Influence of Eugenia uniflora Extract on Adhesion to Human Buccal Epithelial Cells, Biofilm Formation, and Cell Surface Hydrophobicity of Candida spp. from the Oral Cavity of Kidney Transplant Recipients

This study evaluated the influence of the extract of Eugenia uniflora in adhesion to human buccal epithelial cells (HBEC) biofilm formation and cell surface hydrophobicity (CSH) of Candida spp. isolated from the oral cavity of kidney transplant patients. To evaluate virulence attributes in vitro, nine yeasts were grown in the presence and absence of 1000 μg/mL of the extract. Adhesion was quantified using the number of Candida cells adhered to 150 HBEC determined by optical microscope. Biofilm formation was evaluated using two methodologies: XTT (2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide) and crystal violet assay, and further analyzed by electronic scan microscopy. CSH was quantified with the microbial adhesion to hydrocarbons test. We could detect that the extract of E. uniflora was able to reduce adhesion to HBEC and CSH for both Candida albicans and non-Candida albicans Candida species. We also observed a statistically significant reduced ability to form biofilms in biofilm-producing strains using both methods of quantification. However, two highly biofilm-producing strains of Candida tropicalis had a very large reduction in biofilm formation. This study reinforces the idea that besides growth inhibition, E. uniflora may interfere with the expression of some virulence factors of Candida spp. and may be possibly applied in the future as a novel antifungal agent.

Besides inhibition of growth, previous works performed by our group demonstrated the direct interference of E. uniflora components with Candida albicans expression of virulence factors in vitro and in vivo [26,27]. We found that this natural product greatly reduced hypha formation after morphogenesis induction in serum, Spider medium, and N-Acetyl D-Glucosamine (GlcNac) and impaired the ability to secrete phospholipase and proteinase. Oral candidiasis was attenuated in a murine model of infection and several proteins mainly related to cellular structure were differentially expressed in the presence of this natural product, as determined by proteomics analysis [27].
Considering our previous promising findings, we performed in vitro tests to examine the influence of the extract of the leaves of E. uniflora on the expression of some other important Candida spp. virulence factors in vitro, directly involved in oral candidiasis, including adhesion to HBECs, biofilm formation and cell surface hydrophobicity (CSH) of clinical isolates obtained from the oral cavity of kidney transplant patients in Brazil. This study will contribute for an increased knowledge upon the direct interaction of natural products with antifungal properties with the expression of virulence factors in vitro of a reasonable number of Candida different species.
This study aimed to evaluate the influence of the E. uniflora extract on the expression of some virulence factors in vitro of Candida species. In a previous study, Ferreira et al. [23] performed a screening of antifungal activities of medicinal plants from Northeast Brazil, including 30 different vegetal crude extracts and revealed E. uniflora as the most active natural product against Candida spp. Therefore, we selected the extract of this plant to perform the present study. The minimum inhibitory concentration (MIC) of E. uniflora extract was 312.5 µg/mL for C. albicans, C. dublinienis and C. parapsilosis complex species, while C. tropicalis and C. glabrata needed a slightly higher concentration (MIC equal to 625 µg/mL). Therefore, we decided to use a concentration higher than the MIC to perform the experiments, since we found that Candida cells are damaged but not completely unviable at this concentration. For this purpose, a concentration of 1000 µg/mL of the referred natural product was used to perform the in vitro virulence attributes tests in the presence of the extract, a lower concentration than the one used in our group previous publication, because it showed the same effect [26].
The chromatographic fingerprint of E. uniflora extract is shown in Figure 1. In the chromatographic profiling obtained with the E. uniflora extract, a peak correspondent to gallic acid was identified, eluting at a retention time (Rt) of 10.3 min; the presence of the majoritarian compound, the derivative flavonoidic myricitrin, eluted with Rt = 21.18 min. The contents of gallic acid and myricitrin were calculated based with the calibration curves of the respective standards, and the results found were 0.16 g% ± 0.0017 (1.42) and 0.43 g% ± 0.0006 (0.01), respectively. Much research has shown the anti-Candida activity potential of plant-derived compounds such as flavonoids and essential oils [28][29][30]. The antifungal action exerted by plant-derived compounds is related to the amount of the compound present in the extract under analysis [31]. The majoritarian compounds isolated from the E. uniflora extract used in our study presented gallic acid and myricitrin, and in this context several studies have demonstrated the anti-Candida action of these compounds with an MIC range similar to our results [31][32][33][34][35].
Considering the possible adverse effects associated with the use of antifungal agents, of which the action may target eukaryotic fungal cells, it is necessary to analyze the toxicity of natural products with possible antifungal activity [6,36]. Therefore, we performed a cytotoxicity assay of E. uniflora extract against human erythrocytes and HBEC. The possible hemolytic activity of the extract was determined by measuring the lysis of human red blood cells suspension in a spectrophotometric assay. In this experiment, Triton X-100 1% (v/v) was used as a positive control and induced total erythrocytes lysis. We could observe that the E. uniflora extract showed no significant effect on erythrocytes lysis (Figure 2), even with the double of the concentration used to perform all the experiments of the present study (2000 μg/mL; Figure 2). Hemolytic activity of Eugenia uniflora extract to human erythrocytes. Human erythrocytes (5 × 10 7 cells/mL) were incubated for 1 h at 37 °C with different extract concentrations. Triton-X 100 (1%) and washing buffer were used as positive and negative controls, respectively. The hemoglobin released from lysed erythrocytes was measured at 540 nm on a UV-Vis Spectrophotometer. The absorbance values for each sample were subtracted from the negative control and the hemolytic activity (%) was calculated. The experiment was performed in triplicate. The antifungal action exerted by plant-derived compounds is related to the amount of the compound present in the extract under analysis [31]. The majoritarian compounds isolated from the E. uniflora extract used in our study presented gallic acid and myricitrin, and in this context several studies have demonstrated the anti-Candida action of these compounds with an MIC range similar to our results [31][32][33][34][35].
