Propolis from Southwest of Parana produced by selected bees : In uence of seasonality and food supplementation on antioxidant activity and phenolic pro le

Propolis produced by selected bees Apis mellifera were collected from March to June of 2013 and in March of 2015 and analyzed in order to evaluate the in uence of climate, colony of origin, and food supplementation of colonies on the content of total phenolic and avonoid by chromatographic analysis and antioxidant activity by radical scavenging of 2,2-diphenyl-1-picrylhydrazyl hydrate (DPPH) and 2,2’-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) (ABTS) and ferric reducing antioxidant power (FRAP) methods. The Principal Component Analysis (PCA) was carried out with propolis collected in 2013 and two clusters were formed. Propolis produced in the months of March and April showed a higher content of total phenolic compounds (TPC) and antioxidant capacity than those produced in May and June. The results of PCA obtained from samples collected in March of 2013 and 2015 showed two clusters, and propolis collected in 2015 were more bioactive and presented a higher content of TPC. The chromatographic analysis of extracts allowed the identi cation of phenolic acids p-coumaric, ferulic and ca eic with similar chemical pro les that could be closely related to the botanical origin of propolis. It can be concluded that the season and food supplementation of colonies in uenced the chemical composition and the biological activity of samples analysed.

Phenolic compounds being their main components, and they have a high redox potential, which allows them to work as reducing agents MATHEUS A. CALEGARI et al. (Miguel et al. 2010).The antioxidants act as defense agents against free radicals that are naturally present in organisms.Studies indicate that these compounds protect the organism from chronic diseases caused by oxidative stress, such as cancer and metabolic disorders (Wang et al. 2014, De-Melo et al. 2014, Kumazawa et al. 2004, Mohammadzadeh et al. 2007).
Studies described in the literature show that the chemical composition of propolis is varied and complex and it is intimately related to the biodiversity of flora, geographical region and climate where it is collected (Buratti et al. 2007, Gong et al. 2012, López et al. 2014).It is generally composed of resin, vegetable balm, wax, essential oils, pollen, and several other substances (Kalogeropoulos et al. 2009, Luis-Villaroya et al. 2015, Valenzuela-Barra et al. 2015).
Analysis of TPC and total flavonoids content (TFC) and the determination of antioxidant activity, contribute to the determination of both chemical and biological composition of propolis, allowing a relationship to be established with its geographical origin (Falcão et al. 2013).Nevertheless, a detailed discussion could be promoted if chemometric tools were used.Among them, PCA receives attention because it allows the evaluation of dataset, establishing relationships between variables, reducing its dimensions, and maintaining most of the statistical information (Ranđelovic et al. 2015).
Therefore, the goals of this study were to determine the influence of climate, origin of colony, and food supplementation of colonies in the phenolic profile by High-Performance Liquid Chromatography with Diode-Array Detection (HPLC-DAD), TPC and TFC and the antioxidant activity in Brazilian propolis produced by selected bees (by means of the animal model approach for weight at the emergence of Africanized queens) collected from March to June of 2013 and March of 2015, in addition to classifying these samples using the parameters evaluated, and the PCA tool.

SAMPLE COLLECTION
The samples of propolis were acquired from selected colonies of Africanized honey bees (Hymenoptera: Apidae) collected at an apiary within the Unidade de Ensino e Pesquisa de Apicultura (UNEPE) in Federal Technological University of Parana, Dois Vizinhos, Parana, Brazil (latitude: -25,699063°, longitude: -53,095273°, altitude: 546m).The collections were performed between March and June of 2013 and March of 2015.The selection of Africanized bees was done through generations and was based on the weight of queen at emergence.In 2013 the colonies were under systematic selection and in 2015 under relaxed selection.In 2015 the colonies received food supplementation every three days throughout the year.Samples were collected of propolis produced by fifth generation bees and obtained from three distinct colonies.The selection of the colonies used in this study was based on phenotypic reference to the potential of the bees.
The identification used in this study is described in Table I relates to the variables: month of collection, year, color, and colony.

