MATERIALS AND METHODS

Samples of aggregates of pollen of stingless bees. Samples of pollen aggregates from commercial beehives of three stingless bee species (M. beecheii, S. mexicana and T. angustula) were collected from February 2017 to February 2018. The species were identified based on previous regional studies and using taxonomic information of the subfamily (^{Espinoza-Toledo et al., 2018}; ^{Ayala et al. 2013}). The location of the beehives, number of samples per species and dates of collection are shown in Table 1. The pollen aggregate samples were collected using sterile forceps from rational boxes (bee boxes) and then placed in sterile plastic jars that were previously labeled with the corresponding information. The samples were concentrated and stored at -4 ºC at the Biosciences Institute of the Autonomous University of Chiapas, in Tapachula.

Physicochemical analysis. The samples were determined in triplicate using the methods described by the ^{AOAC (2003)} and the following parameters: humidity (method 969.38), ashes (945.38), pH (method 962.19) and free acidity (method 962.19). The determination of free acidity was carried out according to ^{Bogdanov (2002}) and ^{AOAC (2003)} by potentiometric titration with an alkaline buffer up to pH 8.5.

Extraction of total phenols. For this, the protocol of ^{Carpes et al. (2007}) with modifications was used. To prepare the pollen aggregates solution, 1 g of aggregate was mixed with 5 mL of methanol:water (1:1), pH 2. For flavonoids, 80% ethanol left to stand for 24 h was used.

Analysis of phenols content. It was determined following the method of ^{Restrepo et al. (2009}) with the Folin Ciocalteu reagent. From the extract of pollen aggregates, 1 mL was taken to prepare concentrations at 25, 50, 75 and 100%. Each concentration was mixed with 500 μL of Folin-Ciocalteu reagent diluted with water 1/10, to which 400 μL of Na₂CO₃ were added. The mixture was kept in total darkness for 15 min to promote the reaction. Later, the absorbance was recorded at 765 nm. A mixture of methanol:water (1:1) was used as blank (control). The calibration curve to estimate the polyphenols content was created with a solution at concentrations of 0, 50, 100, 150, 200, 250 mg L^-1 in a methanol:water solution (1:1). The results of the total polyphenols content were expressed in equivalent mg of gallic acid per each 1 g of pollen aggregate.

Analysis of flavonoids content. It was determined using the technique of ^{Restrepo et al. (2009}). For this, quercetin dihydrate was used as a standard (1 mg mL^-1). The calibration curve was built with a quercetin/methanol solution at different concentrations: 0, 0.002, 0.004, 0.006, 0.008, 0.01 mg mL^-1. Afterwards, potassium acetate and aluminum trichloride were added, and the solution was kept in total darkness for 30 min. The absorbance was recorded at 415 nm.

Antioxidant capacity. To quantify the antioxidant activity, a discoloration of the radical 2.2´-azino-bis (3-ethylbenzothiazolin-6-sulfonic acid) was carried out following the ABTS methodology (^{Overveld et al., 2000}). The extracts obtained from pollen aggregates, at concentrations of 25, 50, 75 and 100%, were added with 20 mL of ABTS (2 mM) and 80 μL of potassium persulfate 70 mM. They were left to stand in darkness at 25 ºC for 16 h to produce the ABTS+ radical. The solution absorbance (ʎ = 734 nm) was adjusted to 0.800 ± 0.03 absorbance units in a spectrophotometer (Thermo Scientific Model 4001/4), using a 0.01 M phosphate buffer (Na₂PHO₄+KH₂PO₄+NaCl+KCl in 1000 mL of distilled water; pH 7.4 adjusted with NaOH). To obtain the calibration curve, 900 μL of the diluted solution of ABTS+ and 10 μL of Trolox solution (6-hydroxy-2, 5, 7, 8-tetramethylchrome-2-carboxilic acid) were added at different concentrations (0-400 µM in 80% v/v methanol). After 6 min standing, the absorbance was recorded at 734 nm using 80% v/v methanol as blank. The results were expressed as the antioxidant capacity equivalent to Trolox per g of pollen extract.

Antifungal activity. The C. gloeosporioides strain, belonging to the collection of the Biosciences Institute, isolated from Carica papaya fruits, and morphologically identified by Víctor Albores (unpublished data), was reactivated in potato-dextrose-agar (PDA) nutrient medium. The antifungal activity was evaluated by mixing methanolic extracts of the collected pollen aggregates of each bee species with PDA in vitro. The treatments, in triplicate, consisted of mixtures of 100 μL of extracts at 25, 50, 75 and 100% with 20 mL of PDA that was previously prepared. After sowing the strain in 0.5 cm colonial disks in the middle of 52 mm Petri dishes, it was incubated at 32 °C in a stove for 12 days. Mycelial growth was measured every 24 h with a Vernier caliper 0-150 nm capacity for nine days. The C. gloeosporioides colony growth in PDA under the same experimental conditions was used as the absolute control. The inhibition relative controls, in triplicate, were: 1) methanol/water, 2) chlorothalonil at 21 mg mL^-1, and 3) chlorothalonil at 56 mg mL^-1; the last two according to commercial doses applied to C. papaya crops by the Asociación de Fruticultores del Soconusco, Chiapas.

Rate of absolute growth. It was determined using the following equation: μ = (Db-Da) / (tb - ta). Where: Μ is the rate of the colony growth (mm day-1); Db is the diameter of the colony (mm) in time b; Da corresponds to the diameter of the colony in time a; and tb and ta is the time of absolute growth between two evaluations. For the comparative calculation of C. gloeosporioides growth in the four extracts (1-4) with respect to the relative controls (1-3), the following equation was used: DFC = [(D relative control_1-3 - D extract_1-4) / (D relative control_1-3) ] x 100. Where, DFC is the difference in the diameter of the colony (mm) divided by D relative control_1-3 and D extract of the pollen aggregate_1-4 expressed as percent.

Analysis of data. The data of all the physicochemical variables were subjected to a multivariate analysis of variance (MANOVA), considering each species-collection as a treatment. The media comparison was carried out using Hotelling’s test (α=0.05). To determine the separation or groups of collections and species, a linear discriminant analysis (LDA) was performed. A principal component analysis (PCA) was carried out to explain the relative weight of the physicochemical variables in the composition of the pollen aggregates within and among species. The relationship of the mycelial growth with phenols, flavonoids, Trolox and the antioxidant activity was estimated using r-Spearman. An ANOVA analysis by bee species and growth evaluation date was conducted in order to compare the effect of the four concentrations of extracts of pollen aggregates with the relative controls. All the analyses were conducted in INFOSTAT.

RESULTS

The result of the physicochemical analysis of pollen aggregates collected by M. beecheii, S. mexicana and T. angustula is shown in Table 2. Overall, the humidity values ranged from 12.2 to 31.4%, ashes from 0.9 to 5.9%, pH from 2.2 to 3.5, free acidity from 117 to 233.4 meq Kg^-1, phenols from 1.2 to 2.6 mg EAG g^-1, flavonoids from 0.9 to 3.1 µg EQ g^-1, Trolox from 4.3 a 8.1 mg, Trolox g^-1, and the concentration of glucose from 0.01 a 0.3 mg L^-1. The multivariate analysis was significant on bee species in the physicochemical content (p=0.0001). All the species-collection combinations were statistically different (p=0.05); M. beecheii had the highest values, while T. angustula had the lowest. The amount of phenols and flavonoids were positively related to the antifungal activity (p≤0.02), while the antioxidant capacity (Trolox) had a negative r-Spearman (-0.41 a-0.61) (Table 3). In most of the extracts associated with S. mexicana and M. beecheii species, the antioxidant activity had a negative correlation with the antifungal activity.

The discriminant analysis explained 91.1% of variance in the first two components, CAN 1 (80%) and CAN 2 (11.1%). The greatest weight in CAN 1 corresponded to phenols, flavonoids, and free acidity. The formation of groups clearly included the three replications of each species-collection but there was significant variation between intra and inter-species collections. According to MANOVA, the collections from M. beecheii beehives were separated from the other collections and distributed across the positive dimension of CAN 1. The other two species-collections were distributed across CAN 2, which had the greatest weight of pH, ashes, and glucose; S. mexicana had the highest dispersion (Figure 1A).

The analysis of main components explained 58% of the multivariate variance (CP1=43.5; CP2=14.8%) in the two first components. CP1 was mainly explained by flavonoids, pH, and free acidity, while CP2 was determined by ashes and glucose. Phenols had similar weight in both components (Figure 1B). The projection of the collections in this multivariate space unmarked M. beecheii samples from beehives established in an agroecological area where annual agricultural crops prevail (MB3 and MB2). The remaining collections had low dispersion regarding low values of CP1 and CP2.

Antifungal activity. The three bee species and four methanolic extracts that were evaluated showed C. gloeosporioides inhibition compared to the controls. From day three of incubation and up to the end, at day nine, statistical differences were observed in the fungus mycelial growth compared to the methanol-water (p<0.0001) (Table 4) control. From day four onwards, the concentrations at 25, 50, 75 and 100% started to differentiate from the same control. The differences ranged from 20 to 64% depending on the growth time rather than the species (Figure 2). From day five or six, the concentrations at 75 and 100% exceeded or equaled the inhibitory effect of chlorothalonil at its highest concentration (56 mg mL^-1). The S. mexicana extracts at 50% were the only ones that also exceeded the inhibition caused by the fungicide at that concentration. The T. angustula extracts had the lowest effect compared to chlorothalonil because only at a concentration of 100% exceeded the highest doses (Figure 2). The collection-species had an effect on the fungus inhibitory capacity. For example, M. beecheii MB1 collection had a higher effect than MB1 and MB2 (Figure 3). When the final mycelial growth of the relative control was compared to that of the absolute control (5.1 cm), it was found that methanol/water (3.8 cm) reduced the fungus growth by 25.4% and chlorothalonil by 35.3% (3.3 cm, 21 mg mL^-1) and 58.8% (2.1cm, 56 mg mL^-1). Overall, the rate of C. gloeosporioides mycelial growth ranged from 0.027-0.041, 0.022-0.037, 0.012-0.025, and 0.009-0.013 mm h^-1 at concentrations of 25, 50, 75 and 100%, respectively.

Figure 1 Canonical discriminant analysis of the three pollen collector bees (A) and analysis of the main components of the antimicrobial activity of the pollen methanolic extracts (B). The reference information of each pollen extract sample is described in Table 1.  

DISCUSSION

The physicochemical composition of the pollen aggregates collected from commercial stingless bee beehives was heterogeneous at intra- and inter-species level because of the bees’ territorial exploration capacity and pollen collection from different plant species (Table 1). However, it was possible to classify the species in groups, observing that M. beecheii was segregated from the other species. The species effect can be associated with vegetal exploitation niches in order to prevent interspecific competence more than some metabolic effect. Previous reports indicate that this is a single-flower species which prefer fabaceae 45% in contrast with M. solani and S. mexicana (^{Espinoza-Toledo et al., 2018}). The physicochemical composition at the pollen aggregates level has an implication in its quality, and nutritional, therapeutic, antioxidant and antimicrobial value (^{Mărgăoan et al., 2010}; ^{Saavedra et al., 2013}). Physicochemical heterogeneity has also been reported in studies about honey produced by meliponines in Soconusco (Espinoza-Toledo et al., 2018), and the value of these attributes for ecological purposes has been recognized (^{Vit, 2008}).The correlation from the physicochemical variables with the antioxidant capacity of the extracts of pollen aggregates suggests that phenols, flavonoids and free acidity exert a higher inhibitory action on C. gloesporioides mycelial growth (Table 3) (Mărgăoan et al., 2010). This could be associated with the type of phenolic compounds extracted from pollen aggregates, which do not have the property of donating donate electrons and are acidic and prevent a reaction with the ABTS oxidized radical. This compound acts on polyphenols which donate hydrogen atoms, while the ferric complex in FRAP acts on polyphenols that can donate electrons (^{Schaich et al., 2015}). It was not possible to state that the oxidizing capacity observed in the samples of extracts of stingless bee pollen aggregate was caused by a component in particular. The antioxidant capacity is the result of the interaction between molecules that make up each sample (^{Paulino-Zunini et al., 2010}). Based on this, the low relationship obtained with the antifungal activity would be defined by diverse factors, such as the chemical composition of the sample, geographical region and plant from which the pollen was taken (^{Bertrams et al., 2013}; ^{Duran et al., 2011}).

Figure 2 Average values of Colletotrichum gloeosporoides growth in vitro in the presence of methanolic extract at different concentrations of pollen aggregates of Melipona beecheii (Mb), Scaptotrigona mexicana (Sm) and Tetragonisca angustula (Ta). 

Figure 3  Colletotrichum gloeosporoides growth in vitro in the presence of methanolic extract of Melipona beecheii (MB100) (100%) pollen aggregate, as well as the controls methanol/water (M-A) and chlorothalonil 56 mg mL ^-1 (Cl56). 

The lowest relationship between the antioxidant and antifungal capacity in the extracts of T. angustula pollen aggregates could be caused by the sample composition (^{Mărghitaş et al., 2009}). On the other hand, it could also be due to a methodological effect, since, according to the solvent, it is possible to determine what molecule was extracted (^{Muñoz et al., 2015}). The inhibitory effect of the extracts of pollen aggregates is consistent with other reports where it is recognized that the type of biomolecule and its bioactivity can vary making it possible to differentiate regions and taxa, but implying the sample effect (^{Pellati et al., 2011}; ^{Li et al., 2011}; ^{Ruíz-Montañez et al., 2014}).

The extracts of pollen aggregates showed a clear inhibitory effect on C. gloeosporioides compared to chlorothalonil commercial fungicide. The inhibitory action in each extract of pollen aggregate was associated with the extracting action of the solvent and the chemical composition of the pollen samples. The relatively high concentration of phenolic compounds, which may include terpenoids, phenylpropanoids, stilbenes and saponins (^{Soto and Rosales, 2016}), can be involved in the antifungal activity. The hydrophilic extracts of plant cells are strongly related to Alternaria alternata and Fusarium oxysporum cell growth inhibition and conidia germination and can also be toxic in microbial respiration processes (^{Rodríguez-Maturino et al., 2015}). 60% of hydrophilic extracts contain a large amount of flavonoids and phenols (^{Pietarinen et al., 2006}; ^{Okwu and Nnamdi, 2008}).

The antifungal property of the pollen aggregates collected by A. mellifera has been reported to inhibit up to 70% of mycelial growth in Aspergillus niger and 99.9% in A. fumigatus (^{Kacaniova et al., 2012}). Similar effects have been reported in Alternaria, Botrytis and Fusarium (^{Cabrera and Montenegro, 2013}). Regarding antifungal properties of pollen aggregates associated with stingless bees, there are reports of T. angustula and 26% inhibition in Candida albicans (Rojas, 2015). In Soconusco, previous studies using meliponines honey showed 40% more effectiveness than chlorothalonil fungicide in C. gloesporioides (^{Albores-Flores et al., 2018}). In contrast, these results, using extracts of pollen aggregates collected by the same species (M. beecheii, S. mexicana), had a higher effect up to 50% than the fungicide in high commercial doses (56 mg mL^-1), depending on the extract concentration. This study provides the first report of the antifungal properties of meliponines pollen aggregates. In particular, for Soconusco, Chiapas, stingless bees are an alternative for pollination of different economically important crops, besides honey, pollen aggregates and propolis are medical alternatives for treating diabetic ulcers, and are also used in fruit crops pre- and postharvest processes (^{Espinoza-Toledo et al., 2018}; Albores-Flores et al., 2018, ^{Grajales-Conesa et al., 2018}).

CONCLUSIONS

In the pollen aggregates samples, the values of each of the physicochemical properties that were evaluated ranged as follows: pH: 2.2 to 3.5, humidity: 2 to 31 %, ashes: 0.9 a 5.9 %, free acidity: 117 to 133 meq kg^-1, phenols: 1.4 to 2.6 mg EAG g^-1, flavonoids: 0.9 to 3.1 mg EQ g^-1, glucose: 0.01 to 0.3 mg L^-1 and Trolox: 4.3 to 8.0 mg g^-1. The highest values of pH, ashes, glucose and Trolox corresponded to S. mexicana and T. angustula species. The highest values of free acidity, phenols and flavonoids corresponded to M. beecheii.

The bioactive compounds of the pollen aggregates involved in the antifungal action were free acidity, phenols, and flavonoids. The inhibition values reached by M. beecheii were 38-66% higher in C. gloeosporioides than those of chlorothalonil.