Activity of crude extracts from Brazilian cerrado plants against clinically relevant Candida species

Background Medicinal plants have traditionally been used in many parts of the world as alternative medicine. Many extracts and essential oils isolated from plants have disclosed biological activity, justifying the investigation of their potential antimicrobial activity. In this study, the in vitro antifungal activity of six Brazilian Cerrado medicinal plant species were evaluated against clinically relevant Candida species. Methods The crude extract plants were evaluated against American Type Culture Collection (ATCC) standard strains of Candida spp. using disk diffusion method and determining the minimum inhibitory concentration (MIC). The chemical study results were confirmed by HPLC method. Results All six plant species showed antifungal activity. Among the species studied, Eugenia dysenterica and Pouteria ramiflora showed significant inhibitory activity against C. tropicalis at lowest MIC value of 125 and 500 μg/disc, respectively. The Eugenia dysenterica also disclosed MIC value of 125 μg/disc against C. famata, 250 μg/disc against C. krusei and 500 μg/disc against C. guilliermondii and C. parapsilosis. Pouteria torta, Bauhinia rufa, Erythroxylum daphnites and Erythroxylum subrotundum showed activity against the yeast strains with MIC value of 1000 μg/disc. The chemical study of the most bioactive extracts of Eugenia dysenterica and Pouteria ramiflora revealed catechin derivatives and flavonoids as main components. Conclusions All six evaluated plant species showed good antifungal potential against several Candida strains. However, E .dysenterica and P. ramiflora showed the higher inhibitory effect against the non-albicans Candida species. Our results may contribute to the continuing search of new natural occurring products with antifungal activity.


Background
The genus Candida, that includes about 200 different species, is a common commensal microorganism of the human skin and mucosal surfaces. However, some Candida species are opportunistic pathogens, capable of causing superficial or systemic infections, mostly in humans, mainly when debilitated or immunocompromised [1]. The systemic infections due to Candida species are severe infections that can cause high morbidity and mortality [2]. The invasive Candida infection represents the most common infection in hospitals worldwide and has an attributable mortality of 35 % to 60 %, causing public health and economic impact, because of high healthcare cost and prolonged length of hospitalization [3,4]. Among the pathogenic species, Candida albicans is the most common cause of invasive fungal infections in hospital settings, although a growing number of new cases from non-Candida albicans species have been detected. An even greater disparity occurred with Candida glabrata, which has been identified as the second most common cause of invasive fungal infections in North America and fourth in rank order in Latin America [5]. Candida tropicalis and Candida parapsilosis have shown increased prevalence worldwide, as rarer species such as Candida guilliermondii, Candida pelliculosa, Candida kefyr, Candida rugosa, and Candida famata have [6].
The limited number of antifungal agents presently available for the treatment of fungal infections support the need for more effective and less toxic treatment strategies in view of the progressive increase in the frequency of infections by non-albicans Candida species, the elevated resistance to antifungal drugs, and the relative toxicity of the antifungal drugs currently in use [7]. Besides, selection pressure due to the continuous exposure of the infectious agent to azoles seems to be the main cause behind the development of resistance to fluconazole that has been considered the drug of choice against Candida species. In spite of its fungistatic nature a prolonged use of this drug is not advisable considering the side effects, hence the increased demand for agents from natural resources as they could presumably act in a host friendly manner [8,9]. Medicinal plants have traditionally been used in many parts of the world as alternative treatments to traditional medicine [10]. Many extracts and essential oils isolated from plants have been shown to exert biological activity justifying the investigation of their potential antimicrobial activity [11,12]. Plants of the Cerrado, a tropical highland savanna covering almost 800,000 sq mi in the Midwestern region of Brazil have been broadly used in popular medicine [13]. This region is characterized by an enormous range of plants that have been the focus of several research reports [14,15] including the screening of several extracts and essential oils with potential anticandidiasis activity [16,17].
In line with these efforts to increasingly exploit the therapeutic potential of plants of this extremely biodiversity rich region in the present study, we tested the crude extracts of the following botanical species: Eugenia dysenterica DC. Mart. were collected from the Cerrado biome (the Brazilian savanna) in the city of Brasília (Federal District), located in the Midwestern region of Brazil, and its surroundings. Botanical identification was performed by Professors Suelí Maria Gomes and Christopher William Fagg. The voucher herbarium specimens were deposited at the Herbarium of the University of Brasília (UB) ( Table 1). All necessary permits were obtained for the described field studies.

