Antifungal Saponins from the Maya Medicinal Plant Cestrum schlechtendahlii G. Don (Solanaceae)

Bioassay‐guided fractionation of the crude extract (80% EtOH) of the leaves of Cestrum schlechtendahlii, a plant used by Q'eqchi' Maya healers for treatment of athlete's foot, resulted in the isolation and identification of two spirostanol saponins (1 and 2). Structure elucidation by MS, 1D‐NMR, and 2D‐NMR spectroscopic methods identified them to be the known saponin (25R)‐1β,2α‐dihydroxy‐5α‐spirostan‐3‐β‐yl‐O‐α‐l‐rhamnopyranosyl‐(1 → 2)‐β‐d‐galactopyranoside (1) and new saponin (25R)‐1β,2α‐dihydroxy‐5α‐spirostan‐3‐β‐yl‐O‐β‐d‐galactopyranoside (2). While 2 showed little or no antifungal activity at the highest concentration tested, 1 inhibited growth of Saccharomyces cerevisiae (minimum inhibitory concentration (MIC) of 15–25 μM), Candida albicans, Cryptococcus neoformans, and Fusarium graminearum (MIC of 132–198 μM). Copyright © 2015 John Wiley & Sons, Ltd.


INTRODUCTION
The Mesoamerican region of Central America is a biodiversity hotspot of semi-evergreen tropical trees (Myers et al., 2000). Indigenous Maya cultures in this region have strong traditions of healer apprenticeship and continue to use traditional medicines derived from local plants for primary health care. Ethnobotanical studies show that the Maya have an extensive knowledge of useful flora, and quantitative methodologies show that there is a high degree of consensus for many usage categories (Ankli et al., 1999;Amiguet et al., 2005;Bourbonnais-Spear et al., 2005).
The use of plants for infections is one area that shows a high degree of consensus. For dermatological infections, Heinrich (2000) reported an informant consensus factor of 0.52 among Yucatec Maya community. For the Q'eqchi, our own research calculated a consensus factor of 0.68 for treating infections using 96 plant species (Amiguet et al., 2005). The informant consensus factor is calculated by dividing the number of healers who use a plant for a specific ailment by the total number of healers (Amiguet et al., 2005). Laboratory studies have also shown that there is often a sound pharmacological basis for the use of many of these plants. In a lab study of 38 plant species used by the Tzeltal and Tzotzil Maya communities from Chiapas (Mexico), 65% of plants showed antimicrobial activity against Gram-positive Staphylococcus aureus, Gram-negative Escherichia coli bacteria, and the fungal pathogen Candida albicans (Meckes et al., 1995).
The Solanaceae family is of particular interest in the Maya traditional medicine for treatment of mycoses. For example, the Yucatec Maya community healers from San Jose Succotz Belize use Solanum torvum Sw. and Solanum mammosum L. for the treatment of athlete's foot (Arnason et al., 1980). Furthermore, leaves of Cestrum dumetorum Schlecht. are used by the Huastec Maya to treat warts and infected wounds (Alcorn, 1984). A pilot blinded clinical study on tinea pedis infections was undertaken by Lozoya et al. (1992) with Solanum chrysotrichum Schltdl., a plant widely used by Maya communities in Chiapas for severe athlete's foot. A cream containing a 5% MeOH extract of the leaves provided a statistically better response rate than the miconazole control with complete remission in 45% of cases. Further information on the active constituents and mode of action of these antimicrobial plants is required.
In the present study, we examined the antifungal activity of a Q'eqchi' Maya plant Cestrum schlechtendahlii G. Don. This plant is known as ik che (pepper tree) or ik kejen (pepper plant), and crushed leaves are used by the Q'eqchi' healers of Belize for fungal infections such as athlete's foot and thrush. Calderón et al. (2006) reported weak antileishmanial activity of C. schlechtendahlii leaves extract; however, no phytochemical work has been presented on this species. This is the first study to report the antifungal activity of this plant and its phytochemical constituents.