Considering the possible adverse effects associated with the use of antifungal agents, of which the action may target eukaryotic fungal cells, it is necessary to analyze the toxicity of natural products with possible antifungal activity [6,36]. Therefore, we performed a cytotoxicity assay of E. uniflora extract against human erythrocytes and HBEC. The possible hemolytic activity of the extract was determined by measuring the lysis of human red blood cells suspension in a spectrophotometric assay. In this experiment, Triton X-100 1% (v/v) was used as a positive control and induced total erythrocytes lysis. We could observe that the E. uniflora extract showed no significant effect on erythrocytes lysis ( The antifungal action exerted by plant-derived compounds is related to the amount of the compound present in the extract under analysis [31]. The majoritarian compounds isolated from the E. uniflora extract used in our study presented gallic acid and myricitrin, and in this context several studies have demonstrated the anti-Candida action of these compounds with an MIC range similar to our results [31][32][33][34][35]. Considering the possible adverse effects associated with the use of antifungal agents, of which the action may target eukaryotic fungal cells, it is necessary to analyze the toxicity of natural products with possible antifungal activity [6,36]. Therefore, we performed a cytotoxicity assay of E. uniflora extract against human erythrocytes and HBEC. The possible hemolytic activity of the extract was determined by measuring the lysis of human red blood cells suspension in a spectrophotometric assay. In this experiment, Triton X-100 1% (v/v) was used as a positive control and induced total erythrocytes lysis. We could observe that the E. uniflora extract showed no significant effect on erythrocytes lysis (Figure 2), even with the double of the concentration used to perform all the experiments of the present study (2000 μg/mL; Figure 2). Hemolytic activity of Eugenia uniflora extract to human erythrocytes. Human erythrocytes (5 × 10 7 cells/mL) were incubated for 1 h at 37 °C with different extract concentrations. Triton-X 100 (1%) and washing buffer were used as positive and negative controls, respectively. The hemoglobin released from lysed erythrocytes was measured at 540 nm on a UV-Vis Spectrophotometer. The absorbance values for each sample were subtracted from the negative control and the hemolytic activity (%) was calculated. The experiment was performed in triplicate.

Figure 2.
Hemolytic activity of Eugenia uniflora extract to human erythrocytes. Human erythrocytes (5 × 10 7 cells/mL) were incubated for 1 h at 37 • C with different extract concentrations. Triton-X 100 (1%) and washing buffer were used as positive and negative controls, respectively. The hemoglobin released from lysed erythrocytes was measured at 540 nm on a UV-Vis Spectrophotometer. The absorbance values for each sample were subtracted from the negative control and the hemolytic activity (%) was calculated. The experiment was performed in triplicate.
Concerning cytotoxicity to HBEC, the assay was performed by placing the epithelial cells in contact with different concentrations of the extract and subsequently stained with trypan blue dye. Of note, some of the epithelial cells not incubated in the presence of the extract were also unviable. This finding is expected, because they were collect from a human volunteer by scrapping. The results showed that the majority of cells remained viable even after the exposure to different concentrations of the E. uniflora extract (Figure 3). These results, in combination with our previous study using A549 cell line [27] demonstrated that the E. uniflora extract showed to be a safe and nontoxic natural product at the concentrations tested. Concerning cytotoxicity to HBEC, the assay was performed by placing the epithelial cells in contact with different concentrations of the extract and subsequently stained with trypan blue dye. Of note, some of the epithelial cells not incubated in the presence of the extract were also unviable. This finding is expected, because they were collect from a human volunteer by scrapping. The results showed that the majority of cells remained viable even after the exposure to different concentrations of the E. uniflora extract (Figure 3). These results, in combination with our previous study using A549 cell line [27] demonstrated that the E. uniflora extract showed to be a safe and nontoxic natural product at the concentrations tested. We further investigated adhesion to HBEC by the different Candida spp. isolates by determining the number of blastoconidia of each strain adhered to 150 HBEC observed with optical microscopy. C. albicans and NCAC isolates were able to adhere even when previously grown in the presence of the E. uniflora extract. C. albicans was proven to be the most adherent species (mean C. albicans adherence 209 ± 75 vs. mean NCAC adherence 125 ± 76, p < 0.05). Nevertheless, we observed that adhesion of yeast cells to HBEC was strongly reduced when cells were previously incubated with the E. uniflora extract. Comparing the results of the adhesion ability of isolates of Candida spp. grown in the presence of E. uniflora extract with the control group (no treatment), a statistically significant reduction in adhesion for most of the isolates was observed ( Figure 4). Nevertheless, a large variation for all the strains regarding to the reduction of adhesion in the presence of the extract was observed when all the isolates were evaluated. We further investigated adhesion to HBEC by the different Candida spp. isolates by determining the number of blastoconidia of each strain adhered to 150 HBEC observed with optical microscopy. C. albicans and NCAC isolates were able to adhere even when previously grown in the presence of the E. uniflora extract. C. albicans was proven to be the most adherent species (mean C. albicans adherence 209 ± 75 vs. mean NCAC adherence 125 ± 76, p < 0.05). Nevertheless, we observed that adhesion of yeast cells to HBEC was strongly reduced when cells were previously incubated with the E. uniflora extract. Comparing the results of the adhesion ability of isolates of Candida spp. grown in the presence of E. uniflora extract with the control group (no treatment), a statistically significant reduction in adhesion for most of the isolates was observed ( Figure 4). Nevertheless, a large variation for all the strains regarding to the reduction of adhesion in the presence of the extract was observed when all the isolates were evaluated.  