Chemical Composition, Antifungal and Antioxidant Activities of Hedyosmum brasiliense Mart. ex Miq. (Chloranthaceae) Essential Oils

Background: Hedyosmum brasiliense Mart. ex Miq. (Chloranthaceae) is a dioecious shrub popularly used in Brazil to treat foot fungi and rheumatism. This work investigated the chemical composition, antifungal, and antioxidant activities of flowers and leaves of H. brasiliense essential oils; Methods: H. brasiliense male and female flowers and leaves were collected at Ilha do Cardoso (São Paulo) and the essential oils were extracted by hydrodistillation and analyzed by GC/MS and their similarity compared by Principal Component Analysis. Antifungal activity was performed by bioautography and antioxidant potential by 2,2-diphenyl-2-picrylhydrazyl hydrate (DPPH) free radical scavenging and β-carotene/linoleic acid system; Results: The major compounds for all oils were sabinene, curzerene, and carotol, but some differences in their chemical composition were discriminated by Principal Component Analysis (PCA) analysis. Bioautography showed two antifungal bands at Rf’s 0.67 and 0.12 in all samples, the first one was identified as curzerene. The oils presented stronger antioxidant potential in β-carotene/linoleic acid bioassay, with IC50’s from 80 to 180 μg/mL, than in DPPH assay, with IC50’s from 2516.18 to 3783.49 μg/mL; Conclusions: These results suggested that curzerene might be responsible for the antifungal activity of H. brasiliense essential oils. Besides, these essential oils exhibited potential to prevent lipoperoxidation, but they have a weak radical scavenger activity.


Introduction
The family Chloranthaceae includes plants grouped in four genera: Sarcandra Gardner, Chloranthus Sw., Ascarina J.R. Forst. and G. Forst., and Hedyosmum Sw. The genus Hedyosmum comprises 40 species of shrubs and small trees distributed from Mexico, throughout Central America to Bolivia, east of the Guianas and the Antilles [1]. The name Hedyosmum comes from the Greek "pleasant fragrance" and alludes to one of the most remarkable characteristics of these plants, the pleasant aromatic smell coming from all its parts [1,2]. According to Todzia [1], different  Burger) have previously been described [3][4][5][6][7][8], as well as their antioxidant and cytotoxic properties [3,4].
Hedyosmum brasiliense Mart. ex Miq., popularly known as "chá de bugre", "cidreira", or "cidrão", is a dioecious shrub distributed throughout central and southern Brazilian regions and western Paraguay [9,10]. Its leaves have been used in folk medicine to treat migraine, ovarium dysfunction, foot fungi, rheumatism, stomach pain, and as a diuretic [1,2]. Chemical and pharmacological studies with the aerial parts of H. brasiliense indicated analgesic, anxiolytic, antidepressive, and aphrodisiac effects for the crude extracts and for some isolated sesquiterpene lactones [11][12][13][14][15][16]. There are also some reports on the antimicrobial activity for the ethanol extract [17] and the essential oils [18,19], however none of them could attribute the observed bioactivity to a specific compound. The essential oil presented as major compounds specimen sabinene, (Z)-β-guaiene and pinocarvone, for a São Paulo State [18], and pinocarvone, curzerene and carotol, for one from Santa Catarina State [19]. To the best of our knowledge, this is the first report on the chemical composition and biological activities of H. brasiliense flower essential oils. In addition, this is also the first time that the antioxidant activity is described for this species. Therefore, this work aims to investigate the chemical composition of the essential oils from H. brasiliense male and female flowers and leaves collected at Ilha do Cardoso (São Paulo, Brazil) and to compare their antifungal and antioxidant activities, as well as to identify the main antifungal constituents in the essential oils.

