Optimization of solvent extraction and HPLC-DAD method parameters for determination of phenolic compounds in various Brazilian propolis

Abstract Propolis is an organotherapeutic product collected by honey bees that contains a large number of phytoconstituents, possessing excellent pharmacological properties. In this study, phenolic composition and antioxidant capacity were evaluated in different types of Brown, Green and mixture of Red, Green I and Brown II (RGB) Brazilian propolis. Ultrasonic-assisted extraction was performed by direct solvent extraction (DSE) and successive solvent extraction (SSE) methods. The ethanolic extract of Green I propolis by SSE was higher in total phenolics (403.88 mg GAE/g extract), flavonoids (311.57 mg QE/g extract), and DPPH· scavenging activity (82.33 µMol TE/g extract) compared to other Brazilian propolis samples. Moreover, different solvent phases were used in the HPLC-DAD analysis and the results were validated by linearity and sensitivity. The optimized method was proven to be precise revealing good separation of standard phenolic compounds. Thermal stability of the compounds was accessed by using different column oven temperatures (40, 50 and 60 °C), and the results showed that lower temperature (40° C) retains the maximum contents of total phenolics in the 80% ethanolic extract of Green I propolis (253.71 mg/g extract) by DSE compared to other propolis samples. This extract (109.35 mg/g extract) also obtained a higher amount of total phenolic compounds in the SSE compared to ethyl acetate and acetone extracts. Among the phenolic compounds, artepillin C (60.10 mg/g extract) was higher in the SSE of Green I propolis 80% ethanolic extract. Thus, this study optimized the extraction and HPLC analytical conditions for the determination of phenolic composition and antioxidant capacity in different samples of Brazilian propolis.


Introduction
Propolis is a natural resinous substance prepared by bees from bud exudates of some plant sources (Hern andez Zarate et al., 2018).This resinous substance can be used by bees as a sealant to seal cracks in the walls of the hive, preserve moisture and prevent the decomposition of animals and insects killed by bees after intrusion into the hive (Pisarev et al., 2017).Brazil is considered to be one of the world's biggest suppliers of propolis in the world market (Pereira et al., 2002).Propolis can be classified based on the appearance of color, source of collection, and harvesting time.The propolis color may be creamish, yellow, green, light or dark brown.Some propolis samples differ with a friable, hard texture while other samples may be elastic and gummy.Park et al. (2002) classified the Brazilian propolis extracts into 12 groups based on the collections area and color.More recently, according to physico-chemical characteristics and geographic area, Machado et al. (2016) classified Brazilian propolis into 13 different types.The total phenolics contents may vary in different propolis samples obtained from different regions/hives and hence phenolic concentrations of these propolis samples need to be determined.
A total of about 300 phenolic compounds constituting of phenolic acids, flavonoids, flavonols, flavanols and isoflavones are common in Brazilian propolis, which have been considered as major chemo-preventive agents mainly due to their strong antioxidant activities (Gardana et al., 2007;Sawaya et al., 2011).Antioxidants can act as reducing agents, hydrogen donors, singlet oxygen quenchers and metal chelators that inhibit the polymerization chain initiated by free radicals and other subsequent oxidizing reactions which degenerate diseases (Rice-Evans & Miller, 1996).Many studies proved that the CONTACT Narendra Narain narendra.narain@gmail.comM Rajan: design the work, experimental HPLC data analysis, writing and editing the manuscript; TC Batista and CS Oliveira: performing various extractions, experimental analysis of results on different propolis samples; Daniel G de Oliveira: collection and supply of plants and over-all discussion on the topic; N Narain: Over-all design and supervision of the work, results analysis, correcting and finalization of the manuscript.
Supplemental data for this article is available online at https://doi.org/10.1080/00218839.2021.1996111mechanism of the pharmacological property of antioxindats from natural products may be attributed by the inhibition of reactive oxygen species in different body organs and systems such as kidney, liver, nervous and cardiovascular systems (Murugan et al., 2017;Shanmugam et al., 2020).There is a growing interest in safer and natural antioxidants for food applications, and a growing trend in consumer preferences towards natural antioxidants, all of which have given impetus to explore natural sources of antioxidants.Thus, there is a need to extract and estimate the phenolic compounds from natural products.
Phytochemical screening by HPLC is a simple, precise, cost-effective and principal analytical method used for the separation of bioactive compounds from complex mixtures in biological samples (Rajan et al., 2019).The bioactive compounds of phenolics and flavonoids in propolis have been detected by using different solvent phases in HPLC-DAD analyses (Bruschi et al., 2003;Croci et al., 2009;Hakimah et al., 2018;Kongkiatpaiboon et al., 2015;Kosalec et al., 2003;Nunes et al., 2012;Pellati et al., 2011;Pisarev et al., 2017;Regueira Neto et al., 2017;Sha et al., 2009;Tlak Gajger et al., 2017), (Table S1).The phenolic concentration and antioxidant activity of Red propolis were studied earlier by our co-workers (Andrade et al., 2017).However, there are many factors such as solvent type and column oven temperature which also influence the separation of bioactive compounds.Moreover, the inappropriate column oven temperature can degrade the compounds.Thus, there is a need to validate the appropriate solvent phase and the column oven temperature used in HPLC analysis for the detection of different classes of phenolic compounds.Optimization by using different solvents for the extraction of phenolic compounds could increase the validity and reliability of the HPLC method.Thus, the present study was aimed to investigate the optimization of solvent extraction, by using different solvents, solvent phase and column oven temperatures in the HPLC-DAD analysis to determine the phenolic compounds and their antioxidant capacity in various Brazilian propolis samples such as individual Green and Brown propolis samples and mixture of Red, Green and Brown propolis samples.

