Comparison of the Antibacterial Effect of Different Biological Silver Nanoparticles Synthetized and Integrated with Honeys

This study compares the morphologies, zeta potentials, and antibacterial effects a total 12 different microcompounds containing honey and silver nanoparticles, in a novel study of the difference between honey samples in nanoparticle synthesis, as well as the antibacterial interaction that those honey samples can have with the silver nanoparticles synthetized using them. Microcompounds were synthetized by combining silver nitrate solution with a honey sample and performing one of methods of biogenic synthesis: sunlight exposure, basification to pH 5 or basification to pH 10. Samples of each microcompound were also submitted to heat treatment, obtaining thus heated variants. Morphology and size data were obtained by Dynamic Light Scattering (DLS) analysis and Scanning Electron Microscopy (SEM); while zeta potential was measured by Electrophoretic Light Scattering. Broth microdilution, time-kill curves and SEM were used to access the antibacterial effect. Mean diameter of particles inside all microcompounds varied between 100 nm and 150 nm; and the zeta potential varied depending on the honey used. Minimal Inhibitory Concentrations (MIC) of microcompounds were between 15 μM and 500 μM. Time-kill curves showed that microcompounds had a faster and stronger effect against Esch- erichia coli than Staphylococcus aureus . Microcompounds obtained by basification to pH 5 or by sunlight were bactericidal, as they were capable of inhibiting bacterial growth (resulting in an antibacterial efficiency of 100% in 24 hours) at 125 μM against S. aureus and 62.5 μM against E. coli . SEM micrographs showed bacterial cells with lower cell density, blebs and other alteration after microcompound treatment.


Introduction
According to the World Health Organization (WHO) [1] , infections caused by resistant bacteria take 700,000 lives yearly. The selection of resistant bacteria occurs naturally due to the prescription of antibacterial substances; however, the misuse of antibiotics and their inappropriate disposal accelerate the process [1] . Antimicrobial resistance to available drugs is detected in 35 percent of common human infections in countries participating in the Organization for Economic Co-operation and Development (OECD). Additionally, low and middle-income countries present 80 to 90% resistance rates for treatment-bacterium combinations, which led to WHO compiling a list of antibiotic-bacterium combinations that require an alternative [2,3] . Thus, many organizations have established the necessity for developing new antibacterial treatments, including "remodelling" old treatments [1,4,5] .
Due to their positive surface charge, most nanoparticles can interact with the negatively charged structures. Target examples include phosphates and carboxylates present in the external wall of Gram-negative bacteria, and teichoic acid in Gram-positive bacteria [12][13][14][15] . Once inside, they can display multiple antibacterial mechanisms, depending on their morphology, interact with the bacterial membrane and intracellular components such ribosome and bacterial chromosomes, causing oxidative stress and disrupting the bacterial cell's electronic balance, and reach an antibacterial efficacy of 100% [9,11,[14][15][16] . These properties grant metal nanoparticles a more selective antibacterial compared to colloidal silver, enabling their clinical use [11,17] .
Honey samples can present varying physicochemical thus antibacterial properties, according to geographical and botanical origins, soil composition, producing bee, and harvest season [29][30][31][32][33][34] . Notably, the reducing sugars and oxygen peroxide can vary significantly samples. For example, honey produced by Scaptotrigona bipunctata (shown to contain a relevant antibacterial potential) has lower sugar content and greater humidity than the conventional honey produced by Apis mellifera [31,[34][35][36] . Therefore, usage of a different honey sample may alter silver particles' structure as well as their antibacterial property.
Additionally, honey contains antibacterial properties, which can further enhance the antibacterial potential of nanoparticles. For example, the hydrogen peroxide causes oxidative stress in bacteria, and flavonoids exhibit antioxidative, antiviral, and anticancer properties [29,30,[35][36][37][38] . Furthermore, Manuka honey, produced by A. mellifera collecting nectar from the Leptospermum ssp. trees; is used in products such as the Medihoney ® Antibacterial Wound Gel for professional wound care, which is an example of honey's potential in the clinical use of nanoparticles [29,39,40] .
In light of the antibacterial effects of honey and silver nanoparticles, we decided in a novel study to compare the nanoparticles synthesis using two honey samples with contrasting compositions. We intend to assess how the differences in honey composition can affect the morphology and antibacterial effect of the resulting silver nanoparticles, as well as observe the antibacterial interaction those honey samples can have with the particles.
Thusly, we developed and compared different microcompounds made with silver and honey with two honey samples -a honey sample collected from Apis mellifera (nominated Honey Apis mellifera-"HAM") and another one collected from Scaptotrigona bipunctata (nominated honey Scaptotrigona bipunctata-"HSB"). Microcompound synthesis used three different methods for nanoparticle formation: basification, sunlight exposure, and aggregative basification. After synthesis, post-heating of microcompound samplings resulted in different morphological and antibacterial properties. The resulting twelve types of microcompounds are compared in this work in relation to their particle size and antibacterial properties, through Dynamic Light Scattering (DLS) measurement and antibacterial assays, respectively, in order to assess their clinical potential; in addition, this article presents and analyses micrographs of microcompounds and bacteria obtained by Scanning Electronic Microscopy (SEM).

