Study of seasonality and location effects on the chemical composition of essential oils from Eugenia uniflora leaves

The essential oils of Eugenia uniflora leave possess several biological activities but a great variability in their chemical composition is observed. In the present study, gas chromatography and mass spectrometry were applied to examine the essential oil from leaves of four different specimens over the seasons of the year, two of which are located in a habitat of a Brazilian metropolis, and the other two in a natural reserve. The collected data allowed identifying twenty-nine compounds; an aliphatic ketone, sesquiterpenes, fatty acids, hydrocarbons, and phthalate derivatives. Sesquiterpene hydrocarbons and oxygenated sesquiterpenes were the chemical classes prevailing in most samples. The curzerene was observed in higher content (10.5-53.4%) in all samples in which the presence of sesquiterpenes class was confirmed. Phthalate derivatives were identified for the first time in the essential oil of E. uniflora leaves. The occurrence of a common chemical marker for four specimens was not observed. Besides, no compound was observed in the same specimen throughout the seasons, and specimens of the same habitat exhibited different essential oil chemical profiles. According to the multivariate analysis applied to the chemical profile for all individuals in different seasons, four different clusters were identified without dependence on season or location.


INTRODUCTION
The Myrtaceae family is comprised of more than 140 genera and about 5744 species distributed in different regions of the world (The Plant List, 2013a). Eugenia is the largest genus containing about 1011 species, of which 391 occur in Brazilian territory (COPPETEC-UFRJ, 2020;The Plant List, 2013b). Eugenia uniflora ('pitanga'), Eugenia involucrata ('cherry'), and Eugenia caryophyllata ('clove') are the most well-known species due to their *Corresponding author. E-mail : miriamsan@usp.br.

Plant material
E. uniflora leaves (about 60-70 g) from four specimens were collected on the tenth day after the beginning of each season 2018, two of them occurring near the Ayrton Senna highway in the metropolitan city of Sao Paulo, adjacent to the School of Arts, Sciences and Humanities, the University of Sao Paulo (

Leaf essential oils extraction
The hydrodistillation technique was applied to obtain the volatile oil from the fresh leaves using a Clevenger-type apparatus. To extract the essential oils from each sample studied, 50 g of fresh leaves were ground and mixed with 1 L of water. The mixture was transferred to a 3 L round-bottom flask placed on a heating mantle connected to a condenser. After 4 h of heating, the essential oil was collected and dried over anhydrous sodium sulfate (ca. 1 g) (Morais et al., 2019). All these processes were performed once a time with each leaf sample. The essential oils were maintained in sealed vials in a freezer at -27°C to preserve the chemical constituents.

Chromatographic analysis
The chromatographic analyses of the leaf essential oils were performed in a chromatographic system coupled to Shimadzu and Model QP 2020 mass spectrometer. The optimized operating conditions for these analyses were: ZB5HT capillary chromatographic column (30 × 0.25 mm × 0.25), mobile phase flow rate (He): 2.5 ml/min; injector Split ratio 1/25, and the injection volume of 3 µl. The operating temperatures of the equipment are as follows: injector at 260°C, detector at 280°C, and column at programmed temperature starting at 60°C for 1 min, 3°C/min rise to 220°C, later rise from 10°C/min to 280°C and remaining for 1.67 min. The conditions of the mass spectrometer employed were scan detector 1,000; scanning interval of 0.50 fragments and fragments detected in the range of 40 to 550 Da. To identify the chemical composition of the essential oil, two NIST107 and NIST21 libraries were used to compare the mass spectra data. The identification of each peak was assigned only when the similarity was above 80%.
The retention index was calculated by reference to a solution of the C 8 -C 22 homologous series by the Van den Dool and Kratz equation (van Den Dool and Dec. Kratz, 1963). To obtain the relative percentages (%) by peak area normalization of the identified compounds, all samples were analyzed in gas chromatography with a flame ionization detector (GC-FID) using an Agilent CG 6850 system. GC was equipped with the same chromatographic conditions used in the CG-MS analysis.

Principal components analysis (PCA) and hierarchical cluster analysis (HCA) for identified chemical compounds in extracts
PCA and HCA from chemical profile (% w/w) of all essential oil obtained in different seasons and locations (Table 1) were performed. Both analyses aimed to identify similarities and differences among the samples, as well as the chemical reasons for samples' grouping. Scaling and mean-centering to the data were applied to the data before PCA. Wards method was used for HCA. The multivariate data analyses were executed in SIMCA 16 software (Trial version, Sartorius Stedim Data Analytics AB Umeå, Sweden).