When we evaluated the percentage of reduction for each Candida species group separately, C. metapsilosis presented the greatest reduction, followed by C. tropicalis, C. orthopsilosis, C. albicans, C. glabrata, C. parapsilosis and C. dubliniensis. We also highlight strain 79 of C. orthopsilosis, which was the clinical isolate with highest percentage of reduction ( Figure 5). Nevertheless, a limitation in our study is the number of representative strains of each species. We have used strains obtained from a kidney transplant recipients cohort previous investigation from our group and the prevalence of C. albicans was higher than the other species, as described everywhere [37,38]. Therefore, for this reason, the number of NCAC strains is lower and definitely may have influenced our results. When we evaluated the percentage of reduction for each Candida species group separately, C. metapsilosis presented the greatest reduction, followed by C. tropicalis, C. orthopsilosis, C. albicans, C. glabrata, C. parapsilosis and C. dubliniensis. We also highlight strain 79 of C. orthopsilosis, which was the clinical isolate with highest percentage of reduction ( Figure 5). Nevertheless, a limitation in our study is the number of representative strains of each species. We have used strains obtained from a kidney transplant recipients cohort previous investigation from our group and the prevalence of C. albicans was higher than the other species, as described everywhere [37,38]. Therefore, for this reason, the number of NCAC strains is lower and definitely may have influenced our results. A comparison of the average values of the number of Candida cells adhered to 150 HBECs revealed a statistically significant reduction in adhesion for both C. albicans and NCAC species when evaluated separately, when previously grown in the presence of the extract (Table 1). To the best of our knowledge, there are no other publications reporting an impairment of adhesion of Candida spp. to HBEC due to the extract of E. uniflora. Our results showed a trend of a remarkable reduction in adhesion of Candida species when grown in the presence of this natural product, reinforcing that, besides growth inhibition, this extract might directly interfere on Candida spp. adhesion to buccal epithelia. Therefore, it could be particularly useful for prevention of oral candidiasis, because adhesion is also needed for colonization in the oral cavity, which is always considered a previous step to infection [39].
A few other authors have reported antiadhesion effects of other natural products in Candida species. Thaweboon et al. [40] evaluated the effect of Phyllanthus emblica Linn. on Candida adhesion to oral epithelium and denture acrylic. They have found that the ethanolic extract obtained from the fresh fruits of Phyllanthus emblica Linn. was able to reduce adhesion of clinical isolates of C. albicans to buccal epithelial cells.
De Paula et al. [41] showed that eugenol, which is the main active phenylpropanoid component of the essential oil from many aromatic plants, impaired the ability of C. albicans to adhere to Hep2 cells and polystyrene surface. Lopes et al. [42] also demonstrated that phlorotannins from the brown seaweed Fucus spiralis inhibited morphogenesis in C. albicans with the formation of pseudohyphae and limited ability to adhere to epithelial cells. Recently, Matsuura et al. [43] demonstrated that the A comparison of the average values of the number of Candida cells adhered to 150 HBECs revealed a statistically significant reduction in adhesion for both C. albicans and NCAC species when evaluated separately, when previously grown in the presence of the extract (Table 1). To the best of our knowledge, there are no other publications reporting an impairment of adhesion of Candida spp. to HBEC due to the extract of E. uniflora. Our results showed a trend of a remarkable reduction in adhesion of Candida species when grown in the presence of this natural product, reinforcing that, besides growth inhibition, this extract might directly interfere on Candida spp. adhesion to buccal epithelia. Therefore, it could be particularly useful for prevention of oral candidiasis, because adhesion is also needed for colonization in the oral cavity, which is always considered a previous step to infection [39].
A few other authors have reported antiadhesion effects of other natural products in Candida species. Thaweboon et al. [40] evaluated the effect of Phyllanthus emblica Linn. on Candida adhesion to oral epithelium and denture acrylic. They have found that the ethanolic extract obtained from the fresh fruits of Phyllanthus emblica Linn. was able to reduce adhesion of clinical isolates of C. albicans to buccal epithelial cells.
De Paula et al. [41] showed that eugenol, which is the main active phenylpropanoid component of the essential oil from many aromatic plants, impaired the ability of C. albicans to adhere to Hep2 cells and polystyrene surface. Lopes et al. [42] also demonstrated that phlorotannins from the brown seaweed Fucus spiralis inhibited morphogenesis in C. albicans with the formation of pseudohyphae and limited ability to adhere to epithelial cells. Recently, Matsuura et al. [43] demonstrated that the extract of Paulinia cupana (guaraná) was able to reduce adhesion of C. albicans to HBEC, even despite the fact it did not reduce fungal growth, emphasizing the direct interaction of a natural product in adhesion, suggesting its potential use to prevent oral candidiasis.
CSH properties of oral Candida strains of the present study are shown in Figure 6 and Table 2. Our hydrophobicity assay showed that E. uniflora extract induced a statistically significant change in CSH levels of all Candida species tested ( Figure 6). Most Candida species showed variable levels of affinity to the hydrocarbon phase. However, C. albicans strains presented a discrete higher affinity for the hydrocarbon phase than the other NCAC species. This finding may be influenced by the number of strains analyzed in each group. extract of Paulinia cupana (guaraná) was able to reduce adhesion of C. albicans to HBEC, even despite the fact it did not reduce fungal growth, emphasizing the direct interaction of a natural product in adhesion, suggesting its potential use to prevent oral candidiasis. CSH properties of oral Candida strains of the present study are shown in Figure 6 and Table 2. Our hydrophobicity assay showed that E. uniflora extract induced a statistically significant change in CSH levels of all Candida species tested ( Figure 6). Most Candida species showed variable levels of affinity to the hydrocarbon phase. However, C. albicans strains presented a discrete higher affinity for the hydrocarbon phase than the other NCAC species. This finding may be influenced by the number of strains analyzed in each group.  (b) CSH reduction of NCAC strains. Each bar represents mean ± SD of the value of adhesion of all the strains of each species. Experiments were performed in triplicate. * represents a statistically significant difference between the control and test experiment to each strain (Mann-Whitney test, p < 0.05). CSH is an intrinsic property of the external cell wall layer which makes Candida spp. cells close to each other, besides inducing aggregation. This force keeps optimal distance between adhesion molecules and host receptors, which lead to strong binding and finally irreversible adherence to mucosal membrane or other substrates. Adherence is an initial and important step for biofilm formation. Therefore, hydrophobicity might be a contributing factor to biofilm formation [44]. In addition, due to the fact that CSH of Candida spp. can affect cellular behavior as well as adhesion capacity, the reduction of the hydrophobic properties can lead to limitation in yeast colonization [45].
