Profiling the Cymbopogon nardus Ethanol Extract and Its Antifungal Potential against Candida Species with Different Patterns of Resistance

The essential oil of Cymbopogon nardus, citronella, has been extensively studied. However, the chemical and biological properties of the ethanolic extract (EE) of C. nardus have not been evaluated. The aim of this study was to characterize the chemical composition of the EE of C. nardus and its active fraction (FrD). Moreover, the cytotoxic and antifungal properties of these extracts against Candida species with different resistance profiles to conventional drugs were evaluated. The compounds identified in EE were mono-Cand di-C-glycosyl flavones and phenylpropanoid glycosides. Phenylpropanoid glycosides were identified in FrD. EE showed antifungal activity, with minimum inhibitory concentration (MIC) values ranging from 62.5 to 500 μg mL. FrD was more effective against C. glabrata, as evidenced by the lowest MIC value (15.6 μg mL). EE inhibited yeast growth similar to amphotericin-B, as demonstrated by similar time-kill curves. EE inhibited C. albicans hyphae formation and mature biofilm of C. albicans, C. krusei and C. parapsilosis. The results of the chemical and biological analyses of EE and its fractions provided novel information and may contribute to control of infections caused by Candida species.


Introduction
The Cymbopogon genus is an important source of compounds with pharmacological properties. Cymbopogon nardus (L.) Rendle (Poaceae), commonly known as citronella, is native to Ceylon, and is cultivated in subtropical and tropical regions of Asia, Africa, and America. The essential oil and the ethanolic extract (EE) of citronella leaves have been traditionally used as insect repellents. Moreover, in Thailand, an infusion of citronella leaves is used to treat flatulence, dyspepsia, and abdominal cramps. 1 Some plant extracts have been evaluated against fungi 9,10 because of the presence of secondary metabolites with antimicrobial properties, such as phenols, flavonoids, terpenes, and alkaloids. 11 A study 12 described the in vitro fungistatic and fungicidal activity of a hydroethanolic extract of C. nardus against Microsporum canis and Trichophyton rubrum isolated from animals and the home environment.
Classical treatments for fungal infections include polyenes, azoles, and echinocandins. However, these treatments induce significant side effects. The side effects associated with synthetic antifungal agents promote aggravation of the disease state, since the side effects are typically related to hepatotoxicity and renal dysfunction. 13 Moreover, the lack of alternative treatment options is problematic in the case of drug resistance. 14 Candida has emerged as important species associated with opportunistic infections, resulting in a significant public health issue. 15 Several predisposing factors including immunodeficiency, antineoplastic therapy, organ transplantation, endocrine dysfunction, and prolonged antibiotic use increase susceptibility to Candida infection. 16 Candida spp. infections range from superficial infections, such as vulvovaginal candidiasis, esophageal or oropharyngeal candidiasis, and disseminated candidiasis. 17 Candida infections are associated with high morbidity and mortality rates in nosocomial bloodstream infections. 18 Several virulence factors associated with Candida spp. include morphological transition between yeast and hyphae, ability to defend against the host immune system, adhesion, biofilm formation, and production of harmful enzymes such as hydrolytic proteases, phospholipases, and hemolysin. 19 This study aimed to evaluate for the first time the chemical composition and antifungal activity of C. nardus against standard and clinical strains of Candida species with different biological virulence profiles and antifungal susceptibility.

Ethanolic extract (EE) preparation
Dried and powered leaves (500 g) were extracted by sonication in ethanol (99%) (Hexis, Jundiaí, São Paulo, Brazil) in four steps (2.0, 1.5, 1.5, and 0.5 L; 20 min per step) with occasional agitation. All extracted solutions were filtered, mixed, concentrated using a rotary evaporator, dried in a fume hood and then in a desiccator with silica gel. The yield of dried EE was 1.35%.