Extraction procedures
Leaves were dried at room temperature and subsequently powdered in a knife mill (Marconi Laboratory Equipment, Piracicaba SP, Brazil). The ethanol and hexane crude extracts were obtained by the following way: plant material (40 g) was macerated at room temperature for seven days (repeated for three times), first with hexane (2 L), followed by ethanol (2 L). After filtration, the solvents were removed under reduced pressure, using rotary evap-

Screening of antifungal effect
For the experiment, 1000 μg of each extract were dissolved in 1 mL of the respective solvent (hexane, ethanol and distilled water) yielding final stock solutions with a concentration of 1000 μgmL −1 . Six millimeters filter paper discs were soaked with 10 μL of each extract stock solution hexanic, ethanolic and aqueous crude extracts and dried at room temperature. Agar Sabouraud 24 h cultures (35 ± 2°C) of Candida species were used in the preparation of an inoculum diluted in a 0.85 % saline solution, obtaining a fungal suspension equivalent to a 0.5 standard on the McFarland scale (1 × 10 6 to 5 × 10 6 cells per mL). On Petri dishes the fungal inoculum was spread over the surface of Mueller-Hinton + glucose (2 %) + methylene blue (0.5 μgmL −1 ) agar medium in which discs with the extracts and controls were subsequently deposited. The reading of the zone of inhibition was taken after 24 to 48 h of incubation at 35 ± 2°C [18,19]. Fluconazole (25 μg/disc) was used as positive reference standard and, as negative control filter paper discs with ethanol, distilled water and hexane.

Minimal Inhibitory Concentrations (MIC) test
Only extracts that disclosed inhibition zone against the different Candida species underwent the MIC evaluation in the disc-diffusion assay, this technique was adapted by Razmavar et al. [19]. The extract stock solution was twofold diluted by serial dilutions resulting in concentrations ranging from 500 to 15.6 μgmL −1 . After the serial dilutions, 10 μL of each dilution was soaked into 6 mm filter paper discs that were subsequently dried at room temperature. The fungal inoculum was spread over the surface of Mueller-Hinton + glucose (2 %) + methylene blue (0.5 μgmL −1 ) agar and discs were deposited in the culture medium. The plates were incubated at 35 ± 2°C for 24 to 48 h [19]. The MIC (minimum inhibitory concentration) was defined as the lowest concentration of the tested extracts that was able to inhibit any visible microbial growth [20]. Solvent free extract was used as negative control and Fluconazole (25 μg/disc) was used as positive reference standard.

Apparatus and chromatographic conditions
The extracts were Analyzed using LaChrom Elite HPLC system (Hitachi, Japan) liquid chromatograph equipped with L2130 pump, L2200 auto-sampler; L2300 column oven was in set at 25°C. The detector, L2455 DAD (Hitachi, Japan) was set at 280 nm. Separation was performed by Purospher Star reverse phase C18e column (5 μm, 150 mm × 4.6 mm i.d.) in combination with an appropriate guard column (4 × 4; 5 μm particle size) (Merck, Germany). The mobile phase was a linear solvent gradient system consisting of phosphoric acid (1 %) (A) and CH 3 CN (B), at a flow rate of 0.6 mL/min. Data acquisition was performed using EXChrom Elite software (version 3.3.2 SP1, Scientific Software. Inc.). The compounds present in the extract were characterized according to their UV-vis spectra and identified by their retention time in comparison with those of commercial standards (catechin, epicatechin, isoquercitrin myricetin, caffeic acid, chlorogenic acid, ferulic acid, kaempherol, rosmarinic acid, hesperentin, hesperidin, ellagic acid, gallic acid, hyperoside, quercetin, revesratrol, rutin, vitexin and isovitexin) [14].