There are no published reports upon the action of the extract of E. uniflora on CSH of Candida spp. However, some studies evaluating the action of natural products in CSH have been performed. For instance, Zorić et al. [46] evaluated the action of oleuropein in the expression of C. albicans virulence factors. The authors showed that this compound interfered with CSH in C. albicans. Shirley et al. [47] determined in vitro effectiveness of Plantago major extract on the inhibition of C. albicans growth, biofilm formation, and CSH, which decreased at the highest concentrations tested.
Others studies have showed a positive effect on Candida spp. CSH, adhesion to epithelial cells and biofilm formation for a variety of naturally derived products or their constituents: eugenol [41], magnolol, and honokiol [48], the extracts of Brucea javanica and Piper betle [49].
Regarding to the investigation of biofilm formation, according to the Stepanovick et al. [50] classification, we found that 38% (16 strains) of all isolates analyzed were able to form biofilm. Among the clinical isolates of C. albicans, 62% (10 strains) were able to form biofilm, while 38% (6 strains) of the NCAC isolates belonging to C. glabrata, C. tropicalis and C. parapsilosis complex species could also express this important virulence factor in vitro ( Figure 7). Therefore, we only chose biofilm-forming strains to evaluate the effect of E. uniflora on this attribute of virulence of Candida spp.  CSH is an intrinsic property of the external cell wall layer which makes Candida spp. cells close to each other, besides inducing aggregation. This force keeps optimal distance between adhesion molecules and host receptors, which lead to strong binding and finally irreversible adherence to mucosal membrane or other substrates. Adherence is an initial and important step for biofilm formation. Therefore, hydrophobicity might be a contributing factor to biofilm formation [44]. In addition, due to the fact that CSH of Candida spp. can affect cellular behavior as well as adhesion capacity, the reduction of the hydrophobic properties can lead to limitation in yeast colonization [45].
There are no published reports upon the action of the extract of E. uniflora on CSH of Candida spp. However, some studies evaluating the action of natural products in CSH have been performed. For instance, Zorić et al. [46] evaluated the action of oleuropein in the expression of C. albicans virulence factors. The authors showed that this compound interfered with CSH in C. albicans. Shirley et al. [47] determined in vitro effectiveness of Plantago major extract on the inhibition of C. albicans growth, biofilm formation, and CSH, which decreased at the highest concentrations tested.
Others studies have showed a positive effect on Candida spp. CSH, adhesion to epithelial cells and biofilm formation for a variety of naturally derived products or their constituents: eugenol [41], magnolol, and honokiol [48], the extracts of Brucea javanica and Piper betle [49].
Regarding to the investigation of biofilm formation, according to the Stepanovick et al. [50] classification, we found that 38% (16 strains) of all isolates analyzed were able to form biofilm. Among the clinical isolates of C. albicans, 62% (10 strains) were able to form biofilm, while 38% (6 strains) of the NCAC isolates belonging to C. glabrata, C. tropicalis and C. parapsilosis complex species could also express this important virulence factor in vitro ( Figure 7). Therefore, we only chose biofilm-forming strains to evaluate the effect of E. uniflora on this attribute of virulence of Candida spp. Concerning to the effect of the E. uniflora extract on biofilm formation, Candida biofilms were quantified with two different methodologies: CV assay, which measures the total biomass of the biofilm, and the XTT assay, which measures biofilm metabolic activity. Of note, we found a positive correlation between the two methodologies used to quantify biofilm formation (r = 0.9698). We also observed a statistically significant reduced ability to fully express this virulence factor in vitro for all strains. The percentage of reduction ranged from 7 to 79% in isolates of C. albicans and from 3 to 95% in NCAC isolates (Figure 7a,b). This finding is particularly remarkable in our two C. tropicalis strong biofilm producers, which showed a notorious reduction in biofilm formation observed by the two methodologies tested in the presence of E. uniflora (Figure 7b).
There are actually a few published reports upon the action of E. uniflora on micro-organisms' biofilm formation. For instance, Oliveira et al. [51] showed that the hydroalcoholic extract of the fruit and infused leaves of this natural product has antibacterial activity against biofilm-forming bacteria. More recently, Zeuko'o et al. [52] analyzing the anti-Candida biofilm properties of Cameroonian plant extracts, observed that E. uniflora extract reduced biofilm formation of two clinical strains (C. albicans and C. glabrata) obtained from patients with vulvovaginal candidiasis. In fact, the number of studies investigating anti-biofilm properties of natural products in Candida spp. has steadily increased recently [47,53,54].