Plant Material
Hedyosmum brasiliense Mart. ex Miq. (Chloranthaceae) flowers and leaves were collected from male and female specimens at Parque Estadual da Ilha do Cardoso, São Paulo, Brazil (25 • 05 S and 47 • 55 W, 14 m alt) in September of 2015. The taxonomic identity was confirmed by Dr. Inês Cordeiro (Instituto de Botânica, São Paulo, Brazil). The voucher specimens were deposited in the Herbarium at the Instituto de Botânica, São Paulo, Brazil, with the accession No. SP 475335.

Extraction of the Essential Oils
Nine male and female specimens were collected. Leaves and flowers were pooled to provide representative homogeneous samples of the population, separated into three replicates and stored under refrigeration (−22 • C) until extraction. Essential oil was obtained by hydrodistillation for 3.5 h in a Clevenger-type apparatus. The crude oil was separated, dried over anhydrous sodium sulfate and stored in a glass flask at −22 • C until GC-MS analysis and biological activities. The oil yields were calculated based on the oil and fresh plant material weights as mean ± standard deviation of the triplicates [18].

In Vitro Antioxidant Activities
The antioxidant activity was performed by the 2,2-diphenyl-2-picrylhydrazyl hydrate (DPPH) free radical scavenging method [26], adapted for microplates. 178 µL of the essential oils solubilized in methanol at concentrations between 4480-560 µg/mL were added to a 96-well microplate and 72 µL of DPPH (0.3 mmol/L) solubilized in methanol were added. Blank reading was performed with the test sample prior to the DPPH radical incubation. As a negative control, 178 µL of methanol was used. As positive controls, quercetin from 20 to 0.625 µg/mL and Ginkgo biloba extract (Herbarium Laboratório Botânico S.A., Paraná, Brazil) from 40 to 1.25 µg/mL were used. The microplate was incubated in the dark for 30 min at room temperature. Then, the absorbance was measured at a wavelength of 518 nm using a multi-well scanning spectrophotometer (BIOTEK KC4, Winooski, VT, USA). The absorbance was converted to percentage of antioxidant activity (AA) using the formula AA% = 100 − {[(Abs sample − Abs blank ) × 100]/Abs control }. All samples were tested in triplicate and the results are expressed by mean ± standard deviation. The amount of essential oil required to reduce the initial DPPH concentration in the reaction by 50% is referred to as the inhibitory concentration (IC 50 ).
The antioxidant activity was also evaluated by the β-carotene bleaching test, which is based on spectrophotometric measurements of the β-carotene oxidation induced by the lipoperoxidation products from linoleic acid [27] to determine the concentration providing 50% inhibition (IC 50 ). The essential oil (10 µL) solubilized in methanol at concentrations between 10,400-325 µg/mL was added in a 96-well microplate, together with 250 µL of a reactive solution composed by β-carotene and linoleic acid in water saturated with O 2 . Methanol (10 µL) was used as the negative control and as the positive controls, butylhydroxytoluene (BHT) and butylhydroxyanisole (BHA) at concentrations between 130-4.1 µg/mL. The microplate was incubated in the dark at 45 • C and subsequently the absorbance was read at 450 nm (multi-well scanner BIOTEK KC4, Winooski, VT, USA), immediately after adding the reactive solution and every 30 min for 2 h. The absorbance was converted to percentage of antioxidant activity (AA) using the AA% = 100 − {[(Abs final − Abs initial ) × 100]/Abs control }. All samples were tested in triplicate and the results are expressed by mean ± standard deviation.

Chemical Composition of the Essential Oil
The essential oil yields for male and female H. brasiliense flowers were 0.24 ± 0.01 and 0.38 ± 0.01%, respectively. The yield of leaf essential oils, both male and female, presented a yield of 0.33 ± 0.01% (Table 1).
Principal Component Analysis (PCA) data from Table 1 generated a plot, which condensed scores and loadings of the first two components, PC1 and PC2, which explain 61% and 34% of total variance ( Figure 1). Leaves presented higher amounts of sabinene and curzerene than flowers, while flowers presented higher amounts of carotol, as discriminated by PCA ( Figure 1). Furthermore, PCA analysis showed that female flowers and leaves presented higher amounts of 1,8-cineole than the male counterparts ( Figure 1).