Honey Harvesting and Composition Analysis
The HAM and HSB samples originated from the university meliponary (Universidade Estadual de Londrina, Londrina-PR, Brazil) and the Unidade de Conservação Monte Sinai (Mauá da Serra-PR, Brazil). They were col-lected between 2018 and 2019 with glass syringes and metal spatulas; and deposited in sterilized jars for transport. HAM honey was harvested from Africanized A. mellifera Latreille (Hymenoptera: Apidae) hives, and HSB was collected from S. bipunctata Lepeletier 1836 (Hymenoptera: Apidae: Meliponinae) nests. Both have unifloral Eucalyptus sp. as their botanical origin.
Comparison of HAM and HSB composition, color, humidity, ash content, pH, total HAM and HSB composition, color, humidity, ash content, pH, total acidity, and sugar content comparison followed standards and protocols well established by the Association of Official Analytical Chemists International (AOAC International) in 2019 [41] . The color analysis used a Lovibond instrument, the humidity measurement, a digital refractometer (AOAC 969.38 method). The ash content was weighted by heating in an industrial oven at 600 °C for four days (AOAC 920.181 method). The honey pH, lactones, and total acidity were determined by titration of samples with a 0.1 M NaOH solution (Sigma-Aldrich ® Brazil) to a pH value of 8.3 and 10 (AOAC 962.19 method). The reducing and total sugars were analyzed by apparent reducing sugar (AOAC 920.183 method), based on the reduction of copper hydroxide in Fehling's solution. All tests occurred in the Food Analysis Laboratory of the State University of Londrina, using non-sterilized HSB and HAM samples. Before microcompound formation, both HAM and HSB were diluted in water (50% v·v -1 ), sterilized by filtration using a 0.22 µm filter (Millipore ® ), and stored at 4 °C.

Methods of Biogenic Microcompound Synthesis and Measurement by DLS and SEM
Microcompounds were obtained by varying three factors in their synthesis: the chosen honey sample, the mechanism of nanoparticle synthesis, and heating after the nanoparticle formation. The methods used for synthesis in this study followed data and protocols from articles published by González et al. (2016) [20] , Priz (2014) [42] , and Tagad et al. (2013) [43] .
Three prepared samples of "HAM" and "HSB", each at 50% (v·v -1 ), were mixed (50% v·v -1 ) with a solution of 5 mM of AgNO 3 (Sigma-Aldrich ® , Brazil); and afterward, each solution underwent one of the three methods for biogenic particle synthesis selected. As shown in Table 1, solutions containing the microcompounds' substrates were either exposed to sunlight for 10 min; or basified by adding a 1 M NaOH solution (Sigma-Aldrich ® Brazil) until the solution to pH 5.0 or 10.0. All solutions were mixed and left at room temperature for 5 min before storage at 4 °C for 24 h. Samples from each microcompound were then collected and heated separately at 60 °C for 30 min by water bath, and stored at 4 °C before performing the assays described below. The total number of microcompounds tested were twelve. The nomenclature of microcompound type shows its constituent honey, the method used in its synthesis, and whether the microcompound underwent heat treatment as schematized in Figure 1 and shown in Table 1.  Microcompound diameter was determined through dynamic light scattering (DLS) analysis with photon correlation spectroscopy (ZetaSizer NanoZS, Malvern ® Brazil). Zeta potential was determined by electrophoretic mobility inspection of surface charges using the same equipment. Readings for all microcompounds were made in triplicate. Microcompounds in the solutions obtained from the synthesis were washed through centrifugation of 1 mL collected from each sample for 1 h at 24 °C and 31,340.00 g. The resulting pellets were resuspended in 1 mL of deionized water inside microtubes, obtaining thus washed microcompounds. Silver particles inside the microcompounds were assessed through atomic absorption spectroscopy at 400 nm (AAnalyst 700, Perkin Elmer, USA) and compared to reference standard solutions (Specsol, Brazil).
Aliquots of 10 µL from the microtubes containing the washed silver particles were pipetted on polylysine-coated (1%) glass slides. Those were fixed after drying by immersion in a sodium 0.1 M cacodylate buffer solution (pH 7.2) containing 2% glutaraldehyde and 2% paraformaldehyde for 20 h; and then post-fixed in 1% OsO4 for 2 h. The fixed slides were submitted to progressive dehydration by submersion in increasing ethanol gradients (70,80,90, and 100 o GL). Next, samples were critical point-dried using CO 2 (BALTEC CPD 030 Critical Point Dryer), coated with gold (BALTEC SDC 050 SputterCoater). Lastly, the prepared slides were observed with a scanning electron microscope (FEI Quanta 200), and pictures were taken at 20,000×. The diameter of equivalent circle, roundness, and circularity of particles within the microcompounds were analyzed using the ImageJ photoshop software.

Microdilution Assays of Washed Microcompounds
Minimal Inhibitory Concentrations (MIC) of each type of washed microcompound were evaluated following the Clinical and Laboratory Standards Institute (CLSI, 2018) [45] guidelines for serial dilution in 96-well plates. A bacterial strain was cultivated in Mueller-Hinton (MH) broth (Difco ® ) agar for 24 hours before the assay, and then, suspended in a sterile saline solution (NaCl 0.85%, Sigma-Aldrich ® Brazil) at 1.5 × 10 8 CFU·mL -1 (0.5 on the McFarland scale); from which 10 μL were transferred to microtubes containing 990 μL MH (Difco ® ) broth. From those microtubes, aliquots containing 50 μL were pipetted in wells containing 50 μL of a microcompound solution, obtained by serial dilution of a given type of microcompound in MH (Difco ® ) broth. The final concentration of bacterial cells was 5.0 × 10 5 Colony Forming Units (CFU) per millilitre.