MATERIALS AND METHODS
General experimental procedures. Infrared spectra were recorded on a Shimadzu 8400-S FT/IR spectrometer. Optical rotations were registered on a Perkin-Elmer 241 digital polarimeter. NMR spectra were recorded on a Bruker Avance 500 spectrometer in CD 3 OD at 500 ( 1 H) and 125 MHz ( 13 C) using tetramethylsilane as an internal standard. High resolution electrospray ionization mass spectrometry (HRESIMS) was carried out using a Waters Xevo G2 UPLC-QTOF-MS/MS system. Electrospray ionization mass spectrometry (ESI-MS) was carried out also using a Shimadzu LCMS 2020 Series system. Open column chromatography was carried out with silica gel 60 (70-230 mesh, Merck). Thin layer chromatography (TLC) analyses were performed on silica gel 60F254 plates (Merck) with visualization using a ceric sulfate (10%) solution in H 2 SO 4 . For sugar analyses, silica gel 0.25 mm plates (Merck) were used, and visualization was carried out with an anisaldehyde reagent (5% panisaldehyde, 5% concentrated H 2 SO 4 in EtOH).
Plant material. Leaves of C. schlechtendahlii were collected in Jalacte, Belize in May 2010 under ethical approval (permits #H11-11-09, #H03-070-01). Collecting and export permits, as well as phytosanitary certificates, were obtained from the Belize Forest Department (ref. no. CD/60/3/08(33). Plant material was preserved in 70% ethanol immediately after collection. Voucher specimens (UOH19776) have been deposited at the University of Ottawa Herbarium, the Herbario Juvenal Valerio Rodriguez, and the Belize Forest Department.
Phytochemical analyses. Chromatographic analyses of the crude extract and fractions were performed on an Agilent 1100 series HPLC system comprising a quaternary pump, a degasser, an autosampler with 100 μL loop, a column thermostat, and a diode array detector (DAD). The identification of the phenolics was corroborated by comparing the retention time and maximum UV absorption values with authentic commercial standards (Sigma-Aldrich, St. Louis, MO, USA). The analyses were performed using a Luna C18 column (250 mm × 4.6 mm, 5 μm particle size) with column temperature set at 55°C and a flow rate of 1.5 mL/min. The mobile phase A was acetonitrile containing 0.05% trifluoroacetic acid, and B was water containing 0.05% trifluoroacetic acid. Optimized separation was achieved with the following method: initial conditions of 5% A and 95% B with an increasing gradient to 100% A in 25 min; the column was flushed with 100% A for 5 min and then set back to the initial conditions. DAD was set to monitor wavelengths 254, 280, and 330 nm. The active fractions were analyzed using a Shimadzu UPLC-PDA-MS system (Mandel Scientific Company Inc, Guelph, ON, Canada), which consisted of LC30AD pumps, a CTO20A column oven, a SIL-30 AC autosampler, and an LCMS-2020 mass spectrometer with an electrospray ionization source. Briefly, 1 μL of each fraction was injected through the autosampler to an Acquity CSH C18 column (100 × 2.1 mm, 1.7 μm particle size; Waters, Mississauga, ON, Canada) with an Acquity CSH C18 VanGuard Pre-column (5 × 2.1 mm). The mobile phases were H 2 O (A) and acetonitrile (B). The gradient elution method initiated with 5% B then increased to 95% B in 5 min. The column was then washed with 90% B for 3 min and changed back to the initial condition in 1 min. The flow rate was set at 0.8 mL/min with the column temperature set at 50°C. The photodiode array detector was set to monitoring wavelengths from 190 to 400 nm. The mass spectrometer with ESI interface was operating in positive and negative scan modes; the nebulizing gas flow was set at 1.5 L/min, and drying gas flow was at 10 L/min. The desolvation line temperature and heat block temperature were set at 300 and 450°C, respectively. The m/z range of both positive and negative scan is from 150 to 600 with 938 u/s scan speed.
Antifungal disk diffusion assay. Saccharomyces cerevisiae S288C, C. neoformans, and C. albicans D10 were cultured in Sabouraud dextrose broth (Difco) at 30°C. Berberine (95%, Sigma-Aldrich, St. Louis, MO, USA) and ketoconazole (>98%, Sigma-Aldrich, St. Louis, MO, USA) were used as antifungal positive controls and methanol as a negative control. Overnight cultures were grown to an optical density of 1.0 (~1.0 × 10 7 CFU/mL) at 600 nm and diluted 1:100. Aliquots (100 μL) of this inoculum were spread over the surface of Sabouraud agar plates. Paper disks (7.0 mm diameter) were loaded with crude extract (2 mg/disk), berberine (0.5 mg/disk), ketoconazole (90 μg/disk), fractions (0.5 mg/disk), saponin (0.5 mg/ disk), or methanol (carrier solvent) and allowed to air dry. The amended disks were placed treated side down on the prepared medium and incubated in the dark at 30°C for 48 h, prior to measurement of inhibition zones 441 ANTIFUNGAL SAPONINS FROM THE MAYA MEDICINE CESTRUM SCHLECHTENDAHLII with a ruler. All experiments were repeated three times, with three technical repetitions per biological repetition.