Antifungal Activity
The essential oils of H. brasiliense presented clear inhibition zones corresponding to antifungal activity in the limit of detection against Cladosporium cladosporioides and C. sphaerospermum (Figure 2). Female flowers presented the strongest activity for both fungi from 200 μg until 25 μg of essential oil tested, but the other oils presented weaker inhibition, up to 50 μg against C. sphaerospermum and to 25 μg against C. cladosporioides.

Antifungal Activity
The essential oils of H. brasiliense presented clear inhibition zones corresponding to antifungal activity in the limit of detection against Cladosporium cladosporioides and C. sphaerospermum (Figure 2). Female flowers presented the strongest activity for both fungi from 200 µg until 25 µg of essential oil tested, but the other oils presented weaker inhibition, up to 50 µg against C. sphaerospermum and to 25 µg against C. cladosporioides.

Antifungal Activity
The essential oils of H. brasiliense presented clear inhibition zones corresponding to antifungal activity in the limit of detection against Cladosporium cladosporioides and C. sphaerospermum (Figure 2). Female flowers presented the strongest activity for both fungi from 200 μg until 25 μg of essential oil tested, but the other oils presented weaker inhibition, up to 50 μg against C. sphaerospermum and to 25 μg against C. cladosporioides.  After TLC elution of the most active oils, two inhibition bands, at retention factors (R f 's) of 0.67 and 0.12, were observed in all samples (Figure 3). After scraping the active bands, the GC-MS analysis indicated that curzerene (Table 2, Figure 4) was the main compound for the R f 0.67 fractions, comprising 97% and 82% for both flower and leaf oils. The band at R f 0.12 was composed by a complex mixture, with predominance of α-terpineol, α-eudesmol, and ferula lactone I (Figure 4) in different proportions, not allowing to point only one single active compound, as presented in Table 2. After TLC elution of the most active oils, two inhibition bands, at retention factors (Rf's) of 0.67 and 0.12, were observed in all samples (Figure 3). After scraping the active bands, the GC-MS analysis indicated that curzerene (Table 2, Figure 4) was the main compound for the Rf 0.67 fractions, comprising 97% and 82% for both flower and leaf oils. The band at Rf 0.12 was composed by a complex mixture, with predominance of α-terpineol, α-eudesmol, and ferula lactone I (Figure 4) in different proportions, not allowing to point only one single active compound, as presented in Table  2.

Discussion
Although there were previous studies on the essential oils from H. brasiliense leaves, none of them discriminated among male and female plants [18,19,28] and, to the best of our knowledge, this is the first report with flower essential oils. In our study, the essential oil yields were similar for flowers and leaves of H. brasiliense, with the male flowers presenting the lowest oil amounts. Regarding the leaf oil contents, the results were like those found for a Paraná State specimen (Brazil) [28], collected in an area not far from Ilha do Cardoso (São Paulo State, Brazil). However, these values were lower than those found for a specimen collected in Santa Catarina State (Brazil), whose leaves yielded 0.5% (v/w) [19], but this place is further south from our collection site.