Most of the isolates that showed a significant reduction in biofilm formation belong to C. albicans species. However, some isolates of C. glabrata, C. tropicalis, and C. metapsilosis also showed impaired biofilm formation in the presence of E. uniflora extract. Considering that, even for commercially available antifungal drugs, each strain should be tested for MIC determination and that most of the anti-biofilm natural products published in the literature were tested in a very low number of strains, this finding cannot be neglected. Interestingly, we found a remarkable reduction of biofilm formation for the two strong biofilm producer strains of C. tropicalis proving that the vegetal extract may somehow interfere with this specific virulence factor. This finding is of great importance because, besides the fact that adhesion to buccal epithelia and CSH were reduced in the presence of the extract, biofilm formation is a potent virulence factor of Candida species, which confers resistance to antifungal therapies, limiting the penetration of substances through an extracellular matrix [7,55]. Concerning to the effect of the E. uniflora extract on biofilm formation, Candida biofilms were quantified with two different methodologies: CV assay, which measures the total biomass of the biofilm, and the XTT assay, which measures biofilm metabolic activity. Of note, we found a positive correlation between the two methodologies used to quantify biofilm formation (r = 0.9698). We also observed a statistically significant reduced ability to fully express this virulence factor in vitro for all strains. The percentage of reduction ranged from 7 to 79% in isolates of C. albicans and from 3 to 95% in NCAC isolates (Figure 7a,b). This finding is particularly remarkable in our two C. tropicalis strong biofilm producers, which showed a notorious reduction in biofilm formation observed by the two methodologies tested in the presence of E. uniflora (Figure 7b).
There are actually a few published reports upon the action of E. uniflora on micro-organisms' biofilm formation. For instance, Oliveira et al. [51] showed that the hydroalcoholic extract of the fruit and infused leaves of this natural product has antibacterial activity against biofilm-forming bacteria. More recently, Zeuko'o et al. [52] analyzing the anti-Candida biofilm properties of Cameroonian plant extracts, observed that E. uniflora extract reduced biofilm formation of two clinical strains (C. albicans and C. glabrata) obtained from patients with vulvovaginal candidiasis. In fact, the number of studies investigating anti-biofilm properties of natural products in Candida spp. has steadily increased recently [47,53,54].
Most of the isolates that showed a significant reduction in biofilm formation belong to C. albicans species. However, some isolates of C. glabrata, C. tropicalis, and C. metapsilosis also showed impaired biofilm formation in the presence of E. uniflora extract. Considering that, even for commercially available antifungal drugs, each strain should be tested for MIC determination and that most of the anti-biofilm natural products published in the literature were tested in a very low number of strains, this finding cannot be neglected. Interestingly, we found a remarkable reduction of biofilm formation for the two strong biofilm producer strains of C. tropicalis proving that the vegetal extract may somehow interfere with this specific virulence factor. This finding is of great importance because, besides the fact that adhesion to buccal epithelia and CSH were reduced in the presence of the extract, biofilm formation is a potent virulence factor of Candida species, which confers resistance to antifungal therapies, limiting the penetration of substances through an extracellular matrix [7,55].
The property of reduction of biofilm formation by E. uniflora extract on a strong biofilm forming C. tropicalis isolate (strain 77) was analyzed by scanning electron microscopy (SEM). We could confirm that, besides reducing the amount of blastoconidia and hyphae, there is also clear damage generated in the cellular structure in the presence of the extract (Figure 8). We can notice the presence of roughness and a coarse surface in the cell being indicative of a supposed action on the cell wall. The property of reduction of biofilm formation by E. uniflora extract on a strong biofilm forming C. tropicalis isolate (strain 77) was analyzed by scanning electron microscopy (SEM). We could confirm that, besides reducing the amount of blastoconidia and hyphae, there is also clear damage generated in the cellular structure in the presence of the extract (Figure 8). We can notice the presence of roughness and a coarse surface in the cell being indicative of a supposed action on the cell wall. If we consider that we have tested a higher number of strains than several other studies, this study showed that the extract of leaves of E. uniflora acted in an efficient way in reducing adhesion to HBEC, CSH and reducing biofilm formation. This is an interesting finding given that the process of Candida adhesion is the first step to biofilm formation, and the CSH is a property that exerts influence on adhesion [7]. Once again, it is important to emphasize that our highly biofilm-producing strains of C. tropicalis (strains 77 and 30LA) had extremely reduced biofilm formation in the presence of E. uniflora extract.
We next evaluated a possible correlation between the virulence factors determined in vitro using the Spearman coefficient. The results showed a moderate correlation between the adhesion capacity of HBEC and biofilm formation (crystal violet stained; Figure 9), meaning that there is a trend that the most adherent strains to the buccal epithelia are also moderate to strong biofilm producers, with remarkable production of exopolymeric matrix. Therefore, these results emphasize that E. uniflora may reduce the expression of these important virulence factors in Candida spp., specifically because adhesion is the first step for biofilm formation. If we consider that we have tested a higher number of strains than several other studies, this study showed that the extract of leaves of E. uniflora acted in an efficient way in reducing adhesion to HBEC, CSH and reducing biofilm formation. This is an interesting finding given that the process of Candida adhesion is the first step to biofilm formation, and the CSH is a property that exerts influence on adhesion [7]. Once again, it is important to emphasize that our highly biofilm-producing strains of C. tropicalis (strains 77 and 30LA) had extremely reduced biofilm formation in the presence of E. uniflora extract.
We next evaluated a possible correlation between the virulence factors determined in vitro using the Spearman coefficient. The results showed a moderate correlation between the adhesion capacity of HBEC and biofilm formation (crystal violet stained; Figure 9), meaning that there is a trend that the most adherent strains to the buccal epithelia are also moderate to strong biofilm producers, with remarkable production of exopolymeric matrix. Therefore, these results emphasize that E. uniflora may reduce the expression of these important virulence factors in Candida spp., specifically because adhesion is the first step for biofilm formation. Contradictory to these findings, we observed that there was no statistically significant correlation between CSH and the other virulence factors analyzed.