Collection of propolis samples
Different samples of raw propolis (Brown I, Brown II, Green I, Green II and Red) were collected from different apiaries located at Marechal Deodoro, Alagoas state, Brazil, at coordinates 9 45 0 34.454 00 S, 35 50 0 24.986 00 W during summer (dry) season (32 C) of January 2019.The propolis was obtained after the honey harvesting season, by scratching the hive walls and frames, following by the removal of debris of wood and bees.The samples of Brown I, Brown II, Green I, Green II and Red propolis were collected for a period of 7 days in the afternoon time at about 16 h.Later, 30 g of each Red, Green I and Brown II propolis samples were mixed based on the methodology described by da Silva et al. (2007).The collected samples were stored in a refrigerator (À20 C) and maintained at the Laboratory of Flavor and Chromatographic Analysis, Federal University of Sergipe, São Cristovão, Brazil.

Extraction of phenolic compounds from propolis
The extraction of phenolic compounds from various propolis samples was undertaken by two methods after using an initial ultrasonic-assisted extraction method.The equipment Ultrasonic Bath (USC-1400A, Unique, São Paulo, Brazil) with an ultrasonic frequency of 40 kHz, at 30 C was used.For the Direct Solvent Extraction (DSE) method, 1 g of propolis sample was extracted with petroleum ether.Later, the air-dried samples were extracted again with 80% ethanol (50 mL) in an ultra-sonication for 7 hrs.For the Successive Solvent Extraction (SSE) method, the propolis samples (1 g) were successively extracted based on the polarity index (PI) of the solvents (50 mL) using petroleum ether (PI-0.1),chloroform (PI-4.1),ethyl acetate (PI-4.3),acetone (PI-5.1)and 80% ethanol (PI-6.2) in an ultra-sonication.Extraction was realized with each solvent for 7 hr and each time before extracting with the next solvent, the samples were air-dried.
After completion of each extraction, the extracts were filtered with Whatman No.1 filter paper.Later, the filtrates were collected and concentrated in a rotary vacuum evaporator (Rotaevaporator Fisatom, Model 802, Brazil) and then air-dried in a preweighed beaker at room temperature for 12 hr.After drying, the extracts were weighed and the percentage yield was calculated.The dried extracts were dissolved in 15 mL of HPLC grade methanol.