SAMPLE PREPARATION
After collection, the samples were cleaned, crushed with liquid nitrogen, homogenized, weighed, and stored at -6 °C.

PREPARATION OF ETHANOLIC ExTRACT OF PROPOLIS (EEP)
The EEP was prepared as described by Oldoni et al. (2015).Samples (2 g) of propolis were adding 25 mL of ethanol:water 80:20 (v v -1 ) was added.The total volume was transferred to a thermostatically controlled water bath at 70 ºC and held for 45 min.The samples were cooled and filtered using qualitative filter paper.All extracts and analyses were performed in triplicate.

ANTIOxIDANT ACTIVITy USING THE ABTS METHOD
The antioxidant activity was performed through the ABTS method according to Re et al. (1999) where the ABTS •+ radical was formed by the reaction of 7 mmol L -1 of ABTS with potassium persulfate 140 mmol L -1 , incubated at 25 °C in the dark for 16 hours.The radical was diluted with ethanol and resulted in an absorbance value of 0.700 ± 0.200 at 734 nm.Under dark conditions, 3.0 mL of the ABTS •+ solution was added to 30 µL of EEP (16 g L -1 ) and the absorbance was read at 734 nm at a spectrophotometer after 6 min.Ethanol was used as a blank and Trolox was used as reference in concentrations that varied from 100 to 2000 μmol L -1 .The results of antioxidant activity were equivalent to the antioxidant capacity of Trolox per gram of sample (μmol Trolox g -1 ).

ANTIOxIDANT ACTIVITy USING THE DPPH FREE RADICAL SCAVENGING METHOD
The antioxidant activity by the free radical scavenging method DPPH was performed as described by Brand-Williams et al. (1995) where the reaction medium consisted in 0.5 mL of EEP (1.6 g L -1 ), 3 mL of ethanol and 0.3 mL of radical DPPH solution 0.5 mmol L -1 in ethanol.The mixture was incubated without light at room temperature for 45 min.and, later, the absorbance was measured using a spectrophotometer (model UV-VIS Lambda 25, Perkin Elmer) at 517 nm.The ethanol was used as blank and the quantification was performed based on an analytical curve using Trolox as standard in concentrations that varied from 15 to 200 μmol L -1 .
The results were expressed in μmol of Trolox per gram of propolis (μmol Trolox g -1 ).
ANTIOxIDANT ACTIVITy USING THE FRAP METHOD The antioxidant activity by the FRAP was proposed by Benzie and Strain (1996).The FRAP reagent was obtained from a mixture of 25 mL of acetate buffer 0.3 mol L -1 , 2.5 mL of a solution of TPTZ, 10 mmol L -1 and 2.5 mL of iron chloride in an aqueous solution at 20 mmol L -1 , used after preparation.
The method consisted of a mixture of 100 µL of EEP (1.6 g L -1 ) with 3 mL of FRAP reagent.The mixture was homogenized and kept in a water bath at 37 °C for 30 min.and, next, the absorbance in 595 nm was determined by a spectrophotometer (model UV-VIS Lambda 25, Perkin Elmer).The FRAP reagent was used as a blank and an aqueous solution of ferrous sulphate (100 -2000 μmol L -1 ) was used for the calibration curve.The results were expressed as μmol of Iron 2+ ion per gram of propolis (μmol Fe 2+ g -1 ).

TOTAL PHENOLIC CONTENT
The TPC was determined through colorimetric analysis using the Folin-Ciocalteau reagent (Singleton et al. 1998).In a test tube, 0.5 mL of EEP (3.2 g L -1 ) was added, 2.5 mL of Folin-Ciocalteau reagent diluted in 10 g L -1 and 2.0 mL of Na 2 CO 3 at 40 g L -1 .After incubation without light for a period of 2 hours at room temperature, the absorbance was measured in a spectrophotometer (model UV-VIS Lambda 25, Perkin Elmer) at 740 nm.The blank was performed using ethanol:water 80:20 (v v -1 ).
The TPC of the extracts was determined by comparison with a standard calibration curve of gallic acid and represented in equivalent mg of gallic acid (EAG) per gram of propolis (mg EAG g -1 ).