CSH is a factor related with fungal cell walls and is usually considered a good indicator of adhesion ability. Despite some earlier studies have indicated a positive correlation between CSH and adhesion to polystyrene, acrylic, and tissue surfaces, a few others suggested there is no correlation between adhesion to different substrates and CSH, meaning that adhesion of Candida to either biotic or abiotic surfaces is a complex mechanism [44,[56][57][58].
Candida cells with reduced CSH may also present impaired biofilm formation in some circumstances. However, this cannot be recognized as a general rule because meaningfully reduced ability of biofilm formation has also been observed for yeasts that did not show any changes in their hydrophobic properties [43,45].
The extract of E. uniflora used in our study showed in its chemical characterization considerable amounts of phenolic compounds, such as hydrolysable tannins and flavonoids (unpublished data). Similar to azoles, tannins derived from vegetable target lipids of host cell membranes [59]. Moreover, phenolic compounds impair growth and biofilm formation in C. albicans, possibly through the suppression of genes responsible for adhesion and morphogenesis [60]. The presence of phenolic compounds in the extract of E. uniflora may have led to a reduced ability to adhere to HBECs, CSH, and biofilm formation in some of the strains tested.
Taking the results from our previous publication, specifically the impairment of hypha formation and decreased recognition from phagocytic cells [27] and the findings of the present study, we hypothesize that the extract of E. uniflora interacts directly with the cell wall of Candida spp., because several adhesins, such as Hwp1, directly involved with biofilm formation [61] and adhesion [39] are present at this cellular component. When Candida cells were grown in the presence of the extract, accumulation of components of the natural product on the cell wall may have led to decreased adhesiveness, an important step also related to biofilm formation and CSH. Decreased cell fitness may have influenced the secretion of exopolymeric matrix, as observed with our biofilm formed on polystyrene surfaces, quantified with crystal violet staining and analyzed by SEM. We are currently working on the mechanisms of action of E. uniflora, including the investigation of the influence of cell wall perturbators (such as Congo Red, Calcofluor White, and Caffeine) and ergosterol quantification to confirm our hypothesis. Contradictory to these findings, we observed that there was no statistically significant correlation between CSH and the other virulence factors analyzed.
CSH is a factor related with fungal cell walls and is usually considered a good indicator of adhesion ability. Despite some earlier studies have indicated a positive correlation between CSH and adhesion to polystyrene, acrylic, and tissue surfaces, a few others suggested there is no correlation between adhesion to different substrates and CSH, meaning that adhesion of Candida to either biotic or abiotic surfaces is a complex mechanism [44,[56][57][58].
Candida cells with reduced CSH may also present impaired biofilm formation in some circumstances. However, this cannot be recognized as a general rule because meaningfully reduced ability of biofilm formation has also been observed for yeasts that did not show any changes in their hydrophobic properties [43,45].
The extract of E. uniflora used in our study showed in its chemical characterization considerable amounts of phenolic compounds, such as hydrolysable tannins and flavonoids (unpublished data). Similar to azoles, tannins derived from vegetable target lipids of host cell membranes [59]. Moreover, phenolic compounds impair growth and biofilm formation in C. albicans, possibly through the suppression of genes responsible for adhesion and morphogenesis [60]. The presence of phenolic compounds in the extract of E. uniflora may have led to a reduced ability to adhere to HBECs, CSH, and biofilm formation in some of the strains tested.
Taking the results from our previous publication, specifically the impairment of hypha formation and decreased recognition from phagocytic cells [27] and the findings of the present study, we hypothesize that the extract of E. uniflora interacts directly with the cell wall of Candida spp., because several adhesins, such as Hwp1, directly involved with biofilm formation [61] and adhesion [39] are present at this cellular component. When Candida cells were grown in the presence of the extract, accumulation of components of the natural product on the cell wall may have led to decreased adhesiveness, an important step also related to biofilm formation and CSH. Decreased cell fitness may have influenced the secretion of exopolymeric matrix, as observed with our biofilm formed on polystyrene surfaces, quantified with crystal violet staining and analyzed by SEM. We are currently working on the mechanisms of action of E. uniflora, including the investigation of the influence of cell wall perturbators (such as Congo Red, Calcofluor White, and Caffeine) and ergosterol quantification to confirm our hypothesis.

Plant Material and Eugenia uniflora Extract
The leaves of E. uniflora were collected in July 2009 at the city of Ceara-Mirim, in the northeast region of the State of Rio Grande do Norte, Brazil. Of note, we chose this time of the year for plant collection, because our rainy season (when the temperature is also lower in the South Hemisphere) is ideal for flowering, fruiting and fitness of E. uniflora. The plant was identified at the Herbarium of the Federal University of Rio Grande do Norte (Department of Botany, Zoology and Ecology, Biosciences Center) and a voucher specimen was deposited (No. 11763). The leaves of E. uniflora were dried at room temperature and ground into powder. The crude extract was obtained using 10 g of the powder with 100 mL of acetone: water (7:3, v/v) as a solvent by turbo extraction. The resulting extract was filtered and concentrated under reduced pressure (RV10 Basic, IKA ® , Campinas, Sao Paulo, Brazil) at 40 • C, 150 rpm. The concentrated extract was lyophilized to -64 • C, 0.006 mBar (Model 101, Liotop ® , Sao Carlos, São Paulo, Brazil), and the extract was stored at 4 • C.

