Honey samples

A total of 20 polyfloral honey samples (Apis mellifera) were collected in autumn 2018 by beekeepers, at different geographical regions in the state of Guerrero (Mexico) (Figure 1). All samples were stored at −4 °C in amber glass vials until use. A sugar analog (SA) was used as control, and it was composed by sucrose (1.5 g), 7.5 g maltose (7.5 g), 40.5 g fructose (40.5 g), glucose (33.5 g) and 17 mL of distilled water.

Figure 1 Mesoregions of the state of Guerrero, Mexico. 

Figura 1 Regiones del Estado de Guerrero, México. 

Total phenolic content

The total phenolic content (TPC) of honey samples was determined according to the Folin-Ciocalteu’s phenol (FCP) method (^{Singleton et al., 1999}), with slight modifications. Briefly, 5 g of each sample was mixed with 50 mL of distilled water to obtain a stock solution (10 % w/v). The resultant solution (0.5 mL) was mixed with 2.5 mL of FCP reagent (0.2

for 5 min, then 2 mL of sodium carbonate (Na₂CO₃) (0.7 M) were added. The solution was incubated at 25 °C for 2 h, in the dark. Absorbance was measured at 760 nm (VIS Genesys^® 20, Thermo Scientific, NY, USA). TPC was determined using gallic acid as standard and expressed as mg of gallic acid equivalents/100 g of honey (mg GAE/100 g honey).

Total flavonoids content

The total flavonoids content (TFC) of honey samples was determined by the aluminum chloride (AlCl₃) complex formation method (Marghitaş et al., 2009), with slight modifications. Briefly, 1 mL of honey solution (1 mg/mL) was mixed with 0.3 mL of sodium nitrite (NaNO₂) (5 %), and 0.3 mL of AlCl₃ (10 %) was added, after five minutes. Subsequently, after six min, the reaction mixture was neutralized with 2 mL of sodium hydroxide (NaOH) (1 M). The absorbance was measured at 510 nm (VIS Genesys^® 20, Thermo Scientific, NY, USA). TFC was determined using quercetin as standard and expressed as mg of quercetin equivalents/100 g of honey (mg QE/100 g honey).

Antioxidant activity assays

Free-radical scavenging activity. Antiradical activity was determined using the stable radical 1,1-diphenyl-2-picrylhy-drazyl (DPPH^•) method (^{Rodriguez et al., 2012}). Twenty μL of honey dissolved in methanol (10 %) was mixed with 200 μL of DPPH^• solution (150 μM, in 80 % of methanol). The absorbance was measured at 517 nm after 30 min. Results were obtained using Trolox as standard and expressed as µmol of Trolox equivalents/100 g of honey (µmol TE/100 g honey).

Trolox equivalent antioxidant capacity (TEAC) assay. The TEAC assay was conducted following the modified methodology described by ^{Vidal-Gutierrez et al. (2020)}. The radical solution was obtained by mixing 19.3 mg of ABTS in 5 mL of H₂O, with 88 mL of K₂S₂O₈ solution (140 mM) and incubated for 16 h in darkness. Subsequently, the radical absorbance was adjusted to an OD of 0.7 at 730 nm. Five μL of an aqueous honey solution (50 %) mixed with 245 μL of adjusted radical, were incubated for 5 min in the dark. The absorbance was measured at 730 nm. Results were reported as μM Trolox Equivalent (μmol TE)/100 g honey.

Ferric reducing antioxidant power (FRAP) assay. The FRAP assay was performed by the previous reported method (^{Tuberoso et al., 2011}). Briefly, the ferric complex was prepared by mixing 300 mM acetate buffer (pH 3.6), 2,4,6-tri(2-pyridyl)-striazine (TPTZ) (40 mM, dissolved in 40 mM HCl) and 20 mM aqueous ferric chloride (FeCl₃) in a 10:1:1 proportion. Then, 20 ml of an aqueous honey solution (20 %) were mixed with 280 μL of ferric complex The absorbance was measured at 630 nm in a microplate reader after 30 min of incubation in the dark. Results were expressed as μM Fe(II)/ 100 g honey.