TOTAL FLAVONOID CONTENT
The TFC present in EEP was determined by applying the colorimetric method described by Jurd and Geissman ( 1956).An aliquot of 0.5 mL of EEP (16 g L -1 ) was added to a series of tubes identified as with or without nitrate.In the tubes that received nitrate, 4.3 mL of ethanol:water 80:20 (v v -1 ) were added, and in those tubes that had not received nitrate, 4.4 mL of ethanol:water 80:20 (v v -1 ) was added.In all tubes, 0.1 mL of potassium acetate 1 mol L -1 was added.In the series of tubes identified as "with nitrate", 0.1 mL of aluminum nitrate at 10% was added.A control (blank) was also prepared from 4.9 mL of ethanol at 80% and 0.1 mL of potassium acetate.After 40 min., the absorbance was measured in a spectrophotometer (model UV-VIS Lambda 25, Perkin Elmer) at 415 nm.The TFC was determined by comparison with the calibration curve using quercetin as standard, and the results were expressed in mg of equivalent to quercetin (EQ) per gram of propolis (mg EQ g -1 ).

RESULTS AND DISCUSSION
Propolis produced in March and April of 2013 showed a difference in color (Table I) and higher results for antioxidant activity and content of TPC when compared with those produced in May and June (Table II) of the same year, indicating that there is a variation in chemical composition of the material and this variation can be related to the month of production.Among the samples of propolis collected in March of 2013 and 2015 the best results for TPC and TFC content as well as antioxidant activity (  11.8 and 235.6 µmol of Trolox g -1 , respectively (Table II).Kalogeropoulos et al. (2009) evaluated the antioxidant activity of ethanolic propolis extracts originating from Greece and Cyprus through the DPPH radical scavenging method and obtained as response values between 0.33 and 1.11 mmol Trolox g -1 , respectively.Sample D1 produced in 2015 was the most successful in reducing the complex Fe 3+ to Fe 2+ , 1041.0 µmol Fe 2+ g -1 , and presented high TFC (3.03 mg EQ g -1 ) (Table II).However, sample C4, produced in 2013, showed the lowest antioxidant activity through the FRAP method and the lowest TFC, 89.7 µmol of Fe 2+ g -1 and 0.16 mg EQ g -1 , respectively.
A high variability of results was also observed in propolis originating from Maringá -PR, which showed variation from 528 to 2068 μmol Fe 2+ g -1 of propolis for antioxidant activity through the FRAP method, and from 10 to 26 mg EQ g -1 for the TFC (Cottica et al. 2011).In propolis produced in Transylvania, values between 740 and 2540 µmol Fe 2+ g -1 were determined (Mihai et al. 2011).
The lowest and highest contents of TPC were observed for C4 and D3 samples, with 5.36 and 48.2 mg EAG g -1 , respectively.Propolis collected in Anhui, China, showed values for TPC that varied from 174.7to 235.6 μg EAG mg -1 (yang et al. 2011).Matos et al. ( 2014) evaluated the content of TPC in propolis collected in the Northeast of Brazil and obtained values values between 7.68 to 36.78 ± 1.52 mg EAG g -1 , corroborating the results of this study.
It can be concluded that there is a signifi cant diff erence among propolis produced in diff erent months and years by fi fth generation bees.It is evident that propolis produced in 2015 is rich in phenolic compounds which should be directly related to higher antioxidant activity.
PCA is a technique used in multivariate analysis, allowing eased visualization of all information contained in a data set and classifi cation of samples according to similarity, determination of objects showing diff erent properties from others (outliers), and defi nition of important variables for classifi cation (Morlock et al. 2014).In the current study, the fi rst PCA (Figure 1) was performed using results obtained from samples collected in the summer and fall of 2013.Two principal components (PCs) described 92% of the total data variability.PC1 and PC2 described 86.2 and 5.8% of the variability respectively.
There were two general clusters on the 2D PC score (Figure 1a), one group corresponding to samples collected in March and April, with the exception of sample A3, and the other group was composed of samples collected in May and June, with the exception of sample B1.All parameters evaluated had a positive correlation with PC1 and their concentration in sample collected in March and April was significantly higher.According to Figure 1b, PC2 showed a negative correlation with TPC and FRAP while DPPH, ABTS and TFC showed a positive correlation.
Nevertheless, results showed that is impossible to group the samples in relation to the colony, which indicates that the physical-chemical characteristics of the propolis cannot be associated with these classes.