Determination of total phenolics and flavonoids content
The total phenolics content was determined based on the method reported by Shetty et al. (1995).One mL of the extracts of different varieties of propolis samples was diluted with 1 mL of 95% ethanol solution in test tubes.Later, 5 mL of distilled water and 0.5 mL of 1 N Folin-Ciocalteu phenol reagent were added, followed by an addition of 1 mL of 5% sodium carbonate (Na 2 CO 3 ) solution and then the mixture was vortexed.The mixtures were kept in dark for 60 min at room temperature.The total phenolic contents were measured by using spectrophotometer (SpectraMax M2, molecular devices, Brazil) at 725 nm and the results were expressed in gallic acid equivalents (GAE).
The total flavonoid contents were determined using the method described by Meda et al. (2005).About 1 mL of propolis extracts was mixed with 4 mL of deionized water and 300 lL of 5% sodium nitrite for 5 min.Later, 300 lL of aluminum chloride (20 mg in 1 mL methanol) was added.The reaction mixture was allowed to stand for 30 min and then the absorbance was measured at 415 nm using a spectrophotometer (SpectraMax M2, molecular devices, Brazil).The results of total flavonoids contents were expressed as quercetin equivalents (QE).

Ferric-reducing antioxidant power (FRAP) assay
The ferric reducing ability of propolis samples was assayed by the FRAP method (Thaipong et al., 2006).The reaction mixtures of 10 mM TPTZ (in 40 mM HCl), 20 mM ferric chloride hexahydrate and 300 mM acetate buffer (3.10 g sodium acetate trihydrate and 16 mL acetic acid, pH 3.6) were mixed in 1:1:10 ratio, respectively.The propolis extracts (150 lL) were reacted with 2850 lL of FRAP and the solutions were incubated in dark for 30 min.The reading was measured using a spectrophotometer (SpectraMax M2, molecular devices, Brazil) at 593 nm.The calibration curve was plotted using trolox (20 to 800 lM), and the results were expressed as trolox equivalents (TE).

DPPH radical scavenging activity
For this assay, 1.0 mM (0.03943 g) DPPH was dissolved in 100 mL of ethanol, after which an aliquot of 2.9 mL was pipetted and placed in a test tube and 0.1 mL of previously diluted propolis sample was added.Later, the test tubes were kept in dark and the absorbance was measured at 517 nm after 30 min using spectrophotometer (SpectraMax M2, molecular devices, Brazil) against ethanol which served as blank.The analytical standard trolox was prepared at different concentrations of 1.0 to 2.5 mMol/mL to construct the calibration curve and the results were expressed as trolox equivalents (Kwon et al., 2006).

Instrumentation conditions
The chromatographic separation of standard phenolics and propolis samples were carried out on a UFLC (LC-20AD, Shimadzu Corporation, Japan) system equipped with a pump (20AT), a degasser (DGU-20A 3 ), an auto sampler (SIL-20AT) and a diode array detector (SPD-M20A).Chromatograms were recorded and evaluated by the "LC" Solution Software (version 1.24 SP2) of Shimadzu Technologies.The compounds were measured in the range of 190-800 nm and bandwidth of 4 nm using Diode Array Detector.The following different analytical conditions were validated to determine the phenolic compounds in propolis.
Method A. HPLC system connected with a Phenomenex Kinetex C18 column (250 cm Â 4.60 mm, 5 mm).The mobile phase consisted of 0.1% formic acid in water (phase A) and 0.1% formic acid in methanol (phase B).The gradient elution was performed at 0.0 to 15.0 min at 80% B concentration.The gradient elution program was linear at a flow rate of 0.6 mL/min.The column oven temperature was maintained at 40 C. The injection volume was 5 mL.The column was washed with methanol and reequilibrated for 15 min before analysis.
Method B. In this method, the HPLC system was connected with a Phenomenex Kinetex C18 column (250 cm Â 4.60 mm, 5 mm).The mobile phase consisted of 0.1% trifluoroacetic acid in water (phase A) and 100% methanol (phase B).The gradient elution was performed from 0.0 to 20.0 min at 85% B concentration.The elution program was linear at a flow rate of 0.6 mL/min.The injection volume was 5 mL.The column oven temperature was maintained at 40 C. The column was washed with methanol and re-equilibrated for 20 min before analysis.