The concentrations of microcompounds tested ranged from 15.625 μM to 500 μM. Each concentration was tested in triplicate per microcompounds per bacterial strain wells with microcompounds, alongside wells without any treatment (bacterial viability control) and well without any bacterial cell (sterility control), were alongside cultivated. The 96-well plates were incubated for 24 h at 37 °C and; after incubation, the MIC values were defined as the lowest concentration of microcompound that inhibited visible growth. The results obtained were compared to the microcompounds MIC obtained in a previous study (see the Results section) [46] .

Time-kill Curves of Bacteria and SEM Imaging of E. coli Affected by Microcompound Treatment
Investigation bacterial growth inhibition caused by the microcompounds was performed through time-kill assays, in compliance with National Committee for Clinical Laboratory Standards protocols (NCCLS, 1999) [47] . A total of six assays were performed, two using bacteria S. aureus ATCC 29213 and four with E. coli ATCC 8739. Microcompounds were analysed at 125 μM against S. aureus, and at 125 μM and 62.5 μM against E. coli, following MICs modal values obtained. In all assays, microtubes were prepared as well as in the microdilution assays, with an initial inoculum of at 5.0 × 10 5 CFU·mL -1 .
All microtubes were incubated at 37 °C with constant stirring for 24 h. At six time points (0, 2, 5, 7, 10, and 24 h after incubation), aliquots of each sample were collected and serially diluted ten-fold in microtubes with MH (Difco ® ) broth. Three aliquots of 10 µL from each microtube were pipetted in MH (Difco ® ) agar media plates, and the plates were incubated at 37 °C for 18 h. Colonies were counted for all dilutions and treatments.
Based on time-kill differences (see Discussion), HAM pH 5δ, HAM pH 10δ, HSB L and HSB pH 10 microcompounds, alongside a positive (with AgNO 3 ) and a negative (without any treatment) control for cellular damage, were chosen as treatments for the morphological analysis by SEM of treated E. coli ATCC 8739. E. coli samples were prepared by incubation for 4 h at 37 °C in microtubes containing each a microcompound type at 125 μM, in similar conditions to the time-kill curve assay. Afterward, all microtubes were centrifuged for ten minutes at 7,000 g. Bacterial pellets were resuspended in 100 μL of deionized water, deposited in polylysine-coated (1%) glass slides, and submitted to stages of slide preparation (fixation, post-fixation, dehydration, and critical point drying) in the same manner the nanoparticles were (described in the previous section). SEM images were taken at 6,000× and 15,000× magnifications, with the latter being chosen for morphology analysis using ImageJ.
For the experiment, supplemented DMEM from each well was replaced with the same media now containing a type of washed microcompounds (described previously) at concentrations varying from 460 to 3680 μM (or 50 to 400 μM); and cells were cultivated for 24 h. Wells containing non-treated LLC-MK2 cells were also prepared as a positive control. Cell viability was determined by MTT (Merck ® ) reduction following supplier's recommendations, which were also used to determine the cytotoxicity curve and the nanoparticle concentration to inhibit cell growth by 50% (CC50).

Data Analysis
Variables in honey composition analysis are presented as mean values ± standard deviation and compared by Student's independent t-test with Welsh's correction. ImageJ measurements of SEM micrographs were compared by Mann-Whitney or Kruskal-Wallis test followed by Dunn's test; and are presented as mean values ± standard deviation. CFU numbers obtained by the time-kill curve assay are displayed and were analysed in log10. The cytotoxicity curves shown in Figures 11A and 11B were obtained through polynomial regression. All tests were performed with the statistical program GraphPad Prism version 6.02. Table 2 below describes the results obtained from the compositional analysis of HAM and HSB and the statistical differentiation by Student's independent t-test with Welsh's correction. In comparison to the HAM sample, HSB presented a significantly lower (p < 0.05) concentration of total sugars and significantly higher (p < 0.05) humidity, free and total acidity. Table 3 shows the average size measurements obtained by DLS of the microcompounds synthesized using the different methods and honey samples described in this article and the ranks obtained through multiple comparison tests. Average size data is expressed in intensity, volume, number, and hydrodynamic size. The difference among the size means of a given type of microcompound attests to a polydispersity in the microcompounds. HAM microcompounds such as HAM L, HAM Lδ, HAM pH 5, and HAM pH 5δ varied in size, mostly between 350 and 620 nanometres. There is no difference between their sizes (or zeta potentials). In contrast, HAM pH 10 and HAM pH 10δ were more voluminous than the other four HAM microcompounds. Meanwhile, HAM pH 10δ presented a significantly lower average (in number) than HAM pH 10, which can be associated with the heat treatment. Notes: All values (except "whole colour difference") are expressed in mean ± standard error of measurements made in triplicate. Ash content, total reducing sugars, and humidity are expressed in percentage (gg -1 ). Free, lactone, and total acidity are expressed in milliequivalent per kilogram (mEq.kg -1 ). Color and Whole color differences are noted according to the Lovibond scale. Significant differences between variables were analyzed by Welsh's t-test (p < 0.05). Symbols and abbreviations: HAM, Honey samples collected from Apis mellifera honey honeycombs; HSB, Honey samples collected from Scaptotrigona bipunctata honeycombs; 1, HSB average significantly lower than HAM average (p < 0.05); 2, HSB average significantly higher than HAM average (p < 0.05).