Determination of minimum inhibitory concentration (MIC)
The antifungal activity of EE was evaluated by determining the MIC using the microplate dilution technique according to the procedures described by Clinical and Laboratory Standards Institute (CLSI), 20 with modifications. Roswell Park Memorial Institute (RPMI)1640 medium (Sigma-Aldrich ® , Saint Louis, MO, USA) adjusted to pH 7.0 with MOPS (acid 3-[N-morpholino]propanesulfonic acid) buffer (Sigma-Aldrich ® , Saint Louis, MO, USA) was added to each well. Solutions of EE (0.1 mL) were added at concentrations ranging from 1000 to 7.8 µg mL -1 . A suspension (0.1 mL) containing 2.5 × 10 3 yeast mL -1 was added to each well. Amphotericin B (Sigma-Aldrich ® , Saint Louis, MO, USA) and fluconazole (Sigma-Aldrich ® , Saint Louis, MO, USA) were used as antimicrobial positive controls. Controls including culture medium, yeast growth, EE, and solvent were also prepared. The microplates were incubated at 37 °C for 48 h. After incubation, 20 µL of an aqueous 2% solution of 2,3,5-triphenyltetrazolium chloride (TTC) (Sigma-Aldrich ® , Saint Louis, MO, USA) were added, and the plates were incubated at 37 °C for 2 h. 7,21 All experiments were performed in triplicate.
The MIC results of EE strains were used to select the most sensitive strains (one ATCC and one clinical isolate for each species) for evaluation of antifungal activity as previously described. 20

Determination of minimum fungicidal concentration (MFC)
The MFC was determined by adding an aliquot from each well that showed antifungal activity to Petri dishes containing Sabouraud Dextrose Agar (SDA) (DIFCO  , Le Pont de Claix, France). These experiments were performed in triplicate. The MFC was defined as the lowest concentration of EE and Fr that resulted in no visible growth on the solid medium. 7,21 Inhibition of C. albicans hyphae formation C. albicans (ATCC 10231) was cultured for 24 h to obtain filamentous yeast. Then, the yeast was suspended at a concentration of 2.5 × 10³ cells mL -1 in phosphatebuffered saline (PBS, pH 7.2). Twenty microliters of this suspension were added to microplate wells containing RPMI 1640 medium (Sigma-Aldrich ® , Saint Louis, MO, USA) with 10% fetal bovine serum and 1% gentamicin. EE solution was evaluated at concentrations ranging from 1000 to 7.8 µg mL -1 . After 12 and 24 h, reductions in hyphal growth were visualized using an inverted light microscope (400×). Amphotericin B (Sigma-Aldrich ® , Saint Louis, MO, USA) (16 µg mL -1 ) was used as the positive control. Additional controls included fungal growth, solvent, sterile EE solution, and culture medium. 7

Time-kill assay
The time-kill assay was carried out according to Santos-Filho et al. 22

Cytotoxicity assay
To determine cytotoxicity, cells were collected using trypsin/ethylenediaminetetraacetic acid (EDTA) (Vitrocell ® , Campinas, São Paulo, Brazil), centrifuged (2,000 rpm for 5 min), and counted using a Neubauer chamber. The cell concentration was adjusted to 7.5 × 10 4 cells mL -1 in DMEM. Two hundred microliters of this suspension were plated in each well at 1.5 × 10 4 cells per well. The microplates were then incubated at 37 °C in a 5% CO 2 incubator for 24 h to facilitate cell adhesion. Serial dilutions of EE, FrC, and FrD were prepared to obtain concentrations ranging from 3.9 to 1000 µg mL -1 . Diluted solutions were added to the wells after removal of the incubation medium and non-adherent cells. The plates were then incubated for 24 h. Cytotoxicity was determined by addition of 30 µL of resazurin followed by a 6 h incubation period. The plates were analyzed using a microplate reader (BioTek ® , Winoosky, VT, USA) with excitation and emission wavelengths of 530 and 590 nm, respectively. The IC 50 was defined as the highest concentration of each fraction that resulted in at least 50% cell viability. All experiments were performed in triplicate. Five percent of dimethyl sulfoxide (DMSO) were used as the control. 24

UPLC-ESI-QTOF-MSE analysis
Compounds were identified based on retention time, fragmentation pattern, and accurate mass (chemical formula). Figure 1 shows a total ion chromatogram (TIC) of EE obtained in negative mode. Table 1 summarizes the identities of compounds 1-10 in EE as determined by mass spectrometry (high resolution MS and MS/MS n ). Data from the acquired spectra were compared with specialized literature data. [25][26][27][28][29] Based on these spectral comparisons, the following secondary metabolites were identified: two mono-C-(luteolin and apigenin derivatives) and two di-C-glycosyl flavones (luteolin derivatives), and six phenylpropanoid glycosides: three di-O-feruloyl-di-O-acetyl sucrose isomers (e.g., smiglaside A) and three di-O-feruloyl-tri-O-acetyl sucrose isomers (e.g., smiglaside C). The main peak observed in the TIC (t R = 6.90 min) corresponded to a di-O-feruloyl-tri-O-acetyl sucrose isomer.
The spectra of both mono-C-and di-C-glycosyl flavones ( Table 1, compounds 1-4) showed typical sugar moiety fragments resulting from cleavage of the C-hexosyl and which is typical of a hexose substitution in the aglycone moiety. These data supported assignments of luteolin-8-C-glucoside (orientin) for compound 3 and apigenin-8-C-glucoside (vitexin) for compound 4. 26,27 Flavonoid C-glycosylation has almost exclusively been found at positions 6 or 8 29 and according to a previous study 26 the relative intensities of the [M -H -90]fragment ions were 22 and 100% for orientin and isoorientin (luteolin-6-C-glucoside), respectively, and 8 and 42% for vitexin and isovitexin (apigenin-6-C-glucoside), respectively, supporting identification of orientin and vitexin in EE ( Table 1). The fragment ion signal at m/z 285 for compound 3 may correspond to kaempferol (flavonol) or luteolin (flavone) aglycone moieties (Yfragment ion), but data from previous   26 Therefore, the data from Table 1 suggest a 6-C-pentosyl-8-C-hexosyl substitution pattern for compound 1. As discussed for orientin, the presence of the ion fragment signals at m/z 133 for compounds 1 and 2 supports luteolin as the aglycone moiety.  Table 1 supported identification of these compounds as di-O-feruloyldi-O-acetyl sucrose isomers (e.g., smiglaside A). 28  feruloyl moiety 25 which supported identification of these compounds as di-O-feruloyl-tri-O-acetyl sucrose isomers (e.g., smiglaside C). 28 Previous studies have identified phenylpropanoid derivatives (e.g., chlorogenic acid and caffeic acid) and glycosyl flavones (e.g., isoorientin) in the Cymbopogon genus. These compounds are structurally and biosynthetically related to the compounds identified in the of EE C. nardus. However, no studies have evaluated non-volatile secondary metabolites in C. nardus. 30 FrD was the most active EE fraction against Candida strains. TIC data of FrD were compared with data from a previous study 28 and identification of the compounds in FrD was performed identically during evaluation of EE of C. nardus. The identified secondary metabolites included the same six phenylpropanoid glycosides found in EE: three di-O-feruloyl-di-O-acetyl sucrose isomers (e.g., smiglaside A) and three di-O-feruloyl-tri-O-acetyl sucrose isomers (e.g., smiglaside C). Fractionation resulted in collection of each of these compounds in FrD and the main peak observed in the TIC (t R = 6.90 min) corresponded to a di-O-feruloyl-tri-O-acetyl sucrose isomer, as observed in EE.

MIC and MFC determination of EE
The data showed that EE exhibited antifungal activity with MIC values ranging from 62.5 to 500 µg mL -1 , including in isolates resistant to fluconazole and amphotericin-B (Table S1, Supplementary Information (SI) section). The lowest MIC value (62.5 µg mL -1 ) in response to treatment with EE was observed for C. glabrata clinical isolates. EE showed a fungistatic profile against all tested isolates with MFC > 500 µg mL -1 . The solvent and growth controls produced satisfactory results.
These analyses showed that EE was active against all strains, except CK-ATCC and CO1. These results are very important, since EE was able to inhibit different species of Candida, including those resistant to fluconazole, the main antifungal agent used in medical practice. In addition, the results of MFC analysis showed that EE did not induce cell death, but only promoted growth inhibition. These results may be related to fungistatic mechanisms of action.
The antimicrobial potential exerted by plant extracts from the genus Cymbopogon has been observed previously. The study performed by Oloyede 32 evaluated the performance of the aqueous extract of the leaves of C. citratus against Escherichia coli, Staphylococcus aureus, Bacillus cereus, and Salmonella typhi, and showed excellent results.
The antimicrobial activity of C. nardus is reportedly due to properties of its essential oils. Previous studies demonstrated the antifungal potential of the essential oil of C. nardus against Candida species. 7,33,34 However, the antifungal properties and chemical composition of the EE of C. nardus have not been evaluated. Thus, this study was the first to evaluate these parameters, which may be of interest in the pharmaceutical and medical fields.
We highlighted the results obtained from testing of the C. glabrata species. The MIC values for this species were the lowest compared to those for the other strains evaluated in this study, and all strains of this species were resistant to fluconazole. C. glabrata is considered the secondmost pathogenic yeast that affects humans, after only C. albicans. 35 This species is directly involved in invasive fungal infections ranging from local to blood infections. In the case of systemic infections, treatment is challenging due to a dearth of therapeutic options. 19 The antimicrobial performance of products derived from medicinal plants may be explained by the presence of secondary metabolites. Previous studies have demonstrated the antimicrobial activity of secondary metabolites against different types of microorganisms. 11 The major classes of secondary metabolites are phenolic compounds, phenolic acids, quinones, saponins, flavonoids, tannins, phenazine, coumarins, lignans, neolignans, alkaloids, and terpenoids. 36 Chemical analysis of EE and FrD in this study showed the presence of C-and di-C-glycosylated flavones, and glycosylated phenylpropanoid derivatives, which directly exert antifungal activity. The antifungal activity of flavonoids in plant species has been studied extensively. Furthermore, glycosylated phenylpropanoids have been shown to inhibit the growth of several species of Candida. 37

MIC and MFC determination of the Fr
Only FrC and FrD exhibited antifungal activity against of the majority of the Candida strains. FrD showed the lowest MIC value (15.6 µg mL -1 ) against C. glabrata ATCC (Table S2, SI section). FrA, FrB, FrE, FrF, and FrG showed no antifungal activity, with MIC > 500 µg mL -1 .
Comparison between EE and each of the fractions of EE demonstrated that FrD was the most active against C. krusei, C. glabrata, C. tropicalis, and C. orthopsilosis. These results indicated that the phenylpropanoid glycosides identified in this study were the most likely substances responsible for EE antifungal activity, since they were concentrated in FrD after fractionation, and FrD exhibited greater anti-Candida activity than EE. The MIC of FrD against CG-ATCC (15.6 µg mL -1 ) was lower than that of EE (250 µg mL -1 ). Moreover, FrD exhibited greater activity against the C. glabrata clinical isolate (31.2 µg mL -1 ) than EE (62.5 µg mL -1 ).
The mechanisms of action of natural products vary. The cytoplasmic membrane is the most common site of action of secondary metabolites, with action on this structure resulting in extravasation of cellular contents and fungal death. The interaction with genetic material and protein synthesis is also a predisposing factor for promotion of therapeutic actions of natural products. Interaction of genetic material with secondary metabolites promotes changes in deoxyribonucleic acid (DNA), resulting in ineffective transcription, leading to aberrant cellular function and cell death. 36 Phenylpropanoid glycosides may act through formation of intramolecular interactions (for example, hydrogen bonding) which disrupts the physicochemical properties of the fungal cell, including membrane permeability, water solubility, and lipophilicity. 41 Inhibition of C. albicans hyphae formation EE was able to inhibit the transition of C. albicans from yeast to the hyphal form. Microscopic observation of EE-treated fungal cells demonstrated the absence of filamentous cells at concentrations ranging from 250 to 1000 µg mL -1 after 12 and 24 h (Figure 2).
These results are relevant to the pharmaceutical and medical fields because hyphae forming ability is the main risk factor during infections. 42 No previous studies have shown that EE of C. nardus can inhibit hyphae formation in C. albicans.
Several studies evaluating natural products observed prevention of hyphal development and proliferation of C. albicans. Chevalier et al. 43 evaluated the capacity of the aqueous extract of Solidago virgaurea to inhibit C. albicans (ATCC 10231) hyphae formation and showed that this extract inhibited hyphal proliferation.
The observation that EE inhibited hyphae formation suggested that this extract may act through control of yeast morphology, resulting in decreased proliferation, thus facilitating the activity of the active components present in the extract.

Time-kill assay
The effects of EE on Candida growth are shown in Figures 3-8. The results confirmed the fungistatic mechanism observed during evaluation of MFC, since treatment with EE led to reduced numbers of colonies compared to that with control treatments. Furthermore, EE showed activity similar to that of amphotericin-B against all tested strains.
EE inhibited growth of CG ATCC 2001, CG3, CP1, and CO ATCC 96141 to a greater extent at 48 h (final time) than amphotericin B. These results are important because they further demonstrated the inhibitory capacity of EE against different Candida species, especially the C. glabrata strains, which were fluconazole-resistant (MIC > 64 µg mL -1 ). The data found in this work corroborate with study developed by Toledo et al. 7 whereby the essential oil of C. nardus showed similar growth against the same of Candida species.

Effect of EE on mature biofilms of Candida species
The results showed that EE inhibited C. albicans, C. krusei, and C. parapsilosis mature biofilms. The concentrations (fifty times the MIC value) of EE able to eradicate mature biofilms of ATCC strains of C. albicans, C. krusei and C. parapsilosis were 25 mg mL -1 . For clinical strains of C. krusei (CK4) and C. parapsilosis (CP1) the inhibition concentration was 12.5 mg mL -1 of EE. EE showed no biofilm inhibition (> 6.25 mg mL -1 ) against strain of C. albicans (CA3). As this was the first report evaluating EE of C. nardus, these results are of great potential importance in the scientific field.
Biofilms represent a significant impact on public health, especially during establishment of chronic fungal diseases. 46 Biofilm formation is an important virulence factor associated with the Candida species, and treatments that prevent biofilms are limited because biofilms have a complex structure composed of polysaccharide extracellular matrix, which limits targeting of antifungal agents into biofilms. Furthermore, extensive communication among cells resulting in production of virulence-related molecules and the presence of a high fungal burden contribute the lack of efficacy of antifungal drugs. 45 Natural products have been shown to exhibit anti-biofilm potential. A study by Sangetha et al. 47 demonstrated that the methanolic extract of leaves of Cassia spectabilis inhibited C. albicans biofilm formation at 6.25 mg mL -1 . However, this extract did not effectively eliminate mature biofilms.
A study performed by Ramos et al. 48 showed that the methanolic extract of Syngonanthus nitens was ineffective against mature C. krusei biofilms. To overcome this issue, the authors employed a nanoparticle drug delivery system. Based on these findings, the 50 × MIC value observed in our study suggested a satisfactory inhibition profile.

Cytotoxic evaluation
The IC 50 values of EE and FrD are summarized in Table 2. Both EE and FrD exhibited higher IC 50 values against HepG2 cells than MRC-5 cells. The differences in responses between the cells could be related to the greater metabolic capacity of HepG2 cells, which mimic the metabolic status of human liver cells. These cells have the ability to retain the activities of various phase I and phase II enzymes which play important roles in elimination and detoxification of these classes of compounds in vivo. 49

Conclusions
In conclusion, EE from leaves of C. nardus contained compounds that exerted significant antifungal activity. The identified secondary metabolites in EE were phenolic compounds, including C-and di-C-glycosylated flavones, and glycosylated phenylpropanoid derivatives. These metabolites were abundant in FrD, and likely explained the antifungal potency of this fraction. Biological assays showed that EE exhibited activity against several strains of Candida species, including those resistant to fluconazole. Furthermore, EE was able to inhibit the main virulence factors associated with Candida species such as biofilms and C. albicans hyphae formation. In previous study, the essential oil of C. nardus showed superior activity than EE against the main virulence factors of Candida species. However, the EE exhibited lower MIC values than essential oil against planktonic Candida cells.

Supplementary Information
Supplementary data (tables) are available free of charge at http://jbcs.sbq.org.br as PDF file. > 1000 a IC 50 : concentration required to inhibit 50% of cell growth (average ± standard error (SE)); b dimethylsulfoxide. EE: ethanol extract; FrD: fraction D.