Chemical Composition and Biological Activity of Extracts Obtained by Supercritical Extraction and Ethanolic Extraction of Brown, Green and Red Propolis Derived from Different Geographic Regions in Brazil

The variations in the chemical composition, and consequently, on the biological activity of the propolis, are associated with its type and geographic origin. Considering this fact, this study evaluated propolis extracts obtained by supercritical extraction (SCO2) and ethanolic extraction (EtOH), in eight samples of different types of propolis (red, green and brown), collected from different regions in Brazil. The content of phenolic compounds, flavonoids, in vitro antioxidant activity (DPPH and ABTS), Artepillin C, p-coumaric acid and antimicrobial activity against two bacteria were determined for all extracts. For the EtOH extracts, the anti-proliferative activity regarding the cell lines of B16F10, were also evaluated. Amongst the samples evaluated, the red propolis from the Brazilian Northeast (states of Sergipe and Alagoas) showed the higher biological potential, as well as the larger content of antioxidant compounds. The best results were shown for the extracts obtained through the conventional extraction method (EtOH). However, the highest concentrations of Artepillin C and p-coumaric acid were identified in the extracts from SCO2, indicating a higher selectivity for the extraction of these compounds. It was verified that the composition and biological activity of the Brazilian propolis vary significantly, depending on the type of sample and geographical area of collection.


Materials and reagents
Ethanol (HPLC degree) and aseptic acid (HPLC degree) were obtained from Merck Co.

Obtaining and processing of propolis samples
Approximately 700-1000g of propolis samples were donated by the company Apis Nativa Produtos Naturais (Prodapys-Santa Catarina-Brazil), originated from the different regions in Brazil, during the period of July to September 2013. Two samples of red propolis were from the Brazilian northeast (Alagoas and Sergipe), three samples of brown propolis from the south (Santa Catarina, Rio Grande do Sul and Parana) and three samples of green propolis were from the south and southeast of Brazil (Parana and Minas Gerais) (two samples from different regions) ( Table 1). The samples of propolis were crushed in a grinder (Cadence-Brazil) and then sieved (60 mesh), in order to obtain an adequate granulometry (approximately 0.250 mm) to increase the surface area and homogenise the start material in the extraction processes. Small quantities (250g) of propolis were kept in a fridge at -10°C, in bottles protected with laminated paper in inert atmospheric conditions (N 2 ) in order to avoid degradation of the material.

Characterization of raw material
The determinations of humidity, protein and total ash contents were made according to the official methods of Association of Official Agricultural Chemists (AOAC) [40]. The total lipids were extracted and quantified through the cold extraction method described by Bligh & Dyer [41]. The determination of the mineral content was made in a digital flame photometer (DM-62, DIGIMED, São Paulo-Brazil) and the fiber content was obtained through the automatic Table 1. Identification of propolis samples from different regions of Brazil and analysed in this study.