Yeast growth assay. This microdilution method was modified from CLSI M07-A9 (CLSI, 2012) using S. cerevisiae BY4741 and BY4743 cultured overnight in YPD broth (1% yeast extract, 2% peptone, 2% dextrose w/v) at 30°C and distributed into wells of a 96-well flat bottom plate (Costar 3596) at an inoculum size of 5-7 × 10 3 CFU/mL. Test compounds were dissolved at varying concentrations and serial diluted 1:1 across the plate. Wells containing berberine (95%, Sigma-Aldrich, St. Louis, MO, USA) or appropriate quantities of the MeOH carrier solvent were used as positive and negative controls, respectively. Hygromycin B (92%, VWR, Mississauga, ON, Canada) and geneticin (>98%, Sigma-Aldrich, St. Louis, MO, USA) were also used as positive controls. The plate was incubated at 30°C shaking at 600 rpm and absorbance readings taken at 600 nm every 10 min for 24 h (Biotek Powerwave XS2 microplate reader, Winooski, VT, USA). All experiments were repeated three times, with three technical repetitions per biological repetition.
Serial dot dilution assay. Overnight cultures of S. cerevisiae BY4741 and BY4743 grown in YPD medium at 30°C were adjusted to 1.5 × 10 6 CFU/mL and dilutions of 1:10, 1:100, and 1:1000 produced. In a six-well flat bottom plate (Falcon 3046), compound 1 at chosen concentrations was dissolved in YPD agar before solidification. Separate plates were prepared with berberine as a positive control. Once the media had set, 1 μL of each diluted inoculum was separately spotted twice into each well, corresponding to 1.5 × 10 3 , 150, 15, and 1.5 CFU/mL. The plate was then incubated in the dark at 30°C for up to 72 h. The plates were inspected for growth every 24 h.
All experiments were repeated three times, with three technical repetitions per biological repetition.
Fusarium growth assay. Fusarium graminearum was grown in 100 mL of CMC broth (Cappellini and Peterson, 1965) at 28°C with shaking for 3 to 5 days to generate conidia. Mycelium was separated from conidia by filtration through one layer of sterile miracloth. Conidia were then washed with sterile water twice by centrifugation at 4000 rpm for 15 min at room temperature and resuspended in sterile water for storage at 4°C. Conidia were inspected and counted with a hemocytometer prior to use. Conidia were diluted in GYEP broth tõ 1000 CFU/well in a 96-well flat bottom white plate (Costar 3632) and allowed to germinate for 24 h before compounds were added at varying concentrations (Nasmith et al., 2011). Wells containing berberine or appropriate quantities of the MeOH carrier solvent were also included to act as positive and negative controls, respectively. Hygromycin B and geneticin were also used as positive controls. Growth was monitored via fluorescence using a Polarstar Optima Microplate Reader (BMG Labtech, Gmbh, Offenberg, Germany) running FLUOstar OPTIMA Ver. 2.20R2. The plate was incubated at 28°C shaking at 600 rpm; readings (excitation/ emission at 485 nm/520 nm) were taken every 23 min for 72 h. All experiments were repeated three times, with three technical repetitions per biological repetition.

RESULTS
The leaf extract of C. schlechtendahlii (CSE) inhibited growth of all three yeast-like fungi (S. cerevisiae S288C, C. albicans D10, and C. neoformans) in the disk diffusion assay (Table 1). HPLC-DAD analysis of the crude extract identified the presence of caffeic acid (3), p-coumaric acid (4), and rosmarinic acid (5) (Fig. 1), none of which showed antifungal activity in our assays. The crude extract (yield of~13%) was then fractionated using silica gel open glass column chromatography and fractions CSE-XVII to CSE-XXII had antifungal activity (Table 1). Because CSE-XVIII (13.5 g) eluted with 80:20 EtOAc-MeOH was most active, this fraction was chromatographed using another silica gel column resulting in 39 secondary fractions (CSE-XVIII-1 to CSE-XVIII-39). From subfraction CSE-XVIII-13 eluted with 85:15 EtOAc-MeOH, compound 1 (800 mg) was obtained. All secondary fractions and 1 were tested in antifungal disk assays. CSE-XVIII-19 was further chromatographed with a Sephadex LH-20 column to yield compound 2 (3.4 mg). Because of the low yield of 2, this compound was not tested in the disk diffusion assay. A small set of MICs was carried out subsequently with compounds 1 and 2 ( Table 2).