Preparations of standards and samples
The standard phenolic compounds were prepared at 1:1 (w/v) ratio and diluted at different concentrations to construct the linear calibration curves.The crude propolis extracts were dissolved in HPLC grade methanol solvent at 1:5 ratio (w/v).The PVDF membrane filter of 0.45 mm was used to filter the standards/samples.

Linearity
The different concentrations of phenolic compounds were prepared to plot the calibration curves viz., concentration versus peak area for each analyte determined from the regression analysis using the least squares method.The slope, y-intercept and correlation coefficient values were obtained from the linearity equation (Y¼mxþc).

Sensitivity
The limit of detection (LOD) was defined as the lowest detectable concentration of signal-to-noise ratio of 3.3.Limit of quantification (LOQ) was determined at the lowest concentration of signal-to-noise ratio of 10.The values of LOD and LOQ were calculated according to the guidelines of International Conference on Harmonization using the following equations (ICH, 2005).
where r and s are the residual standard deviation of the regression line and slope of the calibration curve, respectively.

Statistical analyses
All the experiments were performed in triplicates and the results were expressed as mean ± standard deviation (SD) values.Analysis of variance (ANOVA) and significant difference between means, determined by Duncan's multiple range tests (p < 0.001).All calculations were performed using SPSS, version 20.0 (IBM SPSS, IBM Corp., Armonk, New York) software.

Results
Extract yield, total phenolics, flavonoids contents and antioxidant capacity In the DSE method, Brown I and RGB propolis samples, 80% ethanolic extracts showed higher total phenolics contents of 478.70 and 699.50 mg GAE/g extract, respectively when compared to other propolis samples.Moreover, in the SSE, 80% ethanolic extracts of all propolis samples showed higher total phenolics contents (214.13 to 403.88 mg GAE/g extract) compared to the samples of ethyl acetate and acetone extracts.
The flavonoid contents were found to be higher in the Green I propolis ethanolic extract (146.69 mg QE/g extract) in the DSE method.The ethyl acetate extracts of Brown I (269.74mg QE/g extract) and RGB (236.48 mg QE/g extract) propolis also showed higher flavonoid contents in the SSE method compared to acetone and 80% ethanol extracts.Comparing the results of various propolis samples, 80% ethanolic extract of Green I propolis by SSE showed higher total phenolics (403.88 mg GAE/g extract) and flavonoids (311.57mg QE/g extract) contents.
The DSE of defatted ethanolic extract of Green I propolis revealed increased antioxidant activities on FRAP (105.26 mg TE/g extract) and DPPH · (108.35mMol TE/g extract) scavenging activities, and this may be due to the higher flavonoids contents in them (Andrade et al., 2017).Furthermore, FRAP (46.34 mg TE/g extract) and DPPH · (93.76 mMol TE/g extract) scavenging activities were observed higher in the 80% ethanolic extract of Brown I propolis which was extracted by SSE method.