High-Performance Liquid Chromatography (HPLC) Analysis of Eugenia uniflora Extract
The extract of E. uniflora (acetone: water, 7:3, v/v) was analyzed by HPLC through an Ultimate 3000 system (ThermoScientific ® , Waltham, Massachusetts, MA, USA). The chromatographic separation was performed using a C 18 column (Dionex ® ). Chromeleon 6.8 (Dionex ® , Sunnyvale, California, CA, USA) software was used for data acquisition and processing. The next steps were performed according to De Araújo et al [62]. The peaks of substances were identified through the retention times and UV spectra. Subsequently, the peaks were confirmed by spiking the sample with the standards.

Strain and Culture Conditions
We evaluated a total of 42 Candida spp. clinical isolates obtained from the oral cavity of kidney transplant recipients, belonging to the Medical and Molecular Mycology Laboratory, Clinical and Toxicological Analyses Department, at the Federal University of Rio Grande do Norte. The isolates were stored at -80 • C in YPD containing 20% glycerol. Cells were defrosted on ice and 100 µL of each strain was added to 5 ml of YPD liquid medium (dextrose 20 g/L, peptone 20 g/L, yeast extract 10 g/L) and incubated in a rotator shaker (TE-420, Tecnal ® Piracicaba, Brazil) at 35 • C, 200 rpm, for 48 h for reactivation and verification of cell viability. Subsequently, 100 µL of cell suspensions was inoculated on the surface of Sabouraud Dextrose Agar (SDA; Oxoid, Basingstoke, Hampshire, UK) using a Drigalsky loop. The strains were identified phenotypically and molecularly as described in Chaves et al. [2].

MIC of Eugenia uniflora Extract
MIC was based on Clinical protocol M27-A2 and Laboratory Standards Institute (CLSI) (Wayne, PA, USA) with adaptations for natural products [63]. The inoculum of all strains tested was obtained from 48 h cultivation in SDA at 35 • C and an initial suspension prepared according to McFarland scale 0.5 standard (1 a 5 × 10 6 cells). Then, two serial dilutions were made, the first in saline solution (1:50) and the second in Mueller-Hinton Growth medium (Difco) (1:30). After that, aliquots of 100 µL of the final inoculum solution were dispensed in microtiter plates of 96 wells containing 100 µL of various concentrations of the E. uniflora extract. The range tested was 10,000 µg/mL to 19.53 µg/mL. Finally, the plates were incubated at 37 • C and test reading taken after 48 h incubation. MIC was considered the lowest concentration of the E. uniflora extract capable of inhibiting 50% of the growth of each strain, taking as reference the respective positive control (treated in the same manner, but without the extract added to yeast cells).

Hemolytic Activity of Eugenia uniflora Extract on Human Erythrocytes
Hemolytic activity was performed according to  with some modifications. A healthy adult donor provided blood for the in vitro experiments. 5 mL of venous blood was collected and placed in EDTA tubes (ethylene-diamine-tetraacetate, Labtest, Lagoa Santa, Brazil) and promptly centrifuged (refrigerated centrifuge, ALC, Model PK121R, Milan, Italy) at 1100× g for 10 min at 4 • C. Plasma was carefully aspirated and the buffy coat was removed. The pellet containing erythrocytes was washed three times in phosphate-buffered saline (pH = 7, PBS) and suspended in PBS to a final suspension of 5 × 10 7 cells/mL counted with a hemocytometer. Two mL of cell suspension was incubated for 1 hour at 37 • C with different E. uniflora extract concentrations (312.5, 625, 1000, and 2000 µg/mL). Triton-X 100 (1%) and washing buffer were used as positive and negative controls, respectively. The amount of hemoglobin released from lysed erythrocytes was measured at 540 nm on a UV-Vis Spectrophotometer (Biochrom Libra S32, Cambrigde, UK). The absorbance values for each sample were subtracted from the negative control and the percentage of hemolytic activity (%) was calculated. The hemolytic activity (%) = (

Cytotoxicity Assay of E. uniflora Extract to HBEC
The cytotoxicity assay of E. uniflora extract to HBEC was performed according to Crowley et al. [65] with some modifications. HBECs were collected as described in Section 3.7.1. A suspension with 5.0 × 10 5 cells/mL was prepared and treated with different E. uniflora extract concentrations (312.5, 625, 1000, and 2000 µg/mL) during 1 h, at 37 • C, while positive control (viable cells) epithelial cells were treated by the same manner, except by the fact that they were not incubated in the presence of the extract. Cells treated with hydrogen peroxide (Sigma-Aldrich, St. Louis, MO, USA) constituted the negative control (dead cells). Cells were washed three times with PBS (pH = 7.0) after incubation and 50 µL of each tube was mixed with an equal volume of trypan blue (Sigma-Aldrich) solution (0.4%). The number of viable cells was determined by counting 150 HBEC with the operator blinded to the nature of the material on the slide. The percentage of cells viability was calculated using the following equation: %Viability = (Number of colorless cells/Total number of cells) × 100. The experiment was performed in triplicate.

Candida spp. Adherence to HBEC in the Presence of Eugenia uniflora
Candida cells were grown overnight to the stationary phase in NGY liquid medium (0.1% Neopeptone (Difco), 0.4% glucose, and 0.1% yeast extract (Difco)) with 1000 µg/mL of concentration of E. uniflora extract at 30 • C. For adhesion assay experiments, HBECs were collected from a unique healthy volunteer (not colonized by yeasts). This procedure was performed in the day of the experiment by gently rubbing a sterile swab on the mucosal surface of the cheeks. Subsequently, the sterile swab was placed in sterile conic tubes containing 5 mL sterile saline at 4 • C until the moment of the experiment. The mixtures containing a ratio of 10 yeast cells per HBEC were incubated at 37 • C for 45 min with shaking (200 rpm) and then cells were vortexed, formalin-fixed, vortexed again to remove nonadherent cells, and transferred to a microscope slide. The number of Candida cells adhering to 150 HBEC was determined with the operator blinded to the nature of the material on the slide [66]. Tests were done in triplicate. The same conditions, except with the absence of the extract added in the NGY liquid medium, were followed for control experiments. Of note, CFU counts after 48 h incubation on SDA plates were performed with Candida spp. cells grown in NGY both in the absence and presence of the extract, to guarantee that test and control experiments contained the same number of viable cells.