Sample Identification
State and region of Brazil Colour Type fibre analyser (A-220, ANKON, New York-USA) [42]. The quantification of the water activity took place using a decagon LabMaster (Novasina, Lachen-Switzerland), with electrolytic cell CM-2. The analyses were performed in triplicate. The Scanning electron microscopy (SEM) was performed in a scanning electron microscope JEOL JSM-6390LV (USA). After drying in an oven (105°C/45 min), the sample of crushed propolis was fixed manually using a tweezer (PELCO1 Tweezers) of aluminum metal surfaces covered with carbon double-sided tape, called stubs. Because of the need for interaction of the electron beam with the sample, it was performed by coating deposition of metallic gold ions (sputtering). The sample was metalized in gold in a "Sputter oater" from Balzers, model SCD 50 (20nm). Then the stubs containing the metallic samples were stored in plastic boxes (storage boxes), duly sealed with parafilm (PAR-AFILM1 M) to prevent moisture absorption. After 24 hours of rest, the samples were analyzed at different magnifications (Voltage 12 kV, Working Distance 12 mm, Spot size 44, Vacuum Mode HV).
Obtaining propolis extracts by supercritical fluid extraction (SFE) and low pressure extraction (LPE) Supercritical extracts using CO 2 as supercritical fluid. The equipment used for obtaining the propolis extracts was the pilot unity called SFT-110 Supercritical Fluid Extractor (Supercritical Fluid Technologies, Inc.), composed by a high pressure bomb (capacity of up to 10,000 psi), extraction cell (capacity of 100 ml), furnace (containing a pre-warmer), static/dynamics valve and restrictor valve, flow meter, flux totalizer (ITRÓN, ACD G1.0, Argentina) and CO 2 cylinder. A CO 2 cylinder with fishing tube was used to ensure that only CO 2 in liquid state was used in the system, a requirement of the SFT-110. The CO 2 was not re-used in the system.
The extraction cell consisted of a packaging using 7.5 g of homogenised propolis sample with 1% ethanol co-solvent (m/m), wool and glass pearls, aimed at avoiding the preferred paths of CO 2 and the total filling of the bed. The extraction conditions were: pressure of 350 bar, temperature 50°C, S/F of 110 (mass of CO 2[solvent] / mass of propolis [solute] ), 1% co-solvent (ethanol m/m), flow of CO 2 of 6 g/min and total time of extraction 2 h 30 min [26,43]. The temperature of the restrictor valve was adjusted at 80°C for all extraction processes. The extracts were collected in glass vials of 50 ml, immersed in ice at room pressure. The vials containing the extracts were protected with aluminium foil in inert atmospheric conditions (N 2 ) in order to avoid degradation of the material, and kept at 5°C until analysis ( Fig 1A).
Low pressure extraction (conventional extraction). The ethanolic extracts of propolis were prepared with the addition of 15 ml ethanol (80%) to 2g propolis. The extraction occurred at 70°C temperature for 30 minutes under constant agitation in a Shaker (MA 420/MAR-CONI-Brazil) incubator, at 710 rpm rotation. Following that, the extract was centrifuged (Centrifuge SIGMA 2-16 KL) at 8800 rpm for 11 min at 5°C. At the end of the centrifugation, the supernatant was transferred to a 50 ml beaker, and 10 ml ethanol (80%) was added to the tube residue, where centrifugation was repeated. The supernatants were homogenised and kept at 50°C until completely dry ( Fig 1B). Afterwards, the extracts were stored in tubes covered in aluminium foil in inert atmospheric conditions (N 2 ) in order to avoid degradation of the material. The material was kept at 5°C until analysis [6].
Chromatographic analysis: identification and quantification of the 3,5-diprenil-4-hidroxicinamic (Artepillin C) and acid 4-hidroxicinamic (p-coumaric acid) 10 mg/ml of propolis extracts obtained in the different conditions of the process were prepared and dissolved in ethanol and placed in ultrasound bath (Sanders, SONICLEAN 6 -Minas Gerais, Brazil-ANVISA 80273140001) for 30 minutes (Electronic timer microprocessor-Temperature 35°C electronically controlled and Ultrasound frequency 40 kHz). The exposure to the ultrasound system only started after reaching 35°C. The samples were filtered in a cellulose ester membrane filter 0.45 μm (Micropore1) for posterior injection in the High Performance Liquid Chromatograph (HPLC).
The conditions for analysis were performed with an elution gradient with a mobile phase of aseptic acid 5% (aqueous phase) and methanol (organic phase) in different proportions and with total analysis time of 70 80:20). The volume of injection was of 10 μL. The equipment was operated at room temperature (25±2°C). The reading of the diode arrangement detector was in the range of 200 to 400 nm and the chromatographic acquisition was defined at 290 nm. The identification of the compounds was performed through the comparison of time of retention and ultraviolet spectrum between the samples and the controls (standard). Aiming to ensure the reliability of the results obtained, a validation took place according to the National Health Surveillance Agency (ANVISA) [44] and National Institute of Metrology, Standardization and Industrial Quality (INMETRO) [45] methodologies. This was done in accordance to the parameters of selectivity, linearity, precision, accuracy, detection limits and quantification limits.

Determination of the total phenolic compounds
The total phenolic content was determined using the Folin Ciocalteau reagent [46][47]. The reaction was prepared with a 0.5 ml aliquot of propolis extract (dissolved in ethanol aimed at obtaining a concentration of 200 μg/ml), 2.5 ml aqueous solution of Folin-Ciocalteau 10% and 2.0 ml sodium carbonate at 7.5%. The mixture was introduced in a thermo-regulated bath at 50°C for 5 minutes; afterwards, the absorbance was measured in a spectrophotometer LAMBDA 25 UV/vis Systems (PerkinElmer, Washington-USA) at 765 nm. The quantity of total phenolic was expressed as Gallic acid equivalents (EAG) (mg EAG/g of sample) through a calibration curve using known solutions to Gallic acid standard in the same conditions (λ = 765 nm). The Folin Ciocalteu method is associated to the appearance of a blue colouring due to the oxidation of phenols in basic medium [48].