Discussion
Although there were previous studies on the essential oils from H. brasiliense leaves, none of them discriminated among male and female plants [18,19,28] and, to the best of our knowledge, this is the first report with flower essential oils. In our study, the essential oil yields were similar for flowers and leaves of H. brasiliense, with the male flowers presenting the lowest oil amounts. Regarding the leaf oil contents, the results were like those found for a Paraná State specimen (Brazil) [28], collected in an area not far from Ilha do Cardoso (São Paulo State, Brazil). However, these values were lower than those found for a specimen collected in Santa Catarina State (Brazil), whose leaves yielded 0.5% (v/w) [19], but this place is further south from our collection site.
The essential oil analysis for the Parana (Brazil) specimen was not complete, most of the oil components were not identified [28]. Nonetheless, with the composition published it was possible to detect 23 common compounds with the São Paulo (Brazil) specimen, but only sabinene (7%) was a common major compound. However, for the Santa Catarina (Brazil) study were found 22 compounds in common [19], having as major compounds α-terpineol (10%), curzerene (9%), pinocarvone (8%), β-thujene (7%), and carotol (6%). Curzerene ( Figure 4A) was the only common major compound with the present analysis, but with 17-18% in the leaves and 11% in flowers.
The essential oils from other Hedyosmum species leaves (H. mexicanum, H. bonplandianum, H. arborescens, and H. angustifolium) have also presented high amounts of sabinene in their composition [4,7,8], as well as β-pinene, pinocarvone, and curzerene in the essential oil from H. colombianum leaves [6], similarly to our study.
PCA plot shows that leaves have a higher percentage of sabinene and curzerene than flowers, which have a more significant amount of carotol, responsible for differentiating them. Besides that, PCA was also able to discriminate male flowers and leaves from the female ones by the 1,8-cineole abundance, which was higher in the female counterparts (Table 1 and Figure 1). These results showed that it is important to analyze separately male and female individuals, part of the variable results on the chemical composition might be explained by not separating the plant genders.
Cladosporium spp. are phytopathogenic filamentous fungi usually chosen for bioautographic assays, as they present high sensitivity and permit the detection of fungitoxic substances by contrast with their dark color [23]. Preliminary assays for the detection limit with bioautography confirmed the antifungal activity for the H. brasiliense specimen from São Paulo against C. cladosporioides and C. sphaerospermum, with the highest activity for the female flower essential oil (Figure 2). In a previous study, the antimicrobial activity of H. brasiliense leaf essential oils has already been confirmed for Gram-positive bacteria and fungi such as the dermatophytes Microsporum canis, M. gypseum, Trichophyton mentagrophytes, and T. rubrum and the yeasts Candida albicans and C. parapsilosis [19].
The bioautography followed by TLC separation has been useful to determine antimicrobial activity of essential oils and to isolate the active compounds. A bioautography study with five aromatic plants (Thymus vulgaris, Lavandula angustifolia Chaix, Eucalyptus globulus Labill., Mentha spicata L., and Cinnamomum zeylanicum Blume) against five bacteria (Pseudomonas syringae, Xanthomonas campestris, Staphylococcus epidermidis, S. saprophyticus, and S. aureus) allowed to attribute the activity of Thymus vulgaris essential oil to thymol, comparing to a pure standard, as well as the activity of C. zeylanicum essential oil to eugenol, by the same principle [29]. A bioautography assay with Mentha x piperita L. essential oil against Candida albicans, followed by TLC separation of the active fraction, indicated that menthol was responsible for this activity [25]. Also, Guerrini et al. [3] detected α-cadinol, α-muurolol, τ-muurolol, and linalool in the active fraction with antibacterial activity against S. aureus using High Performance Thin Layer Chromatography (HPTLC) plates for separation of H. sprucei essential oil. In our study, the bioautography methodology indicated curzerene as the main compound for the active band found at R f 's 0.67 and α-terpineol, α-eudesmol and ferula lactone I for the active band at R f 0.12 ( Table 2). The H. brasiliense leaf essential oil collected at Santa Catarina (Brazil) presented antifungal activity and it was also rich in curzerene (8.9%) and α-terpineol (10.2%) (Figure 4 A,B) [25], supporting our results.
The antioxidant activity of essential oils is already well known [34][35][36][37]. According to Karadag et al. [38], the antioxidant potential is related to compounds capable of protecting a biological system against the potentially harmful effect of reactive oxygen species. The results of a single-assay can give only a narrow view of the essential oil antioxidant properties. Therefore, the antioxidant potential of H. brasiliense essential oils was determined by two different mechanisms of action: DPPH free radical scavenging, that works as electron transfer, and β-carotene/linoleic acid assay, measuring the suppression of lipoperoxides. In the DPPH assay, the IC 50 values for all the oils ranged from 2516.18 to 3783.49 µg/mL (Table 2), indicating a weak electron transfer capacity for the oils when compared to the positive controls quercetin and G. biloba extract, that presented IC 50 values of 2.5 and 13.5 µg/mL, respectively. These results are comparable to those obtained for Salvia tomentosa Mill. essential oils [36], which was rich in β-pinene (39.7%), α-pinene (10.9%) and camphor (9.7%), and it was not considered effective as antioxidant by DPPH and β-carotene/linoleic acid assay. Similarly, the essential oils of seven Artemisia spp. were tested by the same methods and showed weak activities, as their composition were markedly rich in non-phenolic components [39]. The antioxidant capacity of essential oils, by the DPPH method normally, is not very high when compared with that obtained for extracts and fractions rich in phenolic compounds. The major antioxidative plant phenolic compounds known are eugenol, carvacrol, thymol, menthol, and safrole, among others [34]. In a test with 98 pure essential oil components for their antioxidant activity, phenols, such as thymol and carvacrol, were the most active compounds, followed by alcohols mono-and sesquiterpenes and ketones α,β-insaturated, however hydrocarbons presented very low antioxidant activity [40]. On the other hand, our results indicated that the essential oils of H. brasiliense were much more active in the β-carotene bleaching assay, presenting IC 50 values from 71.12 to 180.71 µg/mL (Table 2), indicating their ability in destroying the conjugated diene hydroperoxides, the end products of linoleic acid peroxidation. In this way, the low radical scavenging activity observed for the H. brasiliense essential oils might be explained by the low amounts of phenolic compounds present in the oil, only 1.11 to 3.63% of thymol, and the high amounts of mono-and sesquiterpene hydrocarbons (Table 1). Since antioxidant activity of the entire oil is the result of the interaction of all constituents [41], it is hard to attribute the essential oils activity to a single compound, as other oil components may contribute exhibiting synergistic or antagonistic effects [42].