Validation of solvent phase in the HPLC-DAD
For the identification and quantification of individual phenolic compounds by HPLC-DAD, the different solvent phases were tested such as in Method A: 0.1% formic acid in water (phase A) and 0.1% formic acid in methanol (phase B), Method B: 0.1% trifluoroacetic acid in water (phase A) and 100% methanol (phase B) and Method C: 1% acetic acid in water (phase A) and 1% acetic acid in acetonitrile (phase B).The different classes of compounds were detected in methods A and B at different retention times at different wavelengths and the results are presented in Table S2.
From the data, it can be observed that in the method A, protocatechuic acid (t R ¼ 1.545 min), vanillic acid (t R ¼ 1.529 min), vanillin (t R ¼ Values are determined as mean (n ¼ 3) ±standard deviation.Statistically significant at p < 0.001 where a were detected in the range between 3 and 4 min.
In method B, most of the compounds were detected in the range between 3.765 to 6.633 min.Moreover, in the method, B, chlorogenic acid, p-coumaric acid, succinic acid, daidzein, kaempferol, caffeic acid, kaempferide, chrysin and artepillin C compounds were identified in all propolis samples (Figure S1b).
In method C, a total of 37 phenolic compounds were detected at varying times from 6.114 to 33.479 min in their respective spectra and these showed good partition between the compounds (Table S3).The linearity range was calculated within a given standard concentration versus peak area, which was directly proportional to the standard concentration.A good linear response was obtained for all phenolic compounds (R 2 !0.9833) in the ranges of standard concentrations analyzed in method C.These linearity equations were applied to quantify the concentrations of the phenolic compounds in different solvent extracts of propolis samples.From the data of detection limits of LOD and LOQ, the lowest value of LOD indicated that the validated method is sensitive since LOD values were three times lower than the LOQ values.

Validation of column oven temperature in the HPLC-DAD
The ethanolic extracts from the DSE method were analyzed in the HPLC-DAD at different column oven temperatures to study the effect of temperature on the phenolic compounds.Table 2 presents the data on the 23 identified phenolic compounds in the various propolis samples and their variations at oven temperatures of 40, 50 or 60 C.
The protocatechuic acid was not detected after increasing the column oven temperature at 50 and 60 C in all the propolis samples.However, gallic acid contents in the Brown I (0.05 mg/g extract), Brown II (0.06 mg/g extract), Green I (0.29 mg/g extract), Green II (0.03 mg/g extract) and in RGB (0.04 mg/g extract) was not affected while increasing the temperature.Moreover, at a column oven temperature of 50 C, the rutin compound in the Brown I (0.29 mg/g extract), Brown II (0.28 mg/g extract), Green I (1.62 mg/g extract), Green II (0.16 mg/g extract) and in RGB (0.22 mg/g extract) increased, compared to its contents when the column temperature was either 40 or 60 C. The ferulic acid content in the Brown I (2.46 to 3.76 mg/g extract), Brown II (2.30 to 2.78 mg/g extract), Green I (15.86 to 56.44 mg/g extract), Green II (1.75 to 6.44 mg/g extract) and in RGB (1.55 to 3.34 mg/g extract) increased when the temperature was increased from 40 to 60 C.

Validation of solvent extraction methods
To study the separation efficiency of phenolic compounds in the SSE method, the optimized analytical conditions of method C at column oven temperature of 40 C was used.Table S4 presents the data on phenolic contents variation in different solvent extracts of various propolis samples which resulted in the identification of about 31 phenolic compounds.Contrasting with the different propolis extracts obtained from SSE method, the higher total phenolic compounds concentrations were obtained in the 80% ethanolic extract which were in the order of Green I (109.35mg/g extract) > RGB (97.38 mg/g extract) > Brown II (95.94 mg/g extract) > Green II (33.66 mg/g extract) > Brown I (29.79 mg/g extract).
Comparing the different phenolic compounds obtained from different solvent extracts of SSE method, artepillin C compound was higher in the 80% ethanolic extract in all propolis samples of Brown I (15.05 mg/g extract), Brown II (44.05 mg/g extract), Green I (60.10 mg/g extract), Green II (10.62 mg/g extract) and RGB (42.33 mg/g extract).