Microcompound Size Measurement by DLS
The reactions of involved in the green synthesis of silver nanoparticles (and thus microcompounds), are well described by many studies, alongside the factors involved during the process [1,2,20,[48][49][50][51] . As González et al. (2016) [20] observed, glucose is the primary reducing sugar in honey, which oxidation provides electrons for nanoparticle formation, which aggregate in the silver in nanoparticles. Afterward, as Meshram et al. (2013) [51] noticed, they are then capped by molecules such as gluconic acid and proteins, forming thus the silver microcompound (or nanoparticle) and preventing further aggregation [49][50][51] . The microcompounds obtained in this study varied in size and zeta potential according to the honey sample used, method of biogenic synthesis, and heating, as displayed in Table 3 and the results below.
As shown in Table 2, HAM present a greater amount of reducing sugars than HSB, thusly, HAM microcompounds presented overall smaller measurements than HSB microcompounds, in particular. Authors also have established pH to have an inverse correlation with the average size of silver particles [42][43][44] . Thus, the overall larger sizes of HSB microcompounds can also be related to its higher free acidity.
However, when microcompound synthesis involved basification to pH 10 ( Table 3), microcompound size was greater as shown in most measurements, however HSB pH 10 and HSB pH 10δ were the smallest in number among HSB microcompounds, and smaller than their HAM counterparts. Nanoparticles tend to aggregate at highly alkaline pH values due to the high ionic strength caused by the concentration of hydroxide ions in the solution. As a result, larger structures form, which causes a difference between the average size in number and average size in intensity and volume [42][43][44] . Therefore, the measurements of HSB pH 10 and pH 10δ may be a result of the higher acidity of HSB.
Additionally, the heat treatment increased the average size (in number) of HAM L, HAM pH 5 and HAM pH 10 in number (Table 3); a finding that corroborates with the aggregative effect caused by increased heating time and temperature reported in the literature [42][43][44]52,53] .
In the case of HSB microcompounds, their microcompound size reduction can be related to the fact that heat and acidity are both promoting factors for the brine electrolysis (Table 3). Furthermore, Jiang et al. (2011) [54] show that the size of nanospheres may decrease in certain conditions, such as a temperature between 32 °C and 55 °C; while additional factors, such as acidity, can favor the dissolution of particles in these conditions. The temperature range is close to the temperature of the heat treatment performed in this study. Thus the dissolution reported by Jiang et al. (2011) [54] can be related to the smaller size of HAM pH 10δ in comparison to HAM pH 10.
Therefore, the results imply the possibility that there are two events in contraposition regarding microcompound size: aggregation, favored in microcompounds synthesized with HAM and high alkalinity, and dissolution, favored by factors such as the higher free acidity of HSB. Further studies are, however, necessary to confirm this model. Notes: Size measurements are displayed in mean ± standard deviation. Abbreviations: HAM, Microcompounds with honey samples collected from Apis mellifera honey honeycombs; HSB, Microcompounds with honey samples collected from Scaptotrigona bipunctata honeycombs; Hidro. Size, Hydrodynamic size; L, microcompound obtained by light exposure; Lδ, microcompound obtained by light exposure and heating; pH 10, microcompound obtained by "aggregating basification"; pH 10δ, microcompound obtained by "aggregating basification" and heating; pH 5; microcompound obtained by "basification"; pH 5δ, microcompound obtained by "basification" and heating.
Another possible effect of using a honey sample with a higher concentration of free acidity (such as HSB) for microcompound synthesis is the alteration of its surface charge. Table 4 displays and compares the zeta potential of microcompounds produced with HAM and HSB. Notes: Size measurements are displayed as mean ± standard deviation. Symbols and abbreviations: HAM, Microcompounds with honey samples collected from Apis mellifera honey honeycombs; HSB, Microcompounds with honey samples collected from Scaptotrigona bipunctata honeycombs; L, microcompound obtained by light exposure; Lδ, microcompound obtained by light exposure and heating; pH 10, microcompound obtained by "aggregating basification"; pH 10δ, microcompound obtained by "aggregating basification" and heating; pH 5; microcompound obtained by "basification"; pH 5δ, microcompound obtained by "basification" and heating.
In regards to zeta potential, zeta potentials of HSB microcompounds presented greater modular charges than the corresponding HAM microcompound in four out of six cases. In other words, only in cases which microcompounds were obtained by basification without heat treatment HAM microcompounds obtained a greater charge. Our data also show that HSB microcompounds (HSB L, HSB Lδ, HSB pH 5, HSB pH 5δ) have a positive surface charge, except in microcompounds with high alkalinity (HSB pH 10 and HSB pH 10δ). The results also indicate that high alkaline synthesis processes can grant a negative charge and increase it (as well as provide a high ionic force), as seen with HAM pH 10, HAM pH 10δ, HSB pH 10, and HSB pH 10δ ( Table 4). This characteristic has many pharmacological implications: for example, the pH of many tissues targeted by pathogenic bacteria is acidic; consequently, microcompounds with a positive surface charge can present higher stability in those environments [55][56][57] . Additionally, given that the cell wall of most bacteria is negatively charged, microcompounds with a positive charge potential have a greater tendency to bind to the structure. Authors such as Radovic-Moreno et al. (2012) [56] and Qiao et al. (2018) [58] have developed effective surface-changing antibacterial delivery systems, and Abbaszadegan et al. (2014) [59] noted that nanoparticles with positive charge have a better antibacterial effect against Gram-positive and Gram-negative bacteria.

SEM Images and Measurements of Centrifuged Particles inside Microcompounds
The micrographs in Figure 2, taken at 20000× by SEM, show the round morphology and size of silver particles inside each microcompound, while size averages (as equivalent circular diameter-ECD), roundness, and circularity values are presented in Table 5, alongside significant differences demonstrated by Dunn's and Kruskal-Wallis tests.
