Chemotaxonomy of Indonesian Citrus maxima based on Leaf Essential Oils

Tel and Fax: +62-274-580839 Email: rsusandarini@gmail.com Abstract: Pummelo (Citrus maxima) is one of true citrus species originated from South East Asian regions. Pummelo is known as having high morphological diversity, but lacking of comprehensive publications documenting chemical diversity. The objectives of this study were to reveal essential oil diversity and its implication on the chemotaxonomy of the species. Leaves from forty individual plants were collected from seven districts representing pummelo distribution areas in Indonesia. Isolation of essential oils was done using maceration and solvent extraction procedures. Qualitative and quantitative analysis of the essential oils was conducted using GC-MS. Thirty one compounds were identified, consisted of monoterpenes, sesquiterpenes, diterpenes, ketones, alkenes and fatty acids. Assessment on chemotaxonomical importance of essential oils was carried out using cluster analysis and principal component analysis on volatile compounds. Three chemotypes were recognized, defined as “nerol/loliolide/alloaromadendrenechemotype”, “alpha-selinenechemotype” and “alphapinene/delta-carenechemotype”. Three chemotypes defined in this study was the first reported chemotaxonomic analysis for Indonesian C. maxima.


Introduction
Citrus maxima (Burm.)Merr.,or pummelo, is a fruit crop species originated from Southeast Asian region. From taxonomical perspective, C.maxima is one of "true Citrus species" within the genus Citrus. This taxonomic status has been proven from various studies based molecular data (Pang et al., 2007;Kyndt et al., 2010). C. maxima shows morphological variability and recent study indicated that its variability lead to the recognition of infraspecific grouping (Susandarini et al., 2013). However, there is no chemotaxonomical study focusing on the essential oil variability on this species.
Essential oils are one of the phytochemical compounds used in chemotaxonomic studies (Gonzales et al., 2001;Merle et al., 2004). Reserach on Citrus essential oils are generally conducted in relation to their aromatic properties applied in food industry, cosmetics and perfume (Sawamura et al., 2001;Gogorcena and Ortiz, 2006;Tao et al., 2008). Several studies revealed the use of essential oils from various plant species as food preservatives, antimicrobial agent and natural pesticides.
The potential of essential oils as food preservatives has been reported by Romeo et al. (2010) who studied the antibacterial properties of essential oils from Lippiacitriodora, Cupressussempervirens and Melissa officinalis against Listeria strain. Meanwhile, the application of essential oils in pest management has been noted by Campolo et al. (2014) who reported that a mixture of kaolin and Citrus sinensis essential oil might be an alternative chemical pesticide for durum wheat. Regarding to the potential of essential oils as natural pesticides, Ziaee et al. (2014) investigated the effect of microencapsulation of Cuminum cyminum essential oil on its persistence and insecticidal efficiency againts bettle pests.
Essential oils are generally produced in special glands or secretory tissues and may differ considerably among various plant organs (Nagegowda and Dudareva, 2007). The occurrence of essential oils could be speciesspecific in terms of their accumulating cells, tissues and organs within particular plant species. Previous studies showed that chemical composition of essential oils on some Citrus species were species-specific, or even cultivar-specific (Gonzales et al., 2001;Merle et al., 2004;Tao et al., 2008).
Essential oil compositions from peel and leaves of various Citrus species have been studied for their role as chemical markers for distinguishing taxa at species. Tomi et al. (2008) reported the use of essential oils in determining hybrid status by comparing oil profiles between hybrid plants and their putative parents. Previous studies on the use of essential oils for clarifying taxonomic status and hybrid identification on Citrus have been reported (Sawamura et al., 2001;Gogorcena and Ortiz, 2006;Tao et al., 2008). In this study, the variability of essential oil composition among 40 samples of Indonesian C. maxima cultivars and landraces was examined along with their taxonomic implication in defining infraspecific grouping within the species.