Discussion
In the present study, the solvents usage for the detection of phenolic compounds in various types of Brazilian propolis samples was optimized.Earlier, most of the studies demonstrated that extraction of phenolic compounds in propolis samples was done by ultrasonic-assisted extraction method (Table S1).The ultrasonic-assisted extraction processes have been linked to cavitation, the rapid formation and collapse of air bubbles in ultrasound-treated fluids that produces local increase in pressure and large amounts of energy that ultimately could increase the diffusion rates across the cell wall or its breakdown (Chandrapala, 2015).This produces good extraction of phenolic compounds in lesser time than in conventional methods (Vega arroy et al., 2017).Canadian propolis (Garc ıa-Viguera et al., 1993), Chinese propolis (Cai et al., 2012;Sha et al., 2009), Croatian propolis (Tlak Gajger et al., 2017) and Brazilian Red, Brown and Green propolis (Andrade et al., 2017) samples were extracted through ultrasonic-assisted extraction method and obtained the maximum number of phenolic compounds.
In the present study, two different solvent extraction methods were performed in an ultra-sonication bath to study the extraction efficiency of bioactive compounds.The first attempt was made in this investigation to estimate the phenolic constituents in the defatted Brazilian propolis extracts.In the DSE method, the propolis extracts were dewaxed by petroleum ether solvent, whereas in the SSE method, the petroleum ether and chloroform were used to remove the fatty acids and some carotenoid contents.From the results of extract yield percentage, the polar solvents of ethyl acetate, acetone and 80% ethanol resulted in higher extract recovery than low polar solvents.Furthermore, high polar solvents such as ethyl acetate, acetone and 80% ethanol have the ability to extract more phenolic compounds than low polar solvents of petroleum ether and chloroform (Murugan et al., 2012(Murugan et al., , 2016)).Thus, the high polar solvent extracts of various propolis from each method were used for further HPLC-DAD analysis.The present study also determines the phenolic concentration in individual Green I, Green II, Brown I, and Brown II propolis samples and a mixture of Red, Green I and Brown II propolis samples.Red, Green I and Brown II propolis samples were mixed to study synergetic action on antioxidant activity.Here, Green I and Brown II propolis were selected since they have higher total phenolic compounds in HPLC-DAD analysis compared to Green II and Brown I propolis samples, respectively.

Total phenolics contents and antioxidant capacity
The results showed that total phenolics and flavonoids contents obtained in the DSE were higher compared to SSE method.Previous studies also reported that the Brazilian Green propolis have higher antioxidant capacity due to the presence of higher phenolic contents in them (Machado et al., 2016;Nunes et al., 2012;Paviani et al., 2013;Sawaya et al., 2011).The present study reveals that the dewaxing process leads to extracting/purifying more powerful phenolic structures and antioxidant compounds.
In our previous study, Andrade et al. (2017) demonstrated that Green propolis exhibited higher total phenolics (90.55 mg GAE/g sample) and flavonoids (59.45 mg QE/g sample) contents in the non-defatted 80% ethanolic extract than Brown and Red propolis.Distinctly, lower antioxidant capacity was observed in the Green propolis by the DPPH · (4554.35lmol trolox/g), ABTS ·1 (2214.96lmol trolox/g) and FRAP (604.20 lmol trolox/g) assays.This may be due to the non-defatted ethanolic extract which contains more bioactive compounds of both lipophilic and hydrophilic nature.Sulaiman et al. (2011) also demonstrated in the Iraqi propolis that total phenolics content did not reflect the antioxidant capacity.
The present study demonstrated that defatted ethanolic extract from the SSE of Green I propolis exhibited higher DPPH · scavenging activity.

Compounds
Moreover, the 80% ethanolic extract of Brown I propolis by SSE was higher in ferric reducing ability (46.34 mg TE/g extract) compared to other propolis samples.It may be due to the presence of a higher amount of total phenolics in them.Kasiotis et al. (2017) revealed that the dewaxing the hydroalcoholic extract of Greek (Imathia) propolis by precipitation method revealed that higher phenolic contents (181 mg GAE/g dry extract) influence the higher anti-DPPH (IC 50 ¼1.19mg/mL) activity than in other propolis samples.Hern andez Zarate et al. ( 2018) reported that higher total phenolics (181 mg GAE/g extract) and flavonoids (86 mg QE/g extract) contents influence higher antioxidant activity in the propolis collected from different regions of Mexico.Thus, these findings demonstrate that the dewaxed propolis samples exhibited higher total phenolics content and antioxidant activity.