The chemical composition of the active fraction CSE-XVIII was initially assessed by TLC on silica gel using as an eluent EtOAc:MeOH (80:20). The results of this analysis indicated the absence of UV active chromophores and the presence of oxidizable groups in the components of CSE-XVIII, shown by the formation of a broad band after the development of the plate with a ceric sulfate solution. Moreover, CSE-XVIII composition was also analyzed by HPLC-DAD. During the analysis, no signal was detected using any of the chosen monitoring UV wavelengths (210, 254, 280, and 330 nm), confirming that the compound has no UV absorption and supporting the initial TLC findings. This fraction was then analyzed using UPLC-MS, which showed the presence of three structurally related saponins, a major (pk 2) and two minor (pk 1 and pk 3) with masses of 756, 754, and 610, respectively (Fig. 2). Pk 2 and pk 3 were isolated and identified as compounds 1 and 2, respectively; their structures were verified via NMR. The structure of pk1 was tentatively identified; however, this compound was not isolated because of its low concentration in the plant extract. The purity of compounds 1 and 2 was estimated to be >98% via NMR and UPLC-MS.

DISCUSSION
Initial antifungal assays using disk diffusion assays revealed pronounced inhibition of S. cerevisiae, C. neoformans, and C. albicans. While not quantitative, the disk diffusion assay is an excellent technique for the preliminary identification of antimicrobials as it can detect inhibition by compounds of varying polarity and solubility (Meazza et al., 2003;Galván et al., 2008). The observed antifungal activity of compound 1 validates the traditional use of C. schlechtendahlii by the Q'eqchi' Maya healers of Belize. C. schlechtendahlii (in addition to S. torvum, S. chrysotrichum, and C. dumetorum) is used by Maya healers in the treatment of fungal infections. All of these Solanaceae species contain spirostanol saponins that may have potential applications as antifungal agents. Even though 1 has been isolated from another species in the Solanaceae by Haraguchi et al. (2000), this is the first report of antifungal activity by this compound. Other spirostanol saponins isolated from the Solanaceae as well as from other plant families have also been reported to have antifungal activities (Alvarez et al., 2001;Haraguchi et al., 2000;Shen et al., 2003;González et al., 2004;Yang et al., 2006;Lu et al., 2009). Our observation that compound 1 inhibits growth of the human pathogens C. albicans and C. neoformans suggests it has potential applications in treating mycoses such as ringworm, athletes foot, and onychomycoses fungal infections of the nails. The inhibition by compound 1 of C. albicans D10 is of particular interest because this clinical isolate is resistant to commercial antifungals such as amphotericin B and ketoconazole (Ficker et al. 2003). This indicates that the mode of action is not related to ergosterol biosynthesis and that 1 may be useful in treating recalcitrant mycoses.
Because 1 is also active against F. graminearum, a major cereal pathogen, further experiments could show potential agricultural applications as well. In support of this, the crude extract of Cestrum nocturnum L. leaves was previously reported by Hernández-Albíter et al. (2007) to inhibit the germination of conidiospores of the plant pathogenic fungus Colletotrichum gloeosporioides. Phytochemical analyses of C. nocturnum identified spirostanol saponins with three or more sugar moieties (Ahmad et al., 1995;Mimaki et al., 2001).
Since 1, 1A, and 2 have the same sapogenin, and compound 2 and 1A (sapogenin) showed little or no activity against S. cerevisiae, the presence of the two sugar moieties must be important for the antifungal activity of 1. Although 1A has not been tested previously, its relatively low antifungal activity is consistent with studies of diosgenin (a structurally similar sapogenin) that showed little or no antifungal activity (Imai et al., 1967;Chalenko et al., 1977). In a structure-activity relationship study by Yang et al. (2006), spirostanol saponins containing two or more sugar moieties exhibited inhibition of various pathogenic fungi, suggesting that the number of sugar residues present is important to the antifungal activity.
These results show that extracts of C. schlechtendahlii have antifungal activity and provide pharmacological validation for the traditional use of this plant by Q'eqchi' Maya healers of Belize to treat athlete's foot and other fungal infections. Further in-depth studies using other fungal species and mechanistic experiments would contribute to the assessment of the usefulness of saponins in the treatment of phytopathogenic infections and human mycoses.