Conclusions
The leaf and flower essential oils were dominated by sabinene, 1,8-cineole, curzerene, and carotol. However, the contents of these constituents differed according to the plant part and gender.
This was the first report of antifungal activity of curzerene, explaining partially the activity of H. brasiliense essential oils by bioautography. Moreover, the complex mixture of α-terpineol, α-eudesmol, and ferula lactone I may have acted synergistically to the antifungal activity observed. Bioautography assay helped to detect antifungal fractions in a complex matrix of essential oils and guided to the isolation of target-directed constituents. Further studies are necessary to isolate higher amounts of pure curzerene, in order to test its limit of detection.
Antioxidant activity of essential oils from H. brasiliense was more effective by β-carotene discoloration assay than by DPPH assay, indicating their higher ability in quenching conjugated hydroperoxides than free radicals. The inhibition of lipid peroxidation may have been a result of the interaction of all constituents of the essential oils, since it was not possible to attribute antioxidant activity to a single compound.
Supplementary Materials: The following are available online at www.mdpi.com/2305-6320/4/3/55/s1, Figure S1: Mass spectrum of Non-identified compound 1 (N.I. 1) detected in the essential oil of H. brasiliense from Ilha do Cardoso; Figure S2: Mass spectrum of Non-identified compound 2 (N.I. 2) detected in the essential oil of H. brasiliense from Ilha do Cardoso; Figure S3: Mass spectrum of Non-identified compound 3 (N.I. 3) detected in the essential oil of H. brasiliense from Ilha do Cardoso; Figure S4: Mass spectrum of Non-identified compound 4 (N.I. 4) detected in the essential oil of H. brasiliense from Ilha do Cardoso.