Effect of phenolic compounds on solvent phase
The three methods (A, B and C) were applied in the defatted DSE ethanolic extracts of various propolis samples to identify the phenolic compounds based on the comparison of retention time and spectra of the standard compounds.In the two isocratic methods of A and B, all phenolic compounds were combined to shorten the analysis time but only a small number of compounds were separated in these samples.Similarly, some of the phenolic compounds have been identified in propolis using formic acid/ water/methanol as solvent phase and this may be due to the use of different C18 columns (Cao et al., 2007;Chang et al., 2008;Garc ıa-Viguera et al., 1993).The chromatogram from the method C reveals good separation of compounds in the propolis samples (Figure 1a) as compared to methods A and B. Since, acetonitrile solvent was used in the gradient elution method C, it showed good separation of compounds which makes also the identification easier than in other methods.Stalikas (2007) suggested that using acetonitrile in the mobile phase leads to a better resolution and sharper peak shapes in a shorter analysis time.Method C was applied to the different solvent extracts of propolis samples and the compounds were detected based on their respective wavelengths of the standard compounds.
In earlier works, Machado et al. (2016), Hamasaka et al. (2004), Falcão et al. (2009), Pellati et al. (2011), Falcão et al. (2013) and Kasiotis et al. (2017) also reported that the Green propolis showed higher concentration of phenolic compounds using 0.1% formic acid (phase A) and acetonitrile (phase B) as a solvent phase.In our previous study with UHPLC-QqQ-MS/ MS system using 0.1% formic acid in water (phase A) and 0.1% formic acid in acetonitrile (phase A) as a mobile phase showed Green propolis has higher artepillin C (4.80 mg/g sample), chlorogenic acid (2.81 mg/g extract), kaempferide (6.04 mg/g extract), kaempferol (1.48 mg/g extract), caffeic acid (1.06 mg/ g extract) and p-coumaric acid (5.34 mg/g extract) contents than in Brown and Red propolis samples (Andrade et al., 2017).In method C, acetic acid was used instead of formic acid since acetic acid is of low cost.Moreover, the maximum number of compounds is identified in method C. Correspondingly, Croci et al. (2009) reported that using solvent phase of 10% acetic acid (A) and acetonitrile (B) in HPLC-DAD analysis showed the presence of phenolic compounds such as benzoic acid, 3,4-dimethoxycinnamic acid, p-hydroxybenzoic acid, protocatechuic acid, chlorogenic acid, caffeic acid, o-coumaric acid, p-coumaric acid, salicylic acid, trans-cinnamic acid, syringic acid, ferulic acid, gentisic acid, r-(À)-mandelic acid, s-(þ)-mandelic acid and vanillic acid which could be separated better and these were also identified in the Romanian and Israel propolis.Thus, method C represents a cost-effective methodology that can be used on the laboratory scale.

Effect of phenolic compounds at different column oven temperature
The present study revealed that the temperature influences the phenolic contents analyzed in the various propolis samples.Beelders (2011) and Heidorn (2016) demonstrated that increasing the column oven temperature from 21 to 40 C and 30 to 110 C, respectively improves the separation efficiency and shortens the analysis time.From the chromatogram (Figure 1a-c), it is clearly observed that increasing the column temperature from 40 C to 60 C, the chromatogram peaks shifted and it shortened the analysis time.Similarly, the chromatograms of other propolis samples of Brown I, Brown II, Green II and RGB got also shifted and showed good separation efficiency (Figure S2).From the data it was concluded that column oven at low temperature (40 C) separates the maximum number of phenolic compounds in the Green I propolis compared to that of higher oven temperatures.However, on increasing the column oven temperature to 50 and 60 C, the total phenolic compounds contents in the propolis samples decreased.From the data presented in Table S1, it could be verified that phenolic compounds in propolis samples from New Zealand (Midorikawa et al., 2001), Brazil (Andrade et al., 2017;Bruschi et al., 2003), East Slovakia (Hrobonova et al., 2005), China (Cao et al., 2007;Sha et al., 2009;Yang et al., 2013), Romania and Israel (Croci et al., 2009), Portugal (Falcão et al., 2013), Mangosteen (Kongkiatpaiboon et al., 2015), Egypt (Abu Shady et al., 2016) and Taiwan (Chen et al., 2019) were identified by liquid chromatography at < 40 C column oven temperature.Zhang et al. (2013) optimized the phenolic acids and flavonoids at different column temperatures (28, 30, and 32 C) and their results showed the optimum column oven temperature of 30 C which improves the peak shapes and shortens the analysis time.Thus, it can be concluded that column oven temperature of 40 C is better for the detection of thermos-sensitive molecules in propolis samples.

