [11].
FRAP assay was used to measure the ferric ion (Fe 3+ ) reducing capacity of the samples ( Figure 2B). The overall results showed that all samples from both extraction solutions showed lower ferric ion reducing capacity than standard synthetic antioxidants ( Figure 2B,E). The ethanol extract of S. aethiopis showed higher reducing power at a concentration of 30 µg/mL, the same as BHT. Similarly, S. verticillata has been previously shown to have a reducing activity less than synthetic antioxidant compounds (BHA and BHT) [41]. It was also shown that Salvia species could have different ferric ion reducing activities that could be classified as good, moderate, and poor antioxidant activity depending on the habitat in which they had been grown [54].
DPPH radical scavenging activity was also performed to gain a better understanding of the antioxidant activity of the samples [55]. An antioxidant molecule is able to reduce the stable radical DPPH, having a deep violet color in solution, to the yellow-colored 1,1-diphenyl-2-picrylhydrazyl (DPPH 2 ) [56]. The results showed that ethanol and water extracts of all samples showed moderate radical scavenging activity when compared to standard antioxidants ( Figure 2C,F). All of the Salvia species were able to donate electrons to neutralize the DPPH radical, showing a catalase-like, superoxide dismutase (SOD)-like activity [57]. The highest and lowest activities were observed in S. brachyantha and S. aethiopis, respectively. Among the samples tested S. brachyantha s showed a slightly better antioxidant activity in the CUPRAC and DPPH assay. The reason for this could be due to significant amounts of quinic and trans-caffeic acid found in S. brachyantha (Table 1). The findings of this study are consistent with the literature. It has been found that Brazilian S. officinalis has a DPPH scavenging activity comparable to gallic acid [57]. In addition S. palaestina essential oil showed an increasing DPPH radical scavenging activity in a concentration dependent manner (from 0.122 to 1.35 mg/mL), which was a higher weight concentration than the ones used in this study [58]. Different antioxidant capacities from different species are also to be expected. For instance, a study, where six Salvia species were tested for their antioxidant properties, showed that the DPPH radical scavenging activities of S. aethiopis and S. candidissima were 0% and 49.7%, respectively.

Antimicrobial Activity
Many microorganisms, which cause damage to human health, exhibit drug resistance due to inadequate use of antibiotics. Thus, there is a need for the discovery of new substances from natural sources, including plants. The in vitro antimicrobial activity by the agar disc diffusion method of extracts of S. brachyantha, S. aethiopis, and S. microstegia resulted in a range of growth inhibition patterns against pathogenic microorganisms given in Table 2.
In this study, the strongest antibacterial activity was observed in the extracts of S. brachyantha and S. microstegia against B. subtilis (13 ± 0.47 mm inhibition zone) ( Table 2). However the extracts of S. brachyantha and S. microstegia showed very weak antibacterial activity against S. aureus and K. pneumoniae. Meanwhile, the extract S. aethiopis extract did not show any antibacterial activity against B. subtilis, B. megaterium, and K. pneumoniae. The extract of S. aethiopis showed the highest antifungal activity against Y. lipolytica (13 ± 1.69 mm) and also showed no antifungal activity against C. albicans and S. cerevisiae, but the extract of S. brachyantha showed the highest antifungal activity against C. albicans and Y. lipolytica (10 ± 0.00 mm). The extracts did not show any antifungal activity against S. cerevisiae (Table 2).
It was observed that the antimicrobial effect of the plant extract varies from one plant to another in the different researches carried out in various regions of the world. This may be due to many factors such as, the effect of climate, soil composition, age, and vegetation cycle stage on the quality, quantity, and composition of extracted product of different bacterial strains [59,60].
The degree of antibacterial activity of the samples also changed with the increase in concentration of the samples. Roughly 1-2 mm increases were observed when the concentrations of samples were increased from 30, 60 to 90 µL. Antibacterial activities of S. brachyantha, S. aethiopis, and S. microstegia were also compared with reference antibiotics. As shown in Table 1, the extracts showed similar rates of activity with ampicillin/sulbactam (SAM-20 mg) and amikacin (AK-30 mg), but erythromycin (E-15 mg) and rifampicin (RD-5 mg) showed lower activity.
Similar findings could be found in the literature. For example S. spinosa exhibited potential antimicrobial effects against different tested Gram-negative and -positive strains [61]. Similarly, Salvia chudaei species was found to be effective against seven out of nine bacteria tested [42]. A different study showed that among eleven Salvia species, S. eremophila, S. limbata, S. santolinifolia, and S. sclarea indicated the highest antimicrobial activity whereas S. aegyptiaca and S. aethiopis showed the weakest activity against the tested organisms [17].

Conclusions
We investigated the antioxidant and antimicrobial activity of the phenolic compounds of three Salvia species from Turkey. This study is the first in which the identification and quantification of phenolic acids from S. brachyantha, S. aethiopis and S. microstegia have been accomplished. The results indicated that the methanol extract of S. microstegia contained considerable amounts of kaempferol while S. brachyantha and S. aethiopis were rich in rosmarinic acid. The findings of this study showed that all three species have reasonable and variable antioxidant activity depending on the method and extraction solution used. The compound extracted from S. brachyantha, S. aethiopis, and S. microstegia exhibited a broad spectrum of antimicrobial activity, which could be used as an alternative source for antibiotics. However, pharmacological testing is necessary to isolate and characterize the active compounds. These plant extracts should be investigated in vivo to better understand their safety, efficacy, and properties. Overall these species appear to be promising sources of various bioactive compounds that could be specifically used to treat certain diseases. data analysis, and writing and publication of this report; Ömer Kılıç collected the plant samples from their natural habitats; Yusuf Alan conceived, designed and performed the antimicrobial experiments; Mustafa Abdullah Yılmaz analyzed the phenolic compounds of the plant materials by RP-HPLC/MS/MS; Saleh H. Alwasel supervised and reviewed the manuscript. All authors have read and approved the final manuscript.

Conflicts of Interest:
The authors have declared no conflict of interest.