Antimicrobial assay

Standard strains analyzed were Escherichia coli ATCC^® 25922™ (Gram-negative bacteria) and Staphylococcus aureus ATCC^® 25923™ (Gram-positive bacteria), maintained in tryptone soy broth (TSB) at 4 °C, as well as Candida albicans ATCC^® 90028™ (fungal strain), maintained in Brain Heart Infusion broth (BHI) at 4 °C. The inoculums were prepared in TSB and BHI for bacteria and yeast respectively at 37 °C for 24 h. Cell suspensions were diluted in peptone water (0.1 %), to a 0.5 of McFarland scale concentration (1.5x106 CFU/mL). Minimum inhibitory concentration (MIC) was determined with adapted method by ^{Bueno-Costa et al. (2016)}. Briefly, 10 µL of each 1.5x106 CFU/mL suspension, 90 µL of Trytone TSB or BHI and 100 µL of each honey solution (12.5-400 mg/mL) were added to each pool. Negative control was distilled water, while the antibiotic ciprofloxacin (15 µg/mL) and nystatin (100 IU/mL) were used as positive control. Plates with microdilution were incubated at 37 ^ºC for 24 h and measured at 620 nm in a microplate reader (Thermo Scientific™ Multiskan™ FC, NY, USA).

Statistical analysis

Statistical analysis was performed through the IBM SPSS Statistics, 2020 software. A Principal Component Analysis (PCA), Pearson’s correlation and linear regression analysis were carried out to evaluate the color influence on phenolic composition (TPC and TFC), antioxidant and antibac-terial activity. The significance was set a p < 0.05.

Color intensity

Color of honey samples were high, ranging from 17.0 to 146.0 mm Pfund, with five colors found in the studied honeys (Table 1), White (20 %), Extra Light Amber (10 %), Light amber (25 %), Amber (25 %) and Dark amber (20 %) (Table 1). This is in agreement with ^{Bueno-Costa et al. (2016)}, who reported Light Amber color for Brazilian honeys collected at different zones. Probably the dark tone found in Mexican honey is due to the wild origin of most honey produced in the country. In Guerrero state, Mexico, there is a wide variety of vegetation, which favors the dominance of wild honey, such as Tropical Deciduous Forest, Rain Forest and Coniferous Forest. Light honeys, such as the honey from the Harenna Forest in Ethiopia, showed Pfund scales between 34 and 85 mm (^{Belay et al., 2015}).

Total polyphenol content

Table 2 show the results of TPC in honey samples. The total phenolic compounds were higher in Dark Amber honey (101.5 mg GAE/100 g honey), followed by Amber (66.4 GAE/100 g honey), Light Amber (68 mg GAE/100 g honey), Extra Light (42.5 mg GAE/100 g honey) and White (17.3 mg GAE/100 g honey). These results show a high correlation observed between color and TPC (r = 0.895, p < 0.01) (Table 4) this implies that the amount and type of polyphenolic substances in honey are variable and essentially depend on the floral origin (^{Küçük et al., 2007}). Similarly, the presence of higher phenolic contents in darker color honeys than lighter honeys and their strong correlations are well documented for Cuban (^{Alvarez-Suarez et al., 2010}), Argentina (^{Isla et al., 2011}), and Brazilian (^{Bueno-Costa et al., 2016}) honeys.

Total flavonoid content

The TFC of honeys samples are displayed in Table 2. The results showed that Dark Amber honey possess a high content of total flavonoids (22.45 mg CE/100 g honey), followed by Amber (16.24 mg QE/100 g honey), Light Amber (14.77 mg QE/100 g honey), Extra Light Amber (14.77 mg QE/100 g honey) and White (9.58 mg QE/100 g honey). These results also showed a high correlation between TPC and TFC (r = 0.814, p < 0.01) (Table 4). In addition, these results showed a high correlation between color values of the honey and TFC (r = 0.864, p < 0.01) (Table 4). TFC are also related to the floral sources as discussed previously ^{Bueno-Costa et al. (2016)}.

Antioxidant activity

In this study, three in vitro assays were used to determine antioxidant activity. The DPPH^• radical scavenging activity varied significantly among most honey samples (Table 2). The highest antioxidant activity (24.0 µM TE/100 g honey) was observed in the Dark Amber honey sample (SC13), whereas the lowest activity (2.9 µM TE/100 g honey) were observed in White honey sample (MT7). Higher correla-tions were observed between the DPPH^• activity and TPC (r = 0.945, p < 0.01), DPPH^• activity and TFC (r = 0.792, p < 0.01), and between TPC and TFC (r = 0.814, p < 0.01). These results are similar with the reports of ^{Ferreira et al. (2009)}, ^{Alvarez-Suarez et al. (2010)}, Saxena et al. (2010) and ^{Azonwade et al. (2018)}, who found that there is a positive correlation between DPPH^•, TPC, and TFC.

Free radical scavenged activity of honey samples was also determined through the TEAC assay (Table 2). Results evidenced that evaluated honey exhibited scavenging activity against ABTS•+ radical, between 40.0 and 290 µM TE/100 g honey, Dark Amber showing the highest TEAC values, with a correlation found between color and TEAC (r = 0.824) (Table 4). The results present here are similar with those of other authors who demonstrated a correlation between honey color and TEAC (Alvarez-Suarez et al., 2012; ^{Chen et al., 2017}). Moreover, data obtained by the FRAP evaluation are presented (Table 2). The results obtained in Dark Amber exhibited a high ferric reducing activity (465.7 µM Fe(II)/100 g honey). The antioxidant properties of the Dark Amber and Amber honeys were within the ranges of the antioxidant values reported for Polish honey (^{Kuśet al., 2016}). The differences in antioxidant properties of the honey samples could be due to the variations in phytochemicals of the respective plants and their geographical origins (^{Amarowicz et al., 2004}; ^{Jasicka-Misiak, 2012}). Higher correlations were observed between the FRAP and color (r = 0.903, p < 0.01), TPC (r = 0.888, p < 0.01) and TFC (r = 0.830, p < 0.01). The Dark Amber honeys exhibited higher antioxidant activity (p < 0.01) in all the antioxidant assays.

Antimicrobial activity

All honey samples showed antimicrobial activity against the two bacteria and yeast tested (Table 3). The antibacterial activity was more effective against Gram-positive than Gram-negative bacteria. Thus, more efficient results occurred against S. aureus with averagely range from 12.5 to 75.0 mg/mL. With regard to gram-negative bacteria, the MIC of the studied honeys samples varied from 50.0 to 200 mg/ mL. Other authors reported that Gram-positive bacteria were more sensitive to the honeys antibacterial activities than Gram-negative ones (^{Alvarez-Suarez et al., 2010}; ^{Isla et al., 2011}). Regarding the antifungal activity against C. albicas, the study indicated that fungi were less susceptible than bacteria ranging from 50.3 to 300 mg/mL. The lower susceptibility of fungi to different honey samples, in comparison of bacteria, is documented (Kacˇaniováet al., 2011; ^{Al-Waili et al., 2013}).

However, other factors, in addition to the phenolic composition, such as the presence of hydrogen peroxide, catalase and glucose oxidase, which are known to be present in honeys of diverse origins (^{Stagos et al., 2018}), may have contributed to the antimicrobial activity of the studied honeys. Moreover, the presence of a high content of flavonoids could contribute to its bioactivity. The same samples also showed the best results with regard to antioxidant activities, both of light amber color and showed intermediate compounds of phytochemicals. Moreover, the more antimicrobial properties of some of the honeys could be due to their higher phenolic content and antioxidant properties. The strong relations of antimicrobial properties of honeys with their antioxidant properties and phenolic contents is well discussed (^{Isla et al., 2011}).

PCA analysis

Two major factors were extracted using PCA and the results are shown in Figure 2. These main components (PC1 and PC2) explained 60.9 and 17.5 % of the variability, respectively. The screen graph suggested that PC1 contained most of the information, followed by PC2. The important variables in PC1 were Color, Total Phenol Content (TPC), Total Flavonoid Content (TFC) and Antioxidant Activity with, DPPH^•, TEAC and FRAP and finally PC2 was influenced by the antimicrobial activity against S. aureus (AASA), antimicrobial activity against E. coli (AAEC) and antimicrobial activity against C. albicas (AACA). A regression analysis was performed and Pearson’s correlation coefficients were calculated to determine in detail the correlations between the variations in the biological properties of the samples. The graphs of the first two components clearly indicated that the darker honeys had higher content of phenol, flavonoids and antioxidant capacity than the lighter honeys, and that the antioxidant capacity was strictly related to the total phenolic content (Fig. 2). The multivariate linear analysis showed a high association between variables (p ≤ 0.01). Color was correlated with total phenols (r =0.895), total flavonoids (r = 0.864) and antioxidant capacity (DPPH^• r = 0.809), (TEAC r = 0.824) 1 and (FRAP r = 0.903). No statistically significant correlations were found between color and antimicrobial activity.

Figure 2 Principal component analysis plot of the evaluated parameters. 

Figura 2. Gráfico de componentes principales de los parámetros evaluados.