Results and discussions
The antifungal activity of crude extracts of Eugenia dysenterica, Pouteria ramiflora, Pouteria torta, Bauhinia rufa, Erythroxylum subrotundum and Erythroxylum daphnites against ATCC standards strains of Candida spp. are shown in Table 2. The ethanol extract of Erythroxylum daphnites and all the hexane extract failed to show any inhibitory activity. However, all aqueous extracts and the remaining five ethanol extracts showed a variable inhibitory activity against one or more species of Candida. Among all the extracts that were assessed the aqueous extract of Eugenia dysenterica was the most effective. Except for C. albicans and C. glabatra in which no inhibition could be detected, this extract showed significant inhibitory activity against C. parapsilosis, C. guilliermondii, C. tropicalis, C. krusei and C. famata, with a zone of inhibition extending from 12 to 15 mm. The ethanolic extract Pouteria ramiflora also showed significant antifungal potential against C. glabrata, C. tropicalis and C. krusei, disclosing a zone of inhibition varying between 10 to 23 mm. Similarly, the ethanol extract of Pouteria torta and the aqueous extract of Erythroxylum subrotundum were effective against three Candida species and showed comparable antifungal activity against C. parapsilosis (10 and 11 mm), C. guilliermondii (10 mm) and C. glabrata (15 and 30 mm). Two species, C. tropicalis (12 mm) and C. glabrata (20 mm) were inhibited by aqueous extract of Pouteria ramiflora while the ethanolic extract of Erythroxylum subrotundum and the aqueous extract of Bauhinia rufa, revealed antifungal activity against C. parapsilosis (10 mm) and C. glabrata (20 to 25 mm). Finally, the aqueous extract of Erythroxylum daphnites and the ethanolic extract of Bauhinia rufa showed inhibitory effect only against C. glabrata (24 and 25 mm). In the present study, C. glabrata was the most sensitive specie to the inhibitory activity of the majority of the extracts, whereas C. albicans was the most resistant since none of the assessed extracts showed any effect on its growth. The other species of Candida showed variable degrees of resistance and sensitivity to different extracts that were tested.    Scarce previous reports exist on the antifungal activity of the medicinal plants hereby described. In support of our results Costa et al. [21] described absence of inhibitory activity of E. dysenterica essential oil against C. albicans although it showed effectiveness against the pathogenic yeast Cryptococcus. Boleti et al. [22] reported the antifungal potential of P. torta extract against different yeasts and filamentous fungi and considered that the action of the pouterin, a lectin-like protein found in its seeds, interacting with cell wall carbohydrates of the fungus, results in an inhibitory effect on the growth. As far as we know antifungal activity of P. ramiflora, E. subrotundum, E. daphnites and B. rufa has not be previously reported, although some studies have identified the presence of alkaloids and flavonoids in the genus Erythroxylum [23][24][25] which may partly explain its activity against Candida. Antifungal activity has also been reported for some species Bauhinia [26,27].
In view of the scarcity of data regarding the antifungal activity against Candida yeasts of the plants tested in the present study and in order to identify and confirm the possible medicinal potential of the extracts that had previously disclosed positive results in an initial screening, i.e. extracts that had exhibited an inhibition zone ≥ 10 mm against Candida species, we proceeded to the determination of their final minimal inhibitory concentration (MIC).   The results of MIC values from the diverse plant extracts tested are shown in Table 3. The results confirmed that the aqueous extract of E. dysenterica was the most effective against the different species of Candida. This extract exhibited respectively a MIC 125 μg/disc against C. tropicalis and C. famata, a MIC of 250 μg/ disc against C. krusei and a MIC of 500 μg/disc against C. parapsilosis and C. guilliermondii. On the other hand, the aqueous and ethanolic extract of P. ramiflora showed effectivity against C. tropicalis with a MIC of 500 μg/disc. A MIC of 1000 μg/disc was found on testing the remaining extracts.
In a second step, considering that the lower the MIC of a drug, the greater its effectiveness, the phytochemical composition of E. dysenterica and P. ramiflora extracts, that had showed a MIC < 1000 μg/ disc, was determined. The identification of the individual compounds responsible for the antifungal activity was performed by HPLC. The profiles of the aqueous and ethanolic extracts of P. ramiflora disclosed a large number of similar compounds with five main peaks (Fig. 1) [14]. The peaks 1-5 have characteristic UV/Vis spectra of flavonoids, with λ max between 340 and 370 nm ( Table 4). Identification of the peaks by comparison with commercial standards was not possible. Nevertheless, our P. ramiflora HPLC results showed similar UV/Vis peaks characteristic of the spectra of flavonoids as those found by Arapitsas [28]. Moreover, the peaks found in both extracts showed maximum absorption around 265 and 354 nm (Table 4) corroborating with maximum absorption of isorhamnetin derivatives described by Parejo et al. [29]. Flavonoids are the most studied groups of phenolic compounds, comprising molecules with established antifungal properties [30]. The study of compounds isolated from Calotropis procera Ait (Asclepiadaceae), ie the crude flavonoid fraction and flavonols (Quercetin-3-O-rutinoside, Kaempferol-3-Orutinoside, Isorhamnetin-3-O-rutinoside, 5-Hydroxy-3,7dimethoxyflavone-40-O-b-glucopyranoside), showed inhibitory effects of these compounds on C. tropicalis moreover against C. albicans [31].
On the other hand, the aqueous extract of E. dysenterica showed three main peaks, which have characteristic UV/Vis of catechin derivatives, with maximum absorbance near 280 nm and no absorption at 320 or 350 nm (Table 5, Fig. 2) [32]. Alves et al. [33] previously reported antifungal activity against species Candida of cathechin derivatives extracted from flowers of Castanea sativa, Filipendula ulmaria, Rosa micrantha and Cytius multiflorus, and from fresh leaves of Cistus ladanifer. The polyphenols (catechins and theaflavins) present in black tea also showed antifungal activity against species of Candida [34].

Conclusions
All six evaluated plant species showed good antifungal potential against several Candida strains. E. dysenterica and P. ramiflora showed the higher inhibitory effect against the non-albicans Candida species. The chemical analysis of E. dysenterica and P. ramiflora extracts disclosed presence of polyphenols (flavonoids and catechins), an important chemical class with antifungal activity. Considering the increasing number of infections by non-albicans Candida species and focusing on their importance for public health, our results may contribute to the continuing search of new natural occurring products with antifungal activity.