Materials and Methods
Plant materials for this study were collected from 7 districts in 3 different islands representing pummelo's distribution areas in Indonesia. A total of 40 samples were obtained, consisted of both registered cultivars and landraces. Fully developed leaves from healthy individual trees were washed using water and wiped with 70% ethanol prior to extraction procedure to ensure that the samples were free of contaminations. Plant samples used in this study was listed in Table 1.
Extraction of essential oils was performed using modified maceration and solvent extraction procedure adopted from Merle et al. (2004). Two grams o fresh leaves were manually ground into powder. A volume of 3 mL petroleum benzene was added into leaf powder and after setled down for 10 min the mixture was centrifuged at 12,000 rpm for 10 min. The resulted solution was then poured into glass tube. The left over debris was treated with the same procedure for 2 more times. The bulk of solutions collected from three extraction procedures was air-dried for 24 h. Sample dilution using 500 µL of petroleum benzene was done right before injecting oil samples to the GC.
GC-MS analysis of essential oil was carried out using GC HP5890 Series II and Mass Selective Detector HP5972 Series, fitted with HP-5MS 5%-phenyl methyl silicone capillary column (30 m ×25 mm ×0,25 µm). The analysis was performed using splitless method and run with the following program: injection volume 1 µL, injector temperature at 250°C, MS interface temperature at 280°C, initial temperature at 80°C, initial time 1 min, increment rate of 10°C/min, final temperature at 220°C and final time 16 min. Identification of essential oil compounds was carried out by matching mass spectral fragmentation patterns with those from GC-MS data systems and spectral WILEY library.

Data Analysis
The data used for subsequent numerical taxonomy analysis were composition of essential oil compounds and their relative amounts among samples. Calculation of relative amounts for each compound was done on the basis of peak-area ratios. Taxonomic similarities among samples were formulated using Gower's general similarity coefficient. The resulting similarity matrix was then subjected to cluster analysis using Unweighted Pair-Group Method using Arithmetic averages (UPGMA). Principal component analysis was then employed to discern taxonomic value of each compound in determining the grouping of samples. Cluster analysis and principal component analysis were performed using Multivariate Statistical Program version 3.1 (Kovach, 2007).

Results
A total of 40 samples were collected for leaf essential oils analysis. The resulting chromatogram profiles from GC-MS analysis showed considerable variations. Results of compound identification showed that leaf essential oils of C. maxima consisted of 31 compounds. Determination of compound was based on the matching of mass spectral patterns with the minimum similarity of 85%. The 31 identified compounds were comprised of monoterpenes, sesquiterpenes, diterpenes, ketons, alkanes and fatty acids ( Table 2).
Qualitative comparison among three groups of terpenes showed that sesquiterpenes had the highest proportion. Meanwhile, in terms of relative amount, phytol was the found to be a compound with the highest percentage. From the proportion of six compound groups found in leaf essential oil of C. maxima, it was obvious that monoterpenes and sesquiterpenes were dominant.
In this study only volatile components of essential oils were defined as target compound, primarily due to their biological significance as bioactive compound with various medicinal properties. It was therefore, terpenes became groups of target compound and thus were used in the numerical taxonomic analysis for determining the existence of chemotypes within C. maxima. The determination of the chemotypes was based on the taxonomic affinity of cultivars and landraces. A dendrogram resulted from cluster analysis of 21 volatile compounds figuring taxonomic affinity among 40 C. maxima samples showed two major clusters (Fig. 1). The first cluster, designated as cluster A, consisted of 34 samples whereas the second cluster assigned as cluster B comprised of 6 samples. The first cluster could be divided into two subclusters, A1 and A2.  Table 3.  Dendrogram topology with bifurcating pattern formed at similarity value of 0.624 indicated that cluster A and B had distinctive characteristics that clearly differentiate between them. In this regards, the volatile compounds differentiated those two clusters were categorized as taxonomic distinguishing characters. These characters could be identified by observing the eigen value from three Principal Components (PCs) resulted from principal component analysis (Table 3). The percentage of explained variance on the first principal component were 29.94, while the second principal component explained 49.95% of total varienace. A biplot of PCA results (Fig. 2) showed distribution patterns of samples in two dimentional graph as well as represented relative contribution of each essential oil compunds in differentiating samples into groups of chemotypes.

Eigen values of 21 volatile compounds in C. maxima leaf essential oils Eigen value --------------------------------------------------------------------------
Criteria for particular compound to be categorized as distinguishing character adopted in this study, or herewith as chemical marker, was at least having eigen value of 0.2 (absolute value). Based on the eigen values, five volatile compounds were found to be differentiating characters as indicated by their eigen values. These compounds were nerol, loliolide, allo-aromadendrene, nerolidol and neophytadiene. They represented five major compounds separating cluster A from B and thus considered as distinguishing chemical characters. Referring to the combining results of cluster analysis and principal analysis, it was noticeably recognized the existence of three groups representing three chemotypes within C. maxima, defined here as subcluster A1, subcluster A2 and the smaller cluster B. Since the recognition of chemotypes should be based on chemical markers, then the eigen value from principal component analysis was used for this purpose.