Separation efficiency of phenolic compounds on solvent extraction
The separation efficiency of phenolic compounds in the various propolis samples was studied in the SSE method.Using this extraction method, the polar phenolic compounds were separated into ethyl acetate, acetone and 80% ethanol extracts.This study also revealed that the dewaxed 80% ethanolic extract obtained higher and pure phenolic compounds in the Green I propolis (Figure 2a-c).The chromatograms of other propolis samples of Brown I, Brown II, Green II and RGB also showed good separation efficiency (Figure S3).Higher efficiency in the separation of phenolic compounds can be achieved using SSE method in the HPLC-DAD analysis and this improves the solubility of the compounds.Thus, SSE using ethyl acetate followed by acetone and 80% ethanol could be considered as the most appropriate procedure for the extraction of phenolic compounds from Brazilian propolis samples.El-Hady et al. (2015) and Abd El-Hady et al. (2016) demonstrated that 70% ethanol, ether and ethyl acetate fractions from water extract obtained higher phenolic compounds which were identified by HPLC analysis.Kasiotis et al. (2017) identified 59 phenolic compounds by HPLC-DAD-ESI/MS analysis in the dewaxed (precipitation method) ethanolic extract of Greek propolis.From the data, it was observed that phenolic compounds in the propolis samples could be separated well, based on the polarity of the solvents.Thus, the SSE method is useful for the extract and purify the phenolic compounds from propolis.
From the HPLC-DAD results, the artepillin C content was obtained higher which is recognized as a characteristic constituent in the Green propolis (Lee et al., 2007;Salatino et al., 2005).Riani et al. (2019) also validated the UHPLC-MS method for the identification and quantification of artepillin C compound from the Brazilian Green propolis.

Conclusions
The present study highlighted that the concentrations of the phenolic compounds varied based on the type of propolis.The defatted ethanolic extracts showed higher phenolic compounds and good separation, among them artepillin C compound was higher.The optimized solvent phase (phase A: 1% acetic in water and phase B: 1% accetic acid in acetonitrile) for HPLC-DAD analysis was validated by measuring chromatographic characteristics of linearity and sensitivity and it can be used for further identification of phenolic compounds in propolis samples.Moreover, in the study, temperaturedependent phenolic compounds were estimated in the propolis sample by maintaining the different column oven temperatures.Heat-sensitive (protocatechuic acid) and thermo-stable (gallic acid, rutin and ferulic acid) compounds were observed in the propolis samples when the column oven temperature was maintained at 40 C. Since extraction was done at room temperature (25 C ± 2.0) and hence, some of the compounds may vary when the column temperature is increased.Furthermore, the efficacy of solvents in the separation of phenolic compounds was optimized in the various propolis samples by successive solvent extractions.Thus, based on the requirement of the phenolic compounds from the propolis, the solvent extraction, solvent phase and column oven temperature could be standardized.Overall, this study highlighted the effectiveness of solvent usage for the separation of phenolic compounds from propolis.This study will be useful in designing a unit on-site for further isolation, purification and characterization of bioactive compounds by using a combination of low to high polar solvents.

Table 1 .
Extract yield, total phenolic, flavonoid contents and antioxidant activity of different solvent extracts of various propolis samples.