The variations in the results for samples collected from March to June 2013 can be explained by variations in temperature observed in this period (Table III).Depending on the yield in the production of propolis in UTFPR -campus Dois Vizinhos, samples were collected between summer and fall, as well as in the transitional period between seasons.In southern Brazil there is a gradual decrease in temperature during this period and the change in average temperature directly influences the activity of bees, that is, the lower the temperature, the lower the activity of the bees will be.According to the data obtained by Jiang et al. (2016) activity of bees becomes more energetic when the temperature exceeds 25 °C and relative humidity is between 60% and 70%.It was also observed that the bees leave their hives at sunrise and cease activities at sunset.
In 2013 changes were observed in the average temperatures of 5 °C and 6 °C for the minimum and maximum respectively between the months of March to June, with an average maximum above 25 °C in March and April.The propolis samples collected in the months with the lowest average temperature showed the lowest levels for TPC and TFC, as well as for the antioxidant activity.It is possible that due to the lower activity presented by bees and lower availability of material collected, product quality was affected, and as a result, different biological activities were observed.
In the study by McCall and Primack (1992), the shape of the flower, light, and temperature of the year are the most important variables that influence the rate of visitation of insects to flowers.Malerbo-Souza and Silva (2011) observed in their results that Apis mellifera bees have a different foraging behavior (insects searching for food) during the year.The bees collected less nectar in July 2007 (winter) and more in February 2008 (summer), which correlates with our results.
A second PCA was performed with samples collected in March of 2013 and 2015.The main objective was to compare the influence of year of production and food supplementation in the quality of propolis.Two PCs described 96.3% of the total data variability.PC1 and PC2 described 75.8 and 20.5% of the variability respectively.Two general clusters on the 2D PC score (Figure 2a) were formed related to the samples of 2013 and 2015.
All parameters evaluated had a positive correlation with PC1 and their concentration in samples collected in 2015 was signifi cantly higher.The parameter fl avonoid is related to the D2 sample, while results obtained by the FRAP method were higher for D1 sample.PC2 showed a negative correlation with TPC, ABTS and DPPH while FRAP and TFC showed a positive correlation (Figure 2b).
These results indicate that, despite diff erent methods being used to evaluate the antioxidant activity, there is a strong positive correlation between content of TPC and antioxidant activity by DPPH and ABTS methods (Figure 2b), indicating that the antioxidant activity by these methods could be related to TPC content.
The phenolic profi les of EEP were identifi ed by using High Performance Liquid Chromatography and the samples A2, A4 and D1 (Figure 3) are representative of summer and fall of 2013 and 2015 respectively and presented a chromatographic profi le with diff erences in intensities of absorbance.The phenolic acids p-coumaric, ferulic and caff eic were identified and quantified (Table IV).The chromatographic method used to identify the TPC was eff ective, as it showed a high selectivity and resolution for the majority of signals analyzed.
Samples showed a complex chemical composition with multiple signals mainly eluting with retention times of between 20 and 30 minutes, which indicates compounds with chemical structures of medium polarity (Figure 3).The phenolic acids caffeic, p-coumaric and ferulic were identifi ed in the most extracts and from the identifi ed compounds, which showed the highest levels.These results correlate with those obtained by Oldoni et al. (2015).
It is important to note the similarity of chemical profi les that samples produced in 2013 and 2015 showed, indicating that the botanical origin of the material analyzed is the same.The main diff erence between the propolis extracts was the content of phenolic acids identifi ed in extracts (Table IV and  Figure 3).The extracts D1, D2 and D3, produced in 2015 showed a higher content of phenolic acids caffeic, ferulic and p-coumaric.The D3 extract showed the highest levels of p-coumaric acid (617.69 µg g -1 ) and ferulic acid (305.56 µg g -1 ) and sample D1 showed higher content of caff eic acid (150.84 µg g -1 ) indicating a positive correlation of these compounds with antioxidant activities determined by radical scavenger methods DPPH and ABTS (Table II).In 2013 and 2015 changes were observed in average temperatures of 1 to 3 °C for the minimum and maximum respectively.This variation may have had an impact on bee activity, as well as the daily supplementation provided in 2015.In addition, the temperature can influence the production of secondary metabolites in plants as a function of environmental variables, such as attack by microorganisms, temperature and UV radiation.Resins, flower buds and resinous exudates are sources of material for the production of propolis and a variation in the chemical composition of these materials means changes in the composition of propolis. De