The average size (ECD) among the silver particles inside all types of microcompounds varied between 100 nm and 150 nm; and thus, can be considered a mixture of nanoparticles and fine particles [60][61][62][63] . Significant differences (p < 0.05) found in particle size averages were not corelated to microcompound average sizes measured by DLS. For example, while HSB usage resulted in microcompounds significantly larger overall, the average particle diameter of HSB microcompounds do not have significant differences in four cases. They are significantly smaller (p < 0.05) in the other two groups: in the "L" and "pH 5" groups (without heat treatment).
HAM pH 10 and HAM pH 10δ microcompounds had larger agglomerates (Table 2); however, the particles inside the "larger" HAM pH 10 and HAM pH 10δ microcompounds were significantly smaller (p < 0.05) than those in HAM L, HAM Lδ, HAM pH 5 and HAM pH 5δ ( Table 5). This finding can be related to the inverse correlation between pH and particle size, previously discussed [42][43][44] . Furthermore, the paired comparison of particle sizes between heated and non-heated microcompounds showed that the heat treatments significantly reduced (p < 0.05) the average size of particles inside HAM Lδ and HAM pH 5δ in comparison to HAM L and HAM pH 5.
As for the particles' morphology, the silver nanoparticles were revealed to be somewhat elongated, presenting circularity averages between 0.20 and 0.35, and generally round, presenting roundness averages between 0.64 and 0.74 among all types. These results (shown in Table 5 and Figure 2) show that the effect of a given variable (honey choice, synthesis method…) can vary depending on the other factor involved. For example, even though silver particles in HSB pH 10 presented an even significant- Notes: Micrographs in Figure 4A and 4B were taken at 6000× and 15000×, respectively The red circles show silver particles inside the HSB L and HSB Lδ microcompounds, with the latter portraying higher ratios of roundness and circularity after undergoing the described heat treatment. Blue circles display images of HAM pH 5δ and HAM pH 10δ particles, which exemplify how the higher degree of basification may result in smaller and more elongated particles. Lastly, the yellow circles point to HAB MB and HSB pH 5 particles, demonstrate how using HSB honey in the same synthesis method resulted in particles with smaller average size, roundness, and circularity ratios. Symbols and abbreviations: HAM, Honey samples collected from Apis mellifera honey honeycombs; HSB, Honey samples collected from Scaptotrigona bipunctata honeycombs; L, microcompound obtained by light exposure; Lδ, microcompound obtained by light exposure and heating; pH 10, microcompound obtained by "aggregating basification"; pH 10δ, microcompound obtained by "aggregating basification" and heating; pH 5; microcompound obtained by "basification"; pH 5δ, microcompound obtained by "basification" and heating.
ly higher roundness ratio than the corresponding HAM pH 10, particles inside HAM L were overall rounder than those in HSB L (Table 5). Therefore, honey-mediated synthesis of microcompounds containing silver nanoparticles can be optimized the choices of honey sample, synthesis method, and post-synthesis treatments like heating.
HAM L and HSB pH 10 showed the highest roundness values among the HAM and HSB groups, respectively, while particles inside HAM pH 5δ and HSB Lδ the presented the lowest ones. HAM L particles had significantly higher roundness ratios (p < 0.05) than HAM pH 5's, and HAM pH 10 particles do not have a significantly different roundness (p < 0.05) than either. Meanwhile, among the particles in HSB microcompounds, HSB pH 5δ, HSB pH 10 and HSB pH 10δ were significantly lower (p < 0.05) than all others in HSB. In relation to the "honey sample" used, HAM L and HAM Lδ presented significantly (p < 0.05) higher "roundness" than HSB L and HSB Lδ, respectively. As for the circularity of particles, heating seems to result in more circular HSB nanoparticles (Table 5). Table 6 displays Minimal Inhibitory Concentrations (MICs)-estimated by Ag concentration obtained through atomic absorption spectroscopy -of washed HAM and HSB microcompounds, against S. aureus ATCC 29231 and E. coli ATCC 8739, and compared against microcompound MICs (without centrifugation) described in a previous article [46] .

MIC Assays of before and after Centrifugation
MIC of all non-centrifuged microcompounds in a previous analysis [46] varied between 15 µM and 250 µM (or 1.62-27 μg·mL -1 ) and were similar to MIC findings described in the literature, and thus a promising result [67][68][69][70][71][72][73][74][75][76][77] . However, MIC against E. coli and S. aureus strains increased for most types of microcompounds centrifuged from the honey solution (Table 6). This noticed increase alludes to the relevance of the honey present in the solution as an antibacterial factor, as described by literature [35][36][37][38][39][40] . HSB pH 10 and HSB pH 10δ MIC lowered after microcompound washing, which might be related to two compounded factors. HSB's main antibacterial effect (the oxygen peroxide Notes: Values are displayed as mean ± standard deviation. The diameter was calculated through the formula: A = π(diameter/2)². Ranks were detected by Dunn's multiple comparisons of HAM (or HSB) microcompounds and shown through letters. The "a" rank was attributed to the highest mean, and the other ranks were given letters based on significant difference, and means that share one letter are not significantly different. Symbols indicate significant differences (p < 0.05) detected by the Kruskal-Wallis test. Symbols and abbreviations: ⊕, Significantly greater value (p < 0.05) than corresponding HSB; ⊗, Significantly lower value (p < 0.05) than corresponding HSB microcompound; >, Significantly greater value (p < 0.05) than corresponding heated microcompound; <, Significantly lower value (p < 0.05) than corresponding heated microcompound; HAM, Microcompounds with honey samples collected from Apis mellifera honey honeycombs; HSB, Microcompounds with honey samples collected from Scaptotrigona bipunctata honeycombs; L, microcompound obtained by light exposure; Lδ, microcompound obtained by light exposure and heating; Micro., microcompound; pH 10, microcompound obtained by "aggregating basification"; pH 10δ, microcompound obtained by "aggregating basification" and heating; pH 5; microcompound obtained by "basification"; pH 5δ, microcompound obtained by "basification" and heating.