Discussion
Variations on the GC-MS profiles of essential oils extracted from C. maxima leaves indicated infraspecific variability of the species. The observed variability was due to differences in the composition of essential oil compounds as well as their relative amount among individual samples. Differences in the composition of essential oil within plant species are common. Lόpez et al. (2009) pointed out that composition of essential oil may vary due to differences on geographical origin, altitude, infraspecific variation, soil type, nutrient status of the plants and phenology.
The composition of essential oils which comprised of monoterpenes, sesquiterpenes, diterpenes, ketons, alkanes and fatty acids was in accordance to Mondello et al. (2005) who reported that Citrus essential oils had two components, the volatile and non volatile compounds. The volatile compounds were mainly consisted of monoterpenes and sesquiterpenes with their oxygenated derivates, whereas the non volatile compounds were aliphatic hydrocarbons, aldehydes, ketons, acids and esters. Regarding on the number of compounds, Bakkali et al. (2008) noted that plant essential oils generally consisted of 20 to 60 components with various concentrations. Quantitaive analysis of various compound groups indicated that monoterpenes and sesquiterpenes were dominant. This result was similar to the composition of essential oils from Citrus nobilis, in which monoterpenes and sesquiterpenes were dominant (Liu et al., 2013). Similar report was found on the essential oils of Citrus limetta, with sesquiterpenes being the most abundant compared to other compounds (Colecio-Juárez et al., 2012).
The grouping of samples resulted from cluster analysis, combined with results of principal component analysis demonstrated the existence of three chemotypes within C. maxima based on leaf essential oils. Determination of chemotypes based on the observed variations of essential oil composition has been the subject of discussion by several authors. Grayer et al. (1996) suggested a combination of major essential oil components rather than a single and dominant compound as a basis of defining chemotypes. Similar argument was pointed out by Labra et al. (2004) by mentioning that chemotype determination should not be based on a single essential oil component since plants generally have two or more major compounds of relatively equal amounts. The use of principal component analysis in determining chemotypes has been suggested by Boushama et al. (2006) for Lavanduladentata. In the same way, Wheeler et al. (2007) used principal component analysis as a basis for examining infraspecific variation and defining chemotypes based on leaf essential oils for Melaleucaquinquenervia. From taxonomical perspective, there were criteria for the recognition of chemotype when dealing with variations observable within a species. A chemotype is defined as individuals within a species which have distinctive chemical phenotype differed from the other members of the species (Keefover-Ring et al., 2009).
A number of studies on the use of essential oils for describing chemotypes in various Citrus species have been reported. In C. reticulata three chemotypes were identified based on analysis of leaf essential oils from 58 cultivars. They were recognized as "sabinene/linalool", "c-terpinene/linalool" and "methyl N-methylanthranilate" chemotypes, respectively (Lota et al., 2001). Similarly, analysis on leaf essential oils of C. limon resulted in characterization of two chemotypes, whereas four chemotypes were identified on C. aurantifolia (Lota et al., 2002). Another chemotaxonomic study on C. reticulata based on 13 major compounds of essential oils resulted in the classification of cultivar groups known as "sabinene/delta-carene", "gamma-terpinene" and "sabinene" chemotypes (Kasali et al., 2010). These studies indicated the existence of variations on essential oils at species level and their chemotaxonomic applications on genus Citrus.
Observation on the grouping of samples into three chemotypes did not correspond to their geographical origin. This result implied that infraspecific variation on leaf essential oils defined as chemotypes in this study indicated the existence of a genetic background. This inference was formulated referring to an argument that genetically-determined chemotypes was indicated when individuals representing populations from different geographical origin with different environmental conditions clustered within one chemotype (De Martino et al., 2009;Chauhan et al., 2011). Result of this study corresponded to previous chemotaxonomic study on Citrus cultivars from various countries which confirmed that instead of affected by geographical origin, variations on essential oils suggested differences in chemotypes (Bhuiyan et al., 2009). An important aspect on taxonomic value of essential oil variations within a species was revealed in this study, suggesting the role of essential oils as potential chemical marker in plant classification.

Conclusion
Qualitative and quantitative analysis of leaf essential oils in this study demonstrated the identification of three chemotypes within C. maxima. Results of this study, therefore, supported the recognition of essential oils as chemotaxonomic marker. Phytochemical analysis in this study added scientific value of C. maxima leaf essential oils, especiallytheir potential use as aromatic components in food, aromatherapy and perfume.