CONCLUSIONS
The PCA results provided two clusters for propolis produced in 2013 and it was possible to separate samples produced in warmer months from those produced in colder months.Among the samples collected in March of 2013 and 2015, it was possible to verify more bioactivity in those collected in 2015 from colonies that received food supplementation.
In addition, a positive correlation was verified between the TPC and the antioxidant activity through the DPPH and ABTS radical scavenging.The chromatographic analysis showed a similar chemical profile for propolis produced in different months and years, with identification of phenolic acids p-coumaric, caffeic and ferulic.

Figure 1 -
Figure 1 -Clustering on a 2D PC-score (a) and loading plot (b) based on the variables ABTS, DPPH, FRAP, total contents of phenolic and fl avonoid compounds for propolis collected in 2013.

Figure 2 -
Figure 2 -Clustering on a 2D PC-score (a) and loading plot (b) based on the variables ABTS, DPPH, FRAP, total contents of phenolic and fl avonoid compounds for samples collected in March of 2013 and 2015.
Figueiredo et al. (2015) evaluated the chemical composition of propolis produced in Minas Gerais, Brazil by RP-HPLC and identified p-coumaric acid Li et al. (2016) obtained caffeic and ferulic acids in large quantities by HPLC in aqueous extract of propolis from China.

TABLE I Identification's table of samples of propolis. Code of samples Month of collection (2013)
*In the year of 2015, samples were collected only in March because was observed that propolis collected in March and April of 2013 showed better results and no significant difference.MATHEUS A. CALEGARI et al.
The PCA charts were developed based on the propolis characterization (ABTS, DPPH, FRAP, TFC and TPC).The pre-processing approach of the data used in this study was autoscaling, where each variable is centered in the average and divided by its standard deviation.Samples of propolis produced in the months from March to June 2013 were submitted to multivariate analysis, through the PCA to check the influence of seasonality in quality of material.After that, the results obtained from samples produced in March of 2013 and 2015 were compared with the main objective of verifying the influence of supplementation and year of production in the chemical composition of propolis.
MULTIVARIATE ANALySISMultivariate analysis was performed with Pirouette software.
Table II) were obtained for propolis collected in March of 2015 suggesting that food supplementation of colonies can improve the quality of material.The antioxidant activity through the ABTS radical scavenging varied from 25.5 to 439.2 µmol Trolox g -1 , corresponding to samples C4 and D3 respectively.Samples C4 and D3 also presented the lowest and highest capacity to scavenge the radical DPPH,

TABLE II Total content of phenolic and flavonoid compounds and antioxidant activity determined by the FRAP, ABTS and DPPH methods.
The same letter in the column, are not significantly different at the level of 0.05 according to ANOVA.The value is the average ± standard deviation (n = 4).RSD: Relative Standard Deviation.* equivalents to gallic acid.** equivalent to quercetin.An Acad Bras Cienc (2017) 89 (1) 50 MATHEUS A. CALEGARI et al.