Determination of flavonoid content
The determination of flavonoid content was performed through the reading in a spectrophotometer (LAMBDA 25 UV/Vis Systems-PerkinElmer USA) at 415 nm, using aluminium chloride at 2% in methanol [49] in a 1:1 solution (extract:aluminium chloride). The same procedure was performed using known solutions of quercetin standard to elaborate a standard curve. The quantity of total flavonoids was expressed as quercetin equivalents (EQ) (mg EQ/g of sample).
Determination of anti-oxidant activity in vitro (2,2-Diphenyl-1-picrylhydrazyl-DPPH) The anti-oxidant activity in vitro of propolis extracts obtained in different conditions was evaluated using the reactive 2,2-Diphenyl-1-picrylhydrazyl (also known as the capacity to sequestrate the radical DPPH) [50][51]. Five dilutions of the extracts were prepared (20 to 400 μg/ml) in triplicates. An aliquot of 1 ml of each extract dilution was transferred to assay tubes with 3.0 ml of the ethanoic solution (Ethanol-absolute alcohol 99.8%) of the radical DPPH (0.004%). After 30 minutes incubation in the dark and at room temperature, the reduction of the free radical DPPH was measured through the reading of absorbance in 517 nm spectrophotometer (LAMBDA 25 UV/Vis Systems-PerkinElmer, Washington-USA).
The same procedure was performed with ethanol replacing the sample, considered blank. The capacity to sequestrate free radicals was expressed as the percentage of oxidation inhibition in the radical and calculated according to Eq 1. The IC 50 value (necessary concentration of the extract to sequestrate 50% of DPPH radical) was calculated through the line equation based on the concentrations of extracts and its respective percentages of radical DPPH sequestration. % sequestration ¼ 100 À ½ðfinal absorbance of sample Ã 100Þ = blank absorbance ð1Þ Determination of antioxidant activity in vitro: ABTS method (2,2 0 -azinobis-3-ethylbenzthiazoline-6-sulphonic acid) was added. The final product was incubated for 16 hours, protected from light in order to enable production of the radical cation ABTS •+ . Afterwards, the solution was diluted in ethanol until it reached 700±50 absorbance, reading at 734 nm using a spectrophotometer. Originated from the formation of radicals, 20 μL aliquots of stock solutions of 1 mg.ml -1 , 0.75 mg.ml -1 , 0.5 mg.ml -1 and 0.1 mg.ml -1 taken from the extracts were added to 2 ml of the final solution of ABTS •+ , and after 6 minutes of incubation the samples were read at 734 nm absorbance. The results were expressed in TEAC (antioxidant activity equivalent to Trolox (6-hydroxil-2,5,7,8-tetramethylchromo-2-carboxilic acid)) (Vitamin E).

Antimicrobial activity of the extracts EtOH and SCO 2
The antimicrobial activity of the extracts EtOH and SCO 2 was determined through the Minimal Inhibitory Concentration (MIC) and Minimal Bactericide Concentration (MBC) against Staphylococcus aureus (ATCC 33951 and 25923) and Escherichia coli (ATCC 25922), according to the Clinical and Laboratory Standards Institute [54] and Koo et al. [55]. The strains used were provided by the Bacteria Cultures Collection of the Instituto Oswaldo Cruz-FIOCRUZ (Manguinhos-Rio de Janeiro-Brazil) and initially reactivated in liquid BHI (Brain Heart Infusion) (Sigma-Aldrich Chemical Co.-St. Louis, MO, USA) at 37°C for 24h and then grown in BHI agar plates to inoculum preparation. After bacterial growth, the biomass was removed with the aid of bacteriological loops and suspended in 0.89% NaCl sterile solution, homogenizing the bacterial suspensions until turbidity equivalent to 0.5 McFarland standard scale (equivalent to concentration of 1.5x108 CFU/ml). The volume of 30 μl of the bacterial suspension was inoculated on 30 ml of BHI to give a bacterial concentration of 1-2x105 CFU/ml. In order to determine the MIC, the initial inoculum was 1-2x105 CFU/ml, and the concentrations of the extracts varied from 1600-3.1 μg.ml-1. The MIC was defined as the lowest concentration inhibiting bacterial growth (without visible growth) [55]. As a determining factor of MBC, the surface of the agar BHI (Brain Heart Infusion) was seeded with the samples which did not indicate the presence of visible bacterial growth. MBC was defined as the lowest concentration which did not allow any visible bacterial growth in agar [55]. The assays were realized in triplicate for each concentration of the extracts tested.

In vitro activity of the ethanolic extracts (EtOH) on the strains of tumoral cells B16F10
The cellular strain of murine melanoma, B16F10 (ATCC1 CRL-6475™) was kept in a culture medium RPMI 1640 (Gibco1, Life Technologies, Carlsbad, CA, USA), complete with 10% foetal bovine serum (FBS) (Gibco1) and 1% antibiotic solution of penicillin/streptomycin. It was placed in an incubator at 37°C in environment an of 5% CO 2 . The in vitro tests of the ethanolic extracts (EtOH) performed on these cells followed the procedure: a solution of 1% Trypsin-EDTA was used to detach the culture of confluent cells from the growth bottle; after 5 minutes the solution of trypsin was inactivated by FBS and after adding the culture medium the material was centrifuged for 10 minutes at 1500 RPM; after discarding the supernatant, the pellet of cells was again suspended in a complete RPMI medium and then the inoculum of cells corresponding to the final concentration of 1x10 5 cells/ml was calculated. After distributing the culture medium in plaques of 24 wells, the compounds were added to two concentrations of 50 μg/ml and 100 μg/ml, and the DMSO was used as a diluting control and statistic parameter [56]. The cellular proliferation was determined after 24 and 48 hours of incubation using the method of colorimetric assay adapted from Busatti and Gomes [57], and reading by an ELISA (enzyme-linked immunosorbent assay) reader at 570 nm. The cells used in this study were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA).