production by glucose oxidase) is reported by Brudzynski (2020) [78] to be compromised by high alkalinity; and the partial removal of stabilizing agents, such as gluconic acid, promoted an increase in the silver release [36,38,51,78] . Further studies are necessary in order to confirm this theory.  Tables 1A through 1F compare the antibacterial  potential of microcompounds by displaying the multiple  comparison rankings based on microcompound treatment  and time stamp; while supplementary Tables 2A, 2B, and 2C, meanwhile, show the comparison of the CFU re-duction (compared to the positive control) between each HAM microcompound and its corresponding HSB counterpart.

Time-kill Curve of Bacteria Exposed to Microcompounds
In general, the Gram-negative strain (E. coli ATCC 8739) demonstrated higher sensitivity than the Gram-positive, as observed in the significantly greater (p < 0.05) log10 reduction in assays involving E. coli ( Figures 3E  and 3F and suppl. Tables 1E and 1F). Against S. aureus ATCC 29231, HAM microcompounds presented a significantly lower log10 CFU than control (p < 0.05) only after 4 h ( Figure 3A, suppl. Table 1A). HSB microcompounds, in contrast, presented a lower log10 CFU than control after two hours in all cases (p < 0.05; Figure 3B, suppl. Table 1B) and thus a greater log reduction than HAM microcompounds (suppl. Table 2A). Even after halving treatment concentration for time-kill assays with E. coli, the Gram-negative strain still presented lower CFU counts at the last timestamps (10 h and 24 h).
which notes that the greater stiffness and extension wall of Gram-positive bacteria grant partial protection [69,77,79,80] . However the reducing and stabilizing agent, honey, has a greater effect against Gram-positive strains [12,36,38,[80][81][82] . Thus, the microcompounds described in this article show the potential to have a broader spectrum of activity due to the choice of stabilizing agent for the silver nanoparticles   Notes: Time-kill curves are displayed as mean ± standard deviation in Figure 3, and ranks attributed by Dunn's multiple comparison test in the supplementary material. Bacterial growth without exposure to any treatment is denominated control; and the minimal value of 1.0×10 2 was adopted due to methodological limitations. Graphs in Figure 3 are divided as follows: (A) Time-kill curves of Staphylococcus aureus ATCC 29231 exposed to HAM microcompounds at 125 μM. (B) Time-kill curves of Staphylococcus aureus ATCC 29231 exposed to HSB microcompounds at 125 μM. (C) Time-kill curves of Escherichia coli ATCC 8739 exposed to HAM microcompounds at 125 μM. (D) Time-kill curves of Escherichia coli ATCC 8739 exposed to HSB microcompounds at 125 μM. (E) Timekill curves of Escherichia coli ATCC 8739 exposed to HAM microcompounds at 62.5 μM. (F) Time-kill curves of Escherichia coli ATCC 8739 exposed to HSB microcompounds at 62.5 μM. Statistical analysis of CFU data is displayed in the "Supplementary Data" section. Abbreviations: AgNO 3 , Silver nitrate solution in different concentrations; HAM, Honey samples collected from Apis mellifera honey honeycombs; HSB, Honey samples collected from Scaptotrigona bipunctata honeycombs; L, microcompound obtained by light exposure; Lδ, microcompound obtained by light exposure and heating; pH 10, microcompound obtained by "aggregating basification"; pH 10δ, microcompound obtained by "aggregating basification" and heating; pH 5; microcompound obtained by "basification"; pH 5δ, microcompound obtained by "basification" and heating.
after 4 h at 125 μM ( Figures 3C and 3D; suppl. Tables 1C and 1D) warranted another analysis at 62.5 μM (Figures  3E and 3F; suppl. Tables 1E and 1F). Even then, microcompounds synthesized through light exposure (HAM L, HAM Lδ, HSB L, HSB Lδ) or regular basification (HAM pH 5, HAM pH 5δ, HSB pH 5, HSB pH 5δ) presented a bactericidal effect. HAM pH 10, HAM pH 10δ, HSB pH 10, and HSB pH 10δ, however, obtained by aggregating basification, did not perform well as the other microcompounds. They were only effective in containing the S. aureus and E. coli growth at 125 μM after 24 h, showing a significantly higher CFU at 24 h compared to the beginning of the assay. This difference can be related to both the high ionic force and surface charge presented by those microcompounds (Tables 2 and 3), given that the negative charge hampers electrostatic attraction to bacterial cells and the agglomeration of silver particles in highly alkaline environments goes against the release of silver ions (and thus the antibacterial effect of the particle) [9,14,59] .
As for the comparison in CFU count between honey samples, while HAM and HSB microcompounds presented antibacterial properties against E. coli and S. aureus, HSB microcompounds exhibited a more pronounced antibacterial effect against the S. aureus; while HAM microcompounds caused a more significant reduction in E. coli (Figure 3). This finding might be related to the antibacterial property HSB itself has, as data obtained from a previous study indicate this honey sample to be particularly effective against S. aureus [36] .