Descriptive analysis of primary data
To evaluate the chemical profile of essential oil leaves among individuals of the same habitat, the identified chemical constituents and their contents are presented in Table 1. Thirty compounds present in the essential oil leaf of specimens Eu1, Eu2, Eu3, and Eu4 were identified in the four seasons of the year by CG-MS analysis. The essential oils predominantly contain sesquiterpenes, phthalate derivatives, hydrocarbons, fatty acids, and some of their methyl esters. According to the chemical constituents' data, the essential oils of E. uniflora leaves indicate the expressive presence of sesquiterpenes, while the Wi3 and Wi4 samples showed the identification of only hydrocarbons and/or phthalate derivatives. The Sp1 oil is comprised 38.1% of sesquiterpene hydrocarbons and 49.7% of oxygenated sesquiterpenoids. Nine components in total were identified in the Sp2, which were equal to 61.0% of its content. The chemical profile of the Sp2 oil was constituted mostly by 51.3% of oxygenated sesquiterpenes and 13.4% of sesquiterpene hydrocarbons. The Sp3 essential oil presented 52.1% of oxygenated sesquiterpenes and 4.1% of sesquiterpene hydrocarbons being that germacra-3,7 (11), 9-trien-6one (27.0%) was the major chemical compound. The sp4 oil consisted mainly of 36.5% curzerene, followed by αguaiene (7.7%) and δ-maaliene (2.7%). The sesquiterpene hydrocarbons (49.8%) were the major components in the su1 essential oil composition and 32.0% of oxygenated sesquiterpenes, mainly as ßelemenone (35.6%), spathulenol (16.3 %), germacra-3,7 (11),9-trien-6-one (15.7 %). In the Su2 oil only two components were identified: eudesma-4(14),11-diene (8.4%), and α-selinene (3.6%). Oxygenated sesquiterpenes (51.6%) were the major components in the Su3 oil and sesquiterpene hydrocarbons (36.1%).
Another atypical chemical profile was observed in the Wi4 oil, since only phthalate and fatty acids derivatives were identified. Phthalate derivatives were also detected in Eu3 during fall and winter. Phthalate derivatives have been recognized as contaminants to be utilized as plasticizers and environmental hazards, particularly those with reduced molecular weight, and can be observed in the essential oil constitution (Manayi et al., 2014;Roy, 2020). According to Chen (2004), the difference between the natural and synthetic phthalate derivatives is the abundance of 14C. Phthalic acid dioctyl ester (30) was isolated from Nigella glandulifera and Eichhornia crassipes (Nguyen et al., 2007;Shanab et al., 2010); while bis(2-ethylhexyl) phthalate (29) was isolated from N. glandulifera (Nguyen et al., 2007). There are various bioactivities of these natural phthalates derivatives such as antimicrobial, antioxidant, antitumor, larvicidal, and others (Roy, 2020).    - *Individuals' coding: Sp, Su, Au, Wi stand for spring, summer, autumn, and winter seasons, respectively.1,2,3 and 4 refer to the four different locations of vegetable material harvest, Eu1, Eu2, Eu3, and Eu4, respectively RI: Retention Index; NC: Uncalculated, because the elution of the compound was before the last hydrocarbon of the standard or after de first one.
In our case, the phthalate derivatives in some samples exhibited a high percentage, even higher than the sesquiterpene present in the essential oil. According to Table 1, these compounds' class were the majority in the Wi3 and Wi4 samples. The results obtained from the essential oil analysis of four different individuals, indicate the absence of a biomarker compound, even in individuals from the same habitat. According to Costa et al. (2009), the seasonal study of the essential oils of E. uniflora leaves showed the presence of spathulenol and caryophyllene oxide in dry seasons and seline 1,3,7 (11)-trine-8-one epoxide as the major compound in rainy seasons. However, our data showed the occurrence curzerene in all individuals during spring, but not in other seasons. Moreover, it was the major metabolite or one of it in the composition of the Sp1, Sp2, Sp3 Sp4, Su3, Su4, Au2, Wi1, and Wi2 oils (9.7-53.4%). The individual Eu3 showed a considerable increase in curzerene content from spring to summer (10.5-26.2%) and this was not observed in the other seasons. The highest content was observed during the winter (53.4%) for specimen 2. The data presented in Table 1, it confirmed the predominance of several sesquiterpenes in different seasons for the studied individuals, except in winter for specimens 3 and 4, and autumn for 3. In the winter it was observed the presence of fatty acid derivatives. This is the first report of phthalate derivatives occurrence in the essential oil of E. uniflora leaf. The literature reports the presence of these compounds in the essential oil of other plants such as Silybum marianum seeds, Rizophora flowers (Saranya et al., 2015), and Calycotome villosa subsp. intermedia (Chikhi et al., 2014). There was no similarity between the chemical compositions of specimens from the same habitat. According to De Morais et al. (1996), the chemical composition of essential oils may differ depending on the necessity of the plant to adapt to its habitat and due to chemotypes.