Statistical analysis
The results were expressed in the form of mean ± standard deviation (n = 3). For the statistical analysis of the results, the programme Statistica1 6.0 from StatSoft (Tulsa, USA) was used. Variance analysis (ANOVA) and the Tukey test were used to identify significant differences among the means (p>0.05).

Characterization of raw propolis samples
Propolis is a complex mixture, containing resins, balsamic products, wax, essential oils, pollen, and microelements, besides other components. The samples presented a characteristic aroma, balsamic and/or resinous, malleable (red) to rigid (brown and green) consistencies at room temperature, with a very heterogeneous granulometry. It is noted that the analysis of the physical-chemical composition is of great importance to determine the quality of the studied material, considering the incorporation of this matrix in food products [58][59][60]. On Table 2, the results of the physical-chemical characterization of the different samples of propolis are shown. On Table 3 are the results for the content of certain minerals present in the samples.
The determination of the total ash content is particularly important in samples of propolis commercialized in powder form, as this analysis can identify a possible adulteration of the material through the presence of impurities, or even residues from previously extracted propolis [6]. The samples were within the limit established by the Brazilian legislation (maximum 5%) [61]. Among the microelements analysed and identified in the samples, we can highlight the high contents of potassium on the three samples of green propolis (Table 3). Some studies show aluminium, vanadium, iron, calcium, silicon, manganese, strontium and potassium, as the main microelements present in propolis samples [5,49,[62][63][64]. Performing a comparative evaluation of samples of the same type (in relation to colour), the lower variation identified for the physical-chemical parameters studied was for the red propolis (RSE and RAL). There are a few works reporting the total physical-chemical characterization of propolis, however the values obtained in this study for certain parameters are similar to those found in the literature [24,[64][65][66][67][68]. The variation identified among the samples studied and with those from other studies is easily explained by the type of propolis, flora of the region and period of collection.
On No study evaluating the Brazilian propolis by SEM was identified. On the visual and microscopic analysis no strange substances were identified. In all images, it is possible to observe rugged surfaces covered by layers of wax and extractives. Similar characteristics were identified by Tylkowski et al., [69] for samples of propolis from Bulgaria. It is also important to note that there were similarities identified on the microscopic appearance (profile) of the samples of the same type (colour). For example, in all samples of green propolis, vegetable constituents were found, probably tector and/or glandular trichome and resinous substances from the vegetative apices of Baccharis dracunculifolia (Fig 2C, 2D and 2E) [7,[70][71][72]. Elements which are similar to vegetable parts were also identified on the samples of brown propolis, which probably come from the flora visited by bees, such as species of Copaifera (Fig 2F, 2G and 2H) [73].
Determination of content for phenolic compounds, flavonoids and antioxidant activity of EtOH and SCO 2 activity Table 4 shows the results for the content of total phenolic compounds, flavonoids and antioxidant activity of the extracts from different samples of propolis obtained by the two extraction methods (conventional ethanolic-EtOH and supercritical-SCO 2 ). The content of phenolic compounds varied from 97.97±0.01 (BPR SCO 2 ) to 300.36±0.01 mg EAG/g (RSE EtOH), whereas the content of flavonoids varied from 11.55±0.01 (BPR SCO 2 ) to 58.19±0.01 mg EQ/g (RAL EtOH) among other samples (Fig 3). There is great controversy in relation to the content of flavonoids present in samples of Brazilian propolis, in which phenolic acids are generally a lot more abundant. The antioxidant activity varied from 373.53±0.15 (BPR SCO 2 ) to 31.80 ±0.16 (GMG 1 EtOH) for DPPH (IC 50 ) and from 49.60±4.10 (BPR SCO 2 ) to 98.50±1.40 (RSE EtOH) for ABTS (Trolox 1 mg.ml -1 ) (Fig 4).
It was verified that the extracts obtained from red propolis (RSE and RAL) originated from the Brazilian northeast showed the highest content of phenolic compounds and flavonoids. This confirms the biological potential of this "new" type of propolis [4,[74][75][76]. The highest quantity of total phenols, flavonoids and the best antioxidant activity by ABTS was identified in the extract of red propolis from the state of Sergipe (SER EtOH), with values of 300.36±0.01 mg EAG/g, 57.60±0.01 mg EQ/g and 98.50±1.40%, respectively (Table 4). However, the best results for the antioxidant activity by DPPH was shown on the extract of green propolis from the state of Minas Gerais-GMG 1 EtOH (IC 50 of 31.80±0.16).