Other differences between HAM and HSB microcompounds might also stem from differences in the antibacterial effect of the particles themselves, as HAM microcompounds present a more intense bactericidal effect even in a concentration much lower than the established MIC for HAM [36] . Those findings further suggest the potential for exploration in studies of honey-microcompound interactions, given Rani, Rao, Shamili & Padmaja (2018) [83] have already described a silver-honey combination helpful in wound healing.
No microcompound was the best in all cases, nor was a "factor" in the microcompound production (honey sample used, synthesis method, or heat treatment) that improved the antibacterial effect in all cases. Among all HAM microcompounds, HAM pH 5δ and HAM pH 5 inhibited E. coli growth the fastest ( Figure 3C); and, among HSB microcompounds, the HSB L was particularly effective against the Gram-negative strain ( Figure 3D). However, against S. aureus, HAM Lδ and HSB pH 5δ presented the lowest log10 CFU counts among HAM and HSB microcompounds, respectively, after 24 h (Figures 3A and 3B).
Lastly, the antibacterial efficiency of HAM and HSB microcompounds is comparable to the antibacterial effect of other silver nanoparticles described in literature. Table  7 compares the concentration of silver nanoparticles in microcompounds which reached 100% antibacterial efficiency in our work with concentrations of other silver nanoparticle described in literature. Compared to current literature about the antibacterial effect, of silver nitrate and silver nanoparticles synthetized through other methods, silver nitrate had a slower effect (at a similar concentration), while our microcompounds with silver nanoparticles demonstrated were bactericidal at low concentrations, while they required more time for their antibacterial effect.
In summary, the results reveal that the differences between the antibacterial effects of microcompounds depend on the honey sample used, method of microcompound synthesis, and bacterial model, and that certain combinations, might result in microcompounds with antibacterial effects of particular potential. While this implies the necessity of more studies to assert conclusions, they also show how perspective the analysis of factors involved in the biogenic synthesis of silver microcompounds can be. Figure 4 contains SEM images captured at 6000× and 15000× of E. coli ATCC 8739 submitted to microcompound treatments for 4 h (compared to control). It also pinpoints morphological alterations through arrows in figure. Additionally, results concerning the average bacterial cell size, circularity, and roundness ratios were measured using ImageJ Photoshop, statistically analyzed by Dunn's multiple comparison test, and are exhibited in Table 8.

SEM Images and Measurements of E. coli Affected by Microcompound Treatments
The white arrows in Figure 4A showed treated E. coli cells smaller than control, and this difference is confirmed to be statistically significant (p < 0.05) in the analysis of measurements taken ( Table 8). The application of microcompound treatments also resulted in blebbing as well (light and dark blue arrows, correspondingly); however, its severity varied between treatment was applied, given the significantly lower circularity ratio (p < 0.05) of HAM pH 5δ and AgNO 3 cells. Cell filamentation and cell biofilm formation were observed (red and green arrows), with the latter being present particularly in E. coli treated with HAM pH 10δ and HSB pH 10. There was no significant difference (p > 0.05) in the roundness ratio of the measured cells (Figure 4).
The SEM micrographs in figures 4A and 4B exhibited several morphological alterations of E. coli cells esteeming from the application of microcompounds at 125 µM, which varied depending on the microcompound type used. It is noticeable that cell density is lower in the treated samples images than the control one, and the micrographs of antibacterial micromcompounds treatments (AgNO 3 , HAM pH 5δ, and HSB L) presented fewer bacterial cells than the bacteriostatic ones (HAM pH 10δ and HSB pH 10). Notably, after those treatments, bacterial cells presented severe blebbing (Figure 4, dark blue arrows) centered in the poles, similar to silver's effect in AgNO 3 and the oxidative stress caused by honey treatments [36,[89][90][91] . The average size of bacteria-treated microcompound cells was significantly lower (p < 0.05) than control (Figure 4, white arrows and Table 8). This effect of microcompounds has been observed in bacteria treated with peptidoglycan (ampicillin) and protein (erythromycin) synthesis inhibitors [89] .
Additionally, the elongation of treated E. coli cells observed in cells treated with HSB L, HAM pH 10δ, and HSB pH 10 ( Figure 4, red arrows), denominated "filamentation" in E. coli, is attributed to the inhibition of enzymes critical in septation (such as penicillin-binding protein 3) and to responses to DNA damage, mediated by the SOS pathway or not [89,[92][93][94][95] . In conjunction with other authors, Lee, Kim and Lee (2014) [96] repost that silver nanoparticles can cause E. coli elongation by ROS-related DNA damage, which triggers an SOS response in the bacteria [96][97][98][99] . This corroborates our findings, as the use of HAM pH 10δ and HSB HAM treatments are not bactericidal at 125 µM ( Figures 3C and 3D) and thus allow in turn E. coli to respond to their effects, which results in filamentation ( Figure 4).  Notes: Micrographs in Figure 4A and 4B were taken at 6000× and 15000×, respectively Notes: Micrographs in Figure 4A and 4B were taken at 6000× and 15000×, respectively Notes: Bacterial cells are named after the microcompounds treatment they were submitted to during the 4h of incubation. White arrows in image 8B show examples of bacterial cells smaller than those submitted to control conditions. The following morphological alterations were noted among different treatments: Severe and light blebbing (dark and light blue arrows); cell filamentation-i.e., abnormal growth-(red arrows); and biofilm formation (green arrows). Symbols and abbreviations: AgNO 3 , Silver nitrate; HAM, Microcompounds with honey samples collected from Apis mellifera honey honeycombs; HSB, Microcompounds with honey samples collected from Scaptotrigona bipunctata honeycombs; L, microcompound obtained by light exposure; Lδ, microcompound obtained by light exposure and heating; pH 10, microcompound obtained by "aggregating basification"; pH 10δ, microcompound obtained by "aggregating basification" and heating; pH 5; microcompound obtained by "basification"; pH 5δ, microcompound obtained by "basification" and heating. Lδ, microcompound obtained by light exposure and heating; pH 10, microcompound obtained by "aggregating basification"; pH 10δ, microcompound obtained by "aggregating basification" and heating; pH 5; microcompound obtained by "basification"; pH 5δ, microcompound obtained by "basification" and heating.