CSH
Candida isolates were cultured in YNB medium ("Yeast Nitrogen Base", Difco TM , Detroit, Michigan, MI, USA) with 50 mM of glucose (D-glucose monohidratada P.A., Cinética) for 18 h at 37 • C in a shaker (Tecnal, TE-420, SaoPaulo, Brazil) at 200 rpm, and absorbance was adjusted to OD 600nm = 1.0. For each strain, 2 mL of cell suspension was transferred to four glass tubes, representing one test group and one control group. An aliquot of 0.4 mL of xylene was added to each tube. The test and control tubes were incubated in a water bath (TE-056 Mag, Tecnal, São Paulo, Brazil) at 37 • C for 10 min, then shaken for 30 s, and returned to the water bath for another 30 min for a separation of the aqueous phase and of xylene. The aqueous phase was aspirated and transferred to a clean tube. The absorbance was measured at 520 nm. Hydrophobicity was given by the fraction [(Co − CH)/Co] × 100, where Co represents an OD 520nm of the control tube and CH represents OD 520nm of the test tube [44]. Each strain was tested in triplicate. For the test group cells, 1000 µg/mL of E. uniflora extract was added to the culture medium used to standardize the inoculum.

Candida spp. Biofilm Formation in the Presence of Eugenia uniflora
Biofilms were performed and stained with crystal violet according to Chaves et al. [2]. For the test group cells, 1000 µg/mL of E. uniflora extract was added to the culture medium for inoculum standardization. For the XTT assay, the biofilm-coated wells of microtiter plates were washed twice with 150 µL of PBS and 100 mL of the XTT/menadione (Sigma-Aldrich) solution (1 µM of menadione) was added to each well containing a prewashed biofilm as well as to negative control wells. The plates were covered with aluminum foil and incubated in the dark for 2-3 h at 37 • C. Subsequently, 75-80 mL of the solution was removed (resulting colored supernatant from each well) and transferred into the The preparation of the samples for the SEM documentation was performed from pure cultures of the yeasts, grown in 20 mL of YPD medium contained in 50 mL conical tubes in an orbital shaker (TE-420, Tecnal ® Piracicaba, Brazil) (150-180 rpm) at 30 • C for 18-20 h. Subsequently, cells were washed in PBS for three times at 2500 rpm for 5 min. Thereafter, cells were initially suspended in PBS and adjusted to a concentration of 1 × 10 6 /mL in RPMI, by the McFarland scale 0.5 standard. Next, 300 µL of cell suspension was dispensed into the wells of a 12-well microtiter plate (TPP, Zollstrasse, Trasadingen, Switzerland) in which a 2-mm-thick glass disk was deposited. Plates were incubated a 37 • C for 24 h to allow biofilm formation on glass disks, without shaking. After the incubation period, disks were washed with sterile PBS and fresh medium was added (control experiment). For the evaluation of the action of E. uniflora on biofilm formation, fresh medium with 1000 µg/mL of the natural product was added to the wells (test experiments). Plates containing both control and test experiments were incubated for another 48 h at 37 • C. Plates were washed thrice with PBS and the next steps were equally performed for either treated or untreated biofilms (control experiments). To this end, 1 mL of 3% glutaraldehyde and 2% paraformaldehyde in 0.1 M potassium phosphate buffer, pH 7.4 were added to each well for sample fixation. Subsequently, plates were subjected to six washes with pure buffer solution at 15 min intervals and after were fixed in 1% osmium tetroxide for 16 h. After the incubation period, the material was again subjected to six washes with the same buffer solution and subjected to dehydration with increased concentrations of ethyl alcohol (30%, 50%, 70%, 80%, 90%, 95%, 100%). The samples were then dried in a Critical Point Dryer (CPD-030, Balzers, Oberkochen, Germany), and mounted for gold plating (DENTON VACUM metallizer, DESK II model, Freehold, NJ, USA), analyzed, and electromicrographed in a scanning electron microscope (JEOL, model JSM 5410, Chiyoda, Tokyo, Japan) operated at 15 KV [68].

Statistical Analysis
Data were analyzed using the statistical software "GraphPad, Prism" version 5.0, (GraphPad Software, La Jolla, California, CA, USA). Results were presented as mean ± standard deviation, and differences were analyzed by the Mann-Whitney test, while the Spearman coefficient was used to assess the correlation between the different techniques used to evaluate the expression of virulence factors in vitro. For all the analyses, P was considered a default value of 0.05 and the confidence interval of 95%.

Conclusions
In conclusion, this was the first study to evaluate the effects of E. uniflora extract on adhesion to human buccal epithelia, CSH, and biofilm formation, which are important virulence factors of Candida spp. Our results showed that the extract interferes with adherence of Candida spp. to HEBC and CSH and it is able to reduce biofilm formation for biofilm producing strains. Of note, C. tropicalis highly biofilm-producing strains showed a remarkable reduction in biofilm formation caused by the extract. Taken together, this finding and our previous publication, which also showed limited filamentation and hydrolytic enzymes expression for C. albicans, demonstrated a direct influence of the referred natural product with the expression of virulence attributes of Candida spp. in vitro. Therefore, it may be a viable alternative for the use to prevent oral candidiasis, specifically in denture wearers. Our group is currently working on the elucidation of the chemical compounds responsible for antifungal properties and the mechanism of action for both growth inhibition and interaction with Candida spp. virulence factors in vitro of E. uniflora.