Principal components analysis (PCA) and hierarchical cluster analysis (HCA) for identified chemical compounds in essentials oils
The five more atypical samples were Wi3 (Group 1), Wi4 (Group 1), Su1 (Group 2), Sp1 (Group 3), and Su3(Group 3). The remaining samples showed a similar chemical profile (Figure 1). This finding is also confirmed by the scores plot ( Figure  2A). Overlapping score and loading plots ( Figure  2A and B) with support of biplot graph for the first two principal components (Figure 3) is possible to identify that the Group 1 samples, which are from different locations in winter, were relatively plenty of 4 common compounds myristic acid (21), palmitic acid (23), stearic acid (27) and phthalic acid dioctyl ester (30). However, the Wi3 sample was also abundant in phthalic acid 2,4dimethylpent-3-yl isobutyl ester (25), heneicosane (26), and n-pentacosane (28), while the Wi4 sample was relatively rich in palmitic methyl ester (22) and margaric acid methyl ester (24). Su1 (Group 2) was comprised mainly by 2-butanone (1), guaia-1(10),11-diene (12), spathulenol (15), ß-elemenone (18), and germacra-3,7 (11),9-trien-6-one (20). Group 3 samples from the same location and different seasons (spring and summer) were rich in -elemene (6), germacrene B (14) and viridiflorol (17). A significant number of samples had a high amount of curzerene (Table  1). This volatile oxygenated sesquiterpene has been identified as one of the major constituents in the essential oils of E. uniflora leaves, especially, when are derived from bright red fruit specimens . According to the literature, plants can exhibit changes on their chemical composition by the high CO 2 levels, such as the increase monoterpene concentrations (Idso and Idso, 2000). However, the essential oils from individuals 1 and 3, located in area with high concentration of CO2 by vehicles emission, were not affected the secondary metabolites production. The PCA and  . Score scatter (A) and loading (B) 3D plots corresponding to PCA for the identified compounds in the essential oil samples using relative peak area (%). The four main groups categorized by HCA were also identified using the same color scale (Figure 1).
HCA data showed the season and location are not the major factors to influence the chemical variability of essential oil composition from E. uniflora leaves. This chemical variability can be associated to the ecosystem differences, not only by the biotic and abiotic factors. The secondary metabolites production can depends on the climate changes, which can influence on the soil microflora, all pollinisers and other insects affecting the plant antogeny, adaptation, and including phytochemicals productivities (Thakur et al., 2019).
This should be taking into consideration for commercial applications, specifically in perfumery, where the . Biplot graph for the two first principal components of the chemical profile in "mass percentage (% w/w)" of identified compounds in essential oils. Individuals and chemicals are represented by hexagons and circles, respectively. The coding of figure elements is described in Table 1 and the groups' color in Figure 1 chemical composition of this feedstock has a paramount impact on final product quality attributes (Gallucci et al., 2010). Also, the essential oil of E. uniflora leaves is strongly recommended because of its powerful industrial or pharmaceutical properties, the high variability of essential oils chemistry from different specimens, should be considered during the chemical synthesis or biotechnological products manufacturing (Ochoa-Villarreal et al., 2016). Besides, none of the groups found by the multivariate techniques matched the chemotypes reported for this species (Costa et al., 2009).

Conclusion
In all essential oil samples from E. uniflora leaves, sesquiterpene hydrocarbons and oxygenated sesquiterpenes were prevalent, except for two specimens during the winter and autumn seasons. The curzerene, an oxygenated sesquiterpene, was the most abundant compound in samples with sesquiterpenes. The chemical compositions of specimens from the same habitat are different and without correlations, and there are no direct influence of metabolites production according to the type habitat. Considering the chemical profile variability of essential oils among specimens by principal components analysis, commercial applications should be manufactured using other approaches.