These results indicate that the total concentration of phenolic compounds or flavonoids is not the only factor responsible for antioxidant properties. The chemical nature of the phenolic compounds and, perhaps, the presence of other compounds contribute to the total antioxidant capacity of the extracts [77]. The extracts EtOH and SCO 2 obtained from brown propolis showed the lowest values of phenols, flavonoids and antioxidant activity, therefore showing the lowest biological potential of this type of propolis when compared to the samples of green and/ or red Brazilian propolis evaluated in the study.    (Table 4), when compared to the extraction method for the same sample, as well as for the extracts obtained through the same method and samples of different types. The variations identified among the samples (p>0.05) were already expected, considering that the propolis obtained from different phytogeography regions exhibit very distinct chemical profiles [80][81]. The results found in this study confirm the influence of the origin of the raw material on the composition and characteristics of the extracts.
Chaillou and Nazareno [82] also observed significant (p>0.05) differences on the content of phenolic compounds, flavonoids and antioxidant activity, when evaluating different samples of propolis originated from Santiago del Estero, Argentina, where the averages varied from 92 to 187 mg/g for polyphenols, from 6 to 18 mg/g for flavonoids, and from 49.5 to 65.7% for DPPH. Similar results were also identified by Kumazawa et al. [83], by Kalogeropoulos et al. [84] and Choi et al. [85], when propolis from different geographic origins were evaluated. It stands out that the variability found between the content of compounds and antioxidant activities of propolis from Brazil is attributed to the differences observed in the arboreal species for each geographic area, being justified by the great Brazilian diversity.
Comparatively evaluating the results obtained for the extracts EtOH and SCO 2 of the same sample, it is noted that the majority of extracts EtOH present the best results for the content of total phenolic compounds and flavonoids, with the exception of the samples BSC and BRS, where the extracts SCO 2 shows superior values of these compounds. Generally, the extracts obtained by ethanolic extraction (EtOH) show the best antioxidant activities through the methods DPPH and ABTS.
Similar results were observed by Miguel et al. [86] and Cottica et al. [77], who found higher values of total phenols and flavonoids in EtOH extracts of Portuguese and Canadian propolis, respectively, in relation to the aqueous extracts. Zordi et al. [87] also identified higher Chemical Composition and Biological Activity of Propolis Extracts concentrations in EtOH extracts of Italian propolis when compared to the extracts obtained by SFE in different conditions of process and using CO 2 . Lee et al. [1] also verified that the extracts of Brazilian propolis obtained by SFE showed a lower antioxidant capacity (DPPH) when compared to the extracts obtained by Soxhlet, hot extraction and by ultrasound. However, Laskar et al. [88] reported that the phenolic compounds in aqueous extracts of Indian propolis were in higher concentration when compared to EtOH extracts.
The lower concentrations of phenolic compounds, flavonoids and antioxidant activity identified in the extracts obtained by SFE (SCO 2 ) confirms the higher presence of undesirable substances such as wax, resin and other materials present in propolis, which could indicate a lower biological potential for these extracts [24]. Wax and other organic detritus are removed during the process of ethanolic extraction, and the propolis extracts obtained that way can contain the majority of the antioxidant constituents [84].
Zordi et al., [87] indicated two possible applications for the use of supercritical CO 2 for samples of propolis: obtain lipophilic fractions enriched by specific components, or as a pre-treatment of the raw material to facilitate the additional extraction with ethanol. As described by Biscaia and Ferreira [24], complex natural matrices such as propolis, can result in different products, depending on the method used. Therefore, the viability of the process is related to the yield and quality of the product, in order to improve the biological potential present in the raw material. With that, the efficacy of the extraction method (higher or lower selectivity) for obtaining total phenolic compounds and flavonoids can vary according to the origin and composition of the raw material.

Quantification of Artepillin C and p-coumaric acid in EtOH and SCO 2 extracts
The results of the quantitative analysis of Artepillin C and p-coumaric acid on the extracts of the different samples of Brazilian propolis are shown on Table 5. As expected, the markers Artepillin C and p-coumaric acid were present in all extracts of green propolis [1,7,10,[89][90][91]. The Fig 5 shows the chromatogram of a green propolis samples obtained by ethanolic extraction (GPR EtOH). Artepillin C was also identified in two samples of brown propolis originated from the regions of Santa Catarina (SC) and Paraná (PR). As described by Lee et all. [1], different types of propolis can contain a varied quantity of Artepillin C, however the green propolis, originated from the vegetal species Baccharis dracunculifolia shows a higher quantity of this compound.
The content of p-coumaric acid varied from 5.05±0.10 (BSC SCO 2 ) to 198.09±3.12 μg/ml (GPR SCO 2 ), whereas that of the Artepillin C varied from 58.32±1.00 (BSC EtOH) to 845.05 ±0.12 μg/ml (GPR EtOH) among the extracts. The results show differences among the samples (p>0.05), which are in conformity with its place of origin. The green propolis from Paraná (PR) was the one which presented the highest values of the studied compounds. Kumazawa et al. [83] identified the presence of p-coumaric acid and Artepillin C (43.9 mg/g) in propolis from Brazil when evaluating samples from different countries.
It was also identified that, from the 16 samples analysed, the Artepillin C was only present on the green propolis from Brazil. Tazawa et al. [92] also concluded that the p-coumaric acid and Artepillin C are the main active components of the Brazilian propolis, whereas certain flavonoids are the main constituents of propolis from other countries (China, Japan and Bulgaria, among others). Shimizu et al. [93] identified a high quantity of Artepillin C (21.0 mmol/100g) and p-coumaric acid (7.70 mmol/100g) in the Brazilian propolis from Minas Gerais. Ahn et al. [94] evaluated samples of propolis collected from different regions of China and observed the presence of p-coumaric acid in all samples, which varied from 2.3 to 42.3 mg/g. It becomes important to note that the extraction with supercritical CO 2 (SCO 2 ) was significantly more efficient for obtaining both analysed markers, when compared to the same sample (p>0.05). The extracts obtained by SFE can have a concentration four times higher than that Table 5. Determination of the content of 4-hidroxicinamic acid (p-coumaric acid) and 3,5-diprenil-4-hidroxicinamic (Artepillin C) of extracts from different samples of Brazilian samples obtained by ethanolic extraction (EtOH) and by SFE (SCO 2 ).

Samples
p-coumaric acid (μg/ml) Artepillin C (μg/ml)  of the p-coumaric acid, when compared to the EtOH extracts. Generally, despite this not being the most efficient method for obtaining the total phenolic compounds and antioxidant activity of the samples, it is clear that it provides a higher selectivity to obtain both analysed compounds, Artepillin C and p-coumaric acid in samples of Brazilian green propolis. It is important to highlight that one of the main aspects that should be considered in SFE is the choice of operational conditions in the process of extraction, because that it can provide an additional advantage to the conventional methods, considering the lowest total yield and the highest cost. Besides that, the use of the optimized values for the different conditions can significantly improve the yield and the recovery of the target compound [30].
In this work, the optimized conditions of temperature, pressure, percentage of co-solvent and quantity of sample of a previous work from our group [43], which specifically evaluated the obtainment of Artepillin C and p-coumaric acid. In view of that, a better extraction of both phenolic acids by SFE is justified, when compared to EtOH extraction.
Similar results were identified by other authors, when the extraction of relevant compounds from different natural matrices using SFE and conventional extraction were compared. Those studies have shown the selectivity of SFE, generating products with higher biological value [31,[95][96][97][98][99].
For example, Lee et al. [1] individually investigated organic solvents (conventional methods) and supercritical CO 2 to recover Artepillin C from the Brazilian propolis, and identified that the extracts obtained by SFE (45.3±0.10 mg/ml) showed the highest contents of the relevant compound (Soxhlet = 16.9±0.2 0mg/ml; Hot extraction = 16.4±0.23 mg/ml; Ultrasound = 16.0 ±0.06 mg/ml). Sun et al. [100] extracted the active substance paeonol from Cynanchum paniculatum by SFE and other conventional techniques, identifying 72.02% of the active substance in the extracts obtained by SFE, a highly superior result to that found by other methods of extraction (ultrasound 1.56%, steam distillation 1.64% and Soxhlet 2.74%).
The patent CN 1258511 (Chinese) and BR 1020140320121 (Brazilian) describe the extraction of active compounds from propolis by SFE. There, it is shown that the extracts obtained using CO 2 as supercritical fluid (and ethanol as co-solvent) are rich in different compounds (phenolic acids, flavones and terpenes) [26,101], being, therefore an efficient method for yield and selectivity for the extractive process of relevant compounds from propolis. Extracts of propolis obtained with supercritical fluids are already being sold in the markets of Japan, considering the proof of anti-tumour properties of these extracts [102].
Determination of the antimicrobial activity of the extracts EtOH and SCO 2 Table 6 shows the values of MIC and MBC obtained for the different SCO 2 and EtOH extracts of the propolis samples tested. It was noted that all extracts showed activity against gram-positive bacteria Staphylococcus aureus (ATCC 33951 and 25923) and gram-negative Escherichia coli (ATCC 25922), but this effect was dependent on the origin of the matrix and method of extraction. The control sample did not affect the growth of tested bacteria (data not shown). As expected, the extracts from different samples of propolis showed a higher activity against the gram-positive strains than against the gram-negative strains. These results are in accordance with those from Koru et al. [103], Vardar-Ünlü et al. [104], Kim and Chung [105] and Silva et al. [106], which can easily be explained by the structural differences of the bacterial cellular wall [107][108].
When compared to the extraction method, the EtOH extracts showed the best antimicrobial activities, and as previously shown, these extracts also had the best antioxidant activities and the highest content of total phenolic acids and flavonoids. Propolis samples from different regions of Europe and the Middle East were evaluated in a study performed by Popova et al. [109] and a negative correlation between the concentration of phenolics in the extract and MIC was identified. The test was performed with hydro alcoholic extracts of propolis against Staphylococcus aureus, and it was found that the higher concentration of phenolics, the more powerful the activity against this bacteria. Jug et al. [110] also evaluated the antibacterial and antifungal efficiency of propolis extracts obtained by different extraction techniques and identified that the EtOH extract showed the best antimicrobial potential.
The extract that showed the highest antimicrobial activity in vitro for the three tested strains was the RSE EtOH, which also showed the highest content of phenolic compounds and high values of flavonoids. Among the samples evaluated, the extracts obtained from the samples of red propolis showed the best antimicrobial activities. Koo et al. [55] evaluated extracts of propolis from different types and regions of Brazil (red propolis from Bahia and green propolis from Minas Gerais and Paraná), identified differences in the MIC and MBC for each extract in relation to Streptococcus mutans, S. sobrinus and S. cricetus, and also that the best results were shown by the red propolis from the Brazilian northeast, as identified in this study.
Alencar et al. [18] also identified a noticeable antimicrobial activity for ethanolic and chloroformic extracts of Brazilian red propolis (Alagoas) against the Staphylococcus aureus ATCC 25923 (MIC of 50-100 and MBC of 200-400 -EtOH extract; MIC of 200-400 and MBC of 100-200 -chloroformic extracts) and Staphylococcus mutans UA159. It were concluded that the best antimicrobial activity was found for the extract with the highest concentration of chloroformic total phenols. For the extracts from green propolis, the samples GMG 1 and GPR showed the best antimicrobial potential. The extracts of brown propolis, which showed the lowest antioxidant potentials, presented the highest concentrations for the inhibition of antimicrobial growth for the strains tested. As expected and identified in other studies [18,109,111], the MBC for all extracts was four times superior to the MIC. As already described in other studies, and also identified in this work, regardless of its geographic origin, propolis shows an important antimicrobial activity, since this property is essential for the preservation and maintenance of the hive [84,103,[112][113][114]. Different studies show that the antimicrobial activity of this matrix is mainly due to complex synergic effects between the flavonoids, phenolic acids and its derivatives, which are mainly present in propolis [39,[115][116].
For example, when the propolis extracts originated from Brazil and Bulgaria were evaluated against strains of Staphylococcus aureus, it was found that the Brazilian extracts showed the best antimicrobial potential, which had the highest concentration of phenolic compounds [117]. Although the mechanism of action for the antimicrobial effect of propolis is still not clearly understood and defined, some studies suggest that certain constituents can interfere in the process of bacterial cell division through disorganizing the cytoplasm, causing cellular lysis [118][119]. It was also found in this study, as previously reported by other authors, that the antimicrobial activity of propolis extracts is related to the method of extraction and type of solvent used [110,120]. Therefore, the determination of MIC and MBC is extremely important to evaluate the quality of the extracts and propolis-based products [103,105,108], considering the great variability in its composition.

Determination of antitumoral activity in vitro of EtOH extracts
The present study also investigated the antitumoral activity of the EtOH extracts of the eight samples of propolis against the cellular strains of melanoma murine (B16F10), evaluating the anti-proliferative effects. Generally, the ethanol extracts showed the best results for content of phenolic compounds, flavonoids and antioxidant activity (DPPH and ABTS). Because of that, those extracts were selected for analysis in vitro against human cancer cell lines B16F10. Moreover, the cost associated with these tests do not enables the evaluation of extracts obtained by supercritical fluid extraction. Therefore, comparative and statistical analysis was performed only between different samples of propolis, considering only an ethanol extraction method.
On Fig 6A and 6B are shown the activity on the cellular proliferation of the strain B16F10 after 24 and 48 hours of incubation on both concentrations tested (50 and 100 μg/ml). After 24 Chemical Composition and Biological Activity of Propolis Extracts and 48 hours of incubation, all extracts showed a significant inhibition of cellular proliferation in comparison to the control (p>0.05).
The best results were shown by the extracts derived from the red propolis from the northeast region of Brazil, that is, Sergipe (RSE EtOH) and Alagoas (RAL EtOH). After 48 hours of incubation, the extracts RSE EtOH with concentrations of 50 and 100 μg/ml showed the lowest contents of viable cells, showing absorbance values of 0.018±0.002 and 0.006±0.001 (1x10 5 cells/ml), respectively. Among the green propolis extracts studied, the sample originated from Paraná, in Brazil (GPR EtOH) showed the best results for proliferative inhibition of cells B16F10. This extract also showed the highest concentration of Artepillin C and p-coumaric acid ( Table 6).
The extracts showing the lower potential were those obtained from the brown propolis. The results found for antitumoral activity are in accordance to those previously identified for the content of antioxidant compounds (Table 5). Similar results to this study were found by Franchi-Jr et al. [121], when identifying that the in vitro cytotoxic activity of ethanolic extracts of red propolis against strains of human leukemic cells was superior when compared to the extracts of green propolis. Popolo et al. [122] also identified anti-proliferative effects of the ethanolic extract of the brown propolis from Cuba in cellular lineages of human breast cancer.
The best anti-proliferative effect showed by the red propolis extract, when compared to other samples of propolis may depend on its differentiated composition, for example, the presence of formononetin, the main isoflavones found in this type of propolis [123][124]. Recently, a polyisoprenylated benzophenone (xanthochymol) was also identified in the red propolis [123]. Different studies point to xanthochymol and formononetin as showing activity against tumoral cells [125][126][127][128].
It is likely that the anti-proliferative activity of this type of propolis occurs through the mechanism of cellular circle halt and apoptosis, as indicated in some studies [121,[129][130]. Novak et al. [56] identified that an active fraction of the ethanolic extract of a sample of the Brazilian red propolis (João Pessoa) containing xanthochymol and formononetin showed superior anti-proliferative effects in strains of B16F10 cells, when compared only with the ethanolic extract of the sample.
In previous studies with the red propolis of the same geographic origin (Sergipe-RSE) different antitumoral effects against strains of bladder cancer cells and the presence of formononetin in its composition have been reported [76], and on larynx, uterus and kidney cancer cells [32]. López et al. [125] evaluated samples of red propolis from different regions and identified the presence of formononetin in every sample analysed, among which, four samples were from the state of Sergipe and two from Alagoas (Brazil).
Despite the fact that the green propolis, especially the GPR EtOH, show a lower potential when compared to the extract of red propolis (RSE and RAL), the inhibition in the presence of B16F10 cells was also very significant and relevant (Fig 5), especially when compared to the control (p>0.05). The presence of Artepillin C, a substance found in the Brazilian propolis, is the main constituent of the extracts of green propolis, followed by p-coumaric acid [83,94,131] and also attested by this work.
Carvalho et al. [132] also identified the anti-tumoral effects in vitro of the ethanolic and oily extracts of the green propolis from Paraná (Brazil), against different strains of human tumoral cells: HL-60 (leukaemia), HCT-8 (colon), MDA/MB-435 (breast) and SF-295 (brain). Kimoto et al. [133] identified that Artepillin C showed powerful cytocidal effects and induced levels of apoptosis in all the cellular lines of human leukaemia of different phenotypes evaluated. Kimoto et al. [89] identified that Artepillin C from Brazilian propolis showed cytotoxic effects and inhibited the growth of malign murine tumoral cells (B16F10) in vitro and in vivo, and that the mechanism of action is through the activation of the immunological system. Matsuno et al. [134] reported that the Artepillin C is one of the most effective anti-tumoral compounds in green propolis, and Kimoto et al. [135] showed that this compound is capable of reducing the tumoral load in certain animal models. Akao et al [136] identified that the antitumoral inhibitory effects of the p-coumaric acid were less powerful than those of Artepillin C, and that the compounds available induced apoptosis in the cells characterized by nucleosome and DNA fragmentation analysis. Due to these biological properties, the propolis containing Artepillin C is considered a high quality propolis and the concentration of this component is already been used for the quality control in certain companies [67,132]. From the results identified, and together with other studies, the Brazilian propolis, especially the red and green, can be considered as an important source of active compounds for the development of new drugs with anti-tumoral potential.

Conclusions
In this study, it was determined that the total quantity of phenolic compounds, flavonoids and antioxidant activity are important parameters to evaluate the quality and biological potential of extracts from the Brazilian propolis, especially considering the great Brazilian biodiversity. The results identified significant differences among the samples (p>0.05), which are in conformity with their place of origin. Despite this chemical diversity, all the types of propolis showed a significant antimicrobial activity, and in most cases, it can be assumed that the compounds responsible are the phenolic constituents of propolis. In relation to the extraction method, generally, the ethanolic extraction (EtOH) was the most efficient for obtaining extracts with the highest content of antioxidant compounds and biological activity. However, the extraction with supercritical CO 2 (SCO 2 ) was the most efficient for obtaining Artepillin C and p-coumaric acid, evidencing the higher selectivity of SFE for obtaining both important markers for the Brazilian green propolis.
Finally, it must be noted that, due to its valuable properties and high biological potential, also already evidenced in other studies, the propolis can be considered as an important source of natural antioxidant compounds. New studies about the propolis complete chemical composition are under way.