The honey samples present in the microcompoumds might also contribute to this effect; Brudzynski and Sjaarda (2014) [100] attribute it to Canadian honey samples, while this morphological alteration is not present among the bacteria treated with AgNO 3 or HAM pH 5δ due to how their metabolism comprise by the treatments applied. Lastly, elongation has also been related to an abnormal cell division caused by novobiocin [82,89,[101][102][103] . Thus, from our results in conjunction with established literature, it can be deduced that all microcompound treatments used can compromise bacterial multiplication by different mechanisms.
SEM images also showed biofilm production after HAM pH 10δ or HSB pH 10 treatments (Figure 4, green arrows). Though nanoparticles treatments present antibiofilm properties, this shows that conditions in microcompound synthesis can negatively affect this antibacterial mechanism [104] . Therefore, SEM imagining can be used to verify the presence of resistant mechanisms against silver microcompounds and to observe similarities between the effect of microcompounds and silver nanoparticles. Figure 5 summarizes the morphological alterations observed in E. coli treated with microcompounds and the proposed mechanisms involved based on literature.

Cytotoxicity of Cells Treated with Silver Nanoparticles
The polynomial regression curves in Figure 6 show the viability rates of LLC-MK2 cells treated with HAM and HSB silver microcompounds (compared to control), while Table 9 exhibits the CC50 and selectivity indexes of microcompound treatments. Honey excess was removed before the assays, as can interfere in MTT readings and present in vitro cytotoxicity at higher concentrations (even is though it is safely consumed) [105,106] .  The microcompound treatments did not compromise LLC-MK2 viability at the concentration 300 μM, higher than the antibacterial concentrations analyzed earlier.
Only HAM L showed significantly lower cell viability at that concentration (p < 0.05), and mammal cells were still viable even after applications of HSB L, HSB L, HSB pH 10, and HAM pH 10 and HAM pH 10δ at 2400 μM ( Figure  6). The relation between the microcompound synthesis method and the cytotoxicity of the resulting nanoparticles was dependant on the honey samples used. HSB pH 10 and HSB pH 10δ nanoparticles are more cytotoxic than their HAM counterparts, and HSB pH 10δ was more cytotoxic than HSB pH 10 ( Figure 6).
As for the selectivity ratios, the HAM pH 10, HAM pH 10δ, HSB pH 10, and HSB pH 10δ presented the highest values. The cytotoxicity intensity of silver nanoparticles noted in the studies described can be related to many factors.
For example, authors such as Akter et al. (2017) [107] , Albers et al. (2013) [108] and Gliga et al. (2014) [109] have reported that the cytotoxicity of silver nanoparticles is size-dependent and related to their ion release. In this case, the lower cytotoxicity can be associated with the lower bactericidal effect and high ionic force (which lowers Ag + release) described previously [9,14,59] .
effect as it is the case for nanoparticles. Studies regarding silver nanoparticle toxicity have reported a reduction in cell viability after silver nanoparticles application in concentrations as low as 5 μg·mL -1 , death of human skin epithelial cells at 50 μg·mL -1 to 100 μg·mL -1 ; to compromise human liver cell growth at 5 μg·mL -1 and to induce apoptosis these cells at 80 ugmL -1 [107][108][109][110][111][112] . Thus, the data indicate that the use of nanoparticles in microcompounds might be a safer alternative. This finding is further grounded in other works showing associations between nanoparticles and other materials to reduce the nanoparticle concentration necessary for a bactericidal effect and thus high selectivity ratios [113,114] .

Conclusions
In summary, silver microcompounds presented different characteristics due to the honey sample, method of biogenic synthesis used and application of a heat treatment. They presented similar Minimal Inhibitory Concentrations similar to smaller nanoparticles and much lower than concentrations found cytotoxic in our studies, as well as in literature [67][68][69][70][71][72][73][74][75][76][77][84][85][86][87][88][107][108][109][110][111][112] . Most microcompounds demonstrated themselves to be effective bactericidal agents at 125 µM. In particular, HSB microcompounds synthetized using sunlight were shown to have a selective bactericidal effect; while HAM microcompounds obtained by basification to pH 10 presented the capacity to control bacterial growth while not reducing mammalian cell viability at 2400 µM, which might be related to the excessive aggregation in highly alkaline solutions [9,14,59] . Additionally, our analysis has shown that there are many factors involved between the synthesis of microcompounds with honey and their application against bacterial infections; from the composition of the honey chosen to the target bacterial strain, which can impact the optimization of their usage. Nonetheless, as it is shown by investigating the results gathered with the current literature, there is much potential in the use of the antibacterial